LSCAS Main Book File Qv3441 Student Guide

User Manual:

Open the PDF directly: View PDF .
Page Count: 234 [warning: Documents this large are best viewed by clicking the View PDF Link!]

Front cover
Contents
Trademarks
Course description
Agenda
Unit 1. Virtual I/O Server Configuration
Unit 2. Virtual SCSI Configuration
Unit 3. File-backed Storage Devices
Unit 4. Virtual Fibre Channel Storage Devices
Unit 5. Virtual Ethernet Adapter Configuration
Unit 6. Shared Ethernet Adapter Configuration
Unit 7. Virtual I/O Server Maintenance
Appendix A. Checkpoint solutions

IBM PowerVM



Virtual I/O Server I:

Configuring Virtual Devices

(Course code QV344)

Student Notebook

ERC 1.0

V5.4.0.3

cover

UNIX Software Service Enablement

򔻐򗗠򙳰 Front cover

Student Notebook

July 2012 edition

The information contained in this document has not been submitted to any formal IBM test and is distributed on an “as is” basis without

any warranty either express or implied. The use of this information or the implementation of any of these techniques is a customer

responsibility and depends on the customer’s ability to evaluate and integrate them into the customer’s operational environment. While

each item may have been reviewed by IBM for accuracy in a specific situation, there is no guarantee that the same or similar results will

result elsewhere. Customers attempting to adapt these techniques to their own environments do so at their own risk.

This document may not be reproduced in whole or in part without the prior written permission of IBM.

Note to U.S. Government Users — Documentation related to restricted rights — Use, duplication or disclosure is subject to restrictions

set forth in GSA ADP Schedule Contract with IBM Corp.

Trademarks

IBM® and the IBM logo are registered trademarks of International Business Machines

Corporation.

The following are trademarks of International Business Machines Corporation, registered in

many jurisdictions worldwide:

Linux is a registered trademark of Linus Torvalds in the United States, other countries, or

both.

UNIX is a registered trademark of The Open Group in the United States and other

countries.

Other product and service names might be trademarks of IBM or other companies.

Active Memory™ AIX 6™ AIX®

developerWorks® Express® POWER Hypervisor™

Power Systems™ Power® PowerHA®

PowerPC® PowerVM® POWER6®

POWER7® Redbooks® System p®

System Storage® Tivoli®

Student Notebook

V7.0

TOC

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Contents

Trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

Course description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi

Agenda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii

Unit 1. Virtual I/O Server Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1

Unit objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-3

What is virtual I/O? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4

Virtual devices and scalability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-5

Benefits and considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-6

Virtual devices overview (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-7

Virtual devices overview (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8

Virtual I/O Server overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-9

Virtual I/O Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-10

Virtual I/O Server support portal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-11

Creating a Virtual I/O Server partition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-12

Installing the Virtual I/O Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-13

Installing the Virtual I/O Server from media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-14

Installing from the HMC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-15

Command Line Interface (CLI) (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-16

Command Line Interface (CLI) (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-17

Configuring the Virtual I/O Server (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-18

Configuring the Virtual I/O Server (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-19

Configure networking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-20

Resource Monitoring and Control (RMC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-21

RMC management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-22

RMC reinitialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-23

Checkpoint (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-24

Checkpoint (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-25

Exercise 1 - VIOS configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-26

Unit summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-27

Unit 2. Virtual SCSI Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1

Unit objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3

Virtual SCSI overview (1 of 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-4

Virtual SCSI overview (2 of 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5

Virtual SCSI overview (3 of 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-6

Creating virtual SCSI devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7

Step 1: Virtual SCSI server adapter (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8

Step 1: Virtual SCSI server adapter (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-9

Step 2: Virtual SCSI client adapter (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-10

Step 2: Virtual SCSI client adapter (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-11

Step 3: Verify virtual SCSI server adapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-12

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

iv VIOS © Copyright IBM Corp. 2012

Step 4a: Create storage resources with LVM . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-13

Step 4b: Create storage resources with storage pools (1 of 2) . . . . . . . . . . . . . . .2-14

Step 4b: Create storage resources with storage pools (2 of 2) . . . . . . . . . . . . . . .2-15

List storage resource configuration (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-16

List storage resource configuration (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-17

Step 5: Create the virtual target device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-18

Verify the mapping of virtual targets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-19

Step 6: Virtual SCSI disks (client view) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-20

Manage Virtual devices with cfgassist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-21

Check configuration from HMC (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-22

Check configuration from HMC (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-23

Discover backing storage example (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-24

Discover backing storage example (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-25

Removing virtual devices (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-26

Removing virtual devices (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-27

Managing devices from the HMC (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-28

Managing devices from the HMC (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-29

Create virtual disk example (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-30

Create virtual disk example (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-31

Modify assignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-32

Creating storage pools / volume groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-33

Physical volumes (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-34

Physical volumes (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-35

Optical devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-36

Virtual SCSI storage and LVM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-37

Dynamically adding virtual SCSI storage (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . .2-38

Dynamically adding virtual SCSI storage (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . .2-39

VIOS storage groups (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-40

VIOS storage groups (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-41

Checkpoint (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-42

Checkpoint (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-43

Exercise 2 - Virtual SCSI configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-44

Unit summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-45

Unit 3. File-backed Storage Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-1

Unit objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-3

File-backed virtual devices (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-4

File-backed virtual devices (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-5

Creating file-backed virtual disks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-6

Create FB virtual disks example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-7

View mapping of file-backed virtual disk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-8

View storage pools and backing devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-9

FB virtual disks remove commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-10

File-backed virtual optical devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-11

Create FB virtual optical device (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-12

Create FB virtual optical device (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-13

View repository information (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-14

View repository information (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-15

Student Notebook

V7.0

TOC

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

View virtual optical drive mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-16

Manage repository with cfgassist utility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-17

Viewing FB configuration from the HMC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-18

Manage FB devices with the HMC (1 of 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-19

Manage FB devices with the HMC (2 of 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-20

Manage FB devices with the HMC (3 of 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-21

Adding image file from HMC GUI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-22

Checkpoint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-23

Exercise 3 - File-backed storage devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-24

Unit summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-25

Unit 4. Virtual Fibre Channel Storage Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1

Unit objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3

Virtual Fibre Channel adapter (1 of 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4

Virtual Fibre Channel adapter (2 of 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-5

Virtual Fibre Channel adapter (3 of 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-6

NPIV benefits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-7

NPIV considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-8

NPIV overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-9

Add virtual Fibre Channel adapters (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-10

Add virtual Fibre Channel adapters (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-11

View available N_Ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-12

Map virtual Fibre Channel server adapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-13

View Fibre Channel mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-14

Zone LUNs on the SAN using client WWPNs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-15

Virtual Fibre Channel DLPAR operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-16

Activating all WWPNs to aid SAN mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-17

Logging into N_Ports from the HMC CLI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-18

Alternate procedure to activate WWPNs (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . 4-19

Alternate procedure to activate WWPNs (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . 4-20

Activating WWPNs from the HMC GUI (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-21

Activating WWPNs from the HMC GUI (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-22

Monitoring and tuning virtual Fibre Channel adapters (1 of 2) . . . . . . . . . . . . . . . 4-23

Monitoring and tuning virtual Fibre Channel adapters (2 of 2) . . . . . . . . . . . . . . . 4-24

Checkpoint (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-25

Checkpoint (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-26

Exercise 4 - Virtual Fibre Channel storage devices . . . . . . . . . . . . . . . . . . . . . . . 4-27

Unit summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-28

Unit 5. Virtual Ethernet Adapter Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1

Unit objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3

Virtual Ethernet overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-4

POWER Hypervisor Ethernet switch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5

Virtual LAN (VLAN) overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-6

IEEE 802.1Q VLAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-7

Virtual Ethernet configuration example (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-8

Virtual Ethernet configuration example (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-9

Packets leaving virtual Ethernet hosts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-10

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

vi VIOS © Copyright IBM Corp. 2012

Packet delivered to virtual Ethernet hosts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-11

Virtual Ethernet adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-12

Creating a virtual Ethernet adapter (1 of 6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-13

Creating a virtual Ethernet adapter (2 of 6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-14

Creating a virtual Ethernet adapter (3 of 6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-15

Creating a virtual Ethernet adapter (4 of 6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-16

Creating a virtual Ethernet adapter (5 of 6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-17

Creating a virtual Ethernet adapter (6 of 6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-18

View adapter and interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-19

Adding VLAN access (1 of 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-20

Adding VLAN access (2 of 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-21

Adding VLAN access (3 of 3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-22

Virtual Ethernet switches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-23

Manage virtual switches (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-24

Manage virtual switches (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-25

View virtual LANs (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-26

View virtual LANs (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-27

Checkpoint (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-28

Checkpoint (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-29

Exercise 5 - Virtual Ethernet adapter configuration . . . . . . . . . . . . . . . . . . . . . . . .5-30

Unit summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5-31

Unit 6. Shared Ethernet Adapter Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-1

Unit objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-3

Shared Ethernet adapter overview (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-4

Shared Ethernet adapter overview (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-5

Configure virtual Ethernet adapter for SEA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-6

Create Shared Ethernet adapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-7

View SEA configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-8

SEA with multiple virtual adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-9

SEA with multiple physical adapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-10

Create SEA with cfgassist tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-11

Configuring the VIOS interface (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-12

Configuring the VIOS interface (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-13

Using HEA logical ports for SEA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-14

Troubleshooting SEA "cfgsea"error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-15

Shared Ethernet adapter with VLANs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-16

Configure SEA using the HMC (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-17

Configure SEA using the HMC (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-18

View SEA configuration from HMC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-19

Removing the SEA using the HMC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-20

Threading/non-threading (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-21

Threading/non-threading (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-22

Monitoring SEA traffic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-23

Checkpoint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-24

Exercise 6 - Shared Ethernet adapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-25

Unit summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6-26

Student Notebook

V7.0

TOC

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Unit 7. Virtual I/O Server Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1

Unit objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-2

Maintaining the Virtual I/O Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-3

VIOS backup options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-4

Using backupios for backups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-5

Manual backup of VIOS device mappings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-6

VIOS manual backup examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-7

Restore options with backupios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-8

Backup and restore with viosbr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-9

viosbr command examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-10

Updating the Virtual I/O Server (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-11

Updating the Virtual I/O Server (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-12

View installed levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-13

Sizing the Virtual I/O Server (1 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-14

Sizing the Virtual I/O Server (2 of 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-15

Monitoring Virtual I/O Server resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-16

Checkpoint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-17

Exercise 7 - Virtual I/O Server Maintenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-18

Unit Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-19

Appendix A. Checkpoint solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

viii VIOS © Copyright IBM Corp. 2012

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

TMK Trademarks

The reader should recognize that the following terms, which appear in the content of this

training document, are official trademarks of IBM or other companies:

IBM® and the IBM logo are registered trademarks of International Business Machines

Corporation.

The following are trademarks of International Business Machines Corporation, registered in

many jurisdictions worldwide:

Linux is a registered trademark of Linus Torvalds in the United States, other countries, or

both.

UNIX is a registered trademark of The Open Group in the United States and other

countries.

Other product and service names might be trademarks of IBM or other companies.

Active Memory™ AIX 6™ AIX®

developerWorks® Express® POWER Hypervisor™

Power Systems™ Power® PowerHA®

PowerPC® PowerVM® POWER6®

POWER7® Redbooks® System p®

System Storage® Tivoli®

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

x VIOS © Copyright IBM Corp. 2012

Student Notebook

V7.0

TOC

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Course description

IBM PowerVM Virtual I/O Server I: Configuring Virtual Devices

Purpose

This course describes the concepts and implementation details of

virtual I/O devices on IBM Power Systems servers. The course details

the installation and configuration of the Virtual I/O Server and the

implementation of virtual SCSI devices, virtual Fibre Channel devices,

virtual Ethernet adapters, and Shared Ethernet adapters. The Virtual

I/O Server software is part of the PowerVM system feature. Students

will configure virtual SCSI disks using physical and logical volumes,

optical media devices, and files as backing storage. Students will learn

how to create and use the Optical Media Repository and configure

networking with virtual Ethernet adapters and shared Ethernet

adapters. In addition, students will configure virtual Fibre Channel

adapters, which support N_Port Identifier Virtualization (NPIV).

Students will use both the HMC and the Virtual I/O Server command

line interface to configure partitions to implement and manage virtual

devices.

Audience

The audiences for this training include System p and AIX/Linux

technical support individuals, AIX or Linux system administrators,

system architects and engineers, pre-sales technical support staff and

performance analysis individuals.

Prerequisites

Student must already know the basics of configuring and managing

logical partitions. This prerequisite can be met by attending the

following course:

• AHQV332 Managing Logical Partitions (LPARs) on Power Systems

with HMC V7

Objectives

After completing this course, you should be able to:

• Configure virtual SCSI devices that are backed by physical

volumes, logical volumes, optical media devices, and file-backed

devices

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

xii VIOS © Copyright IBM Corp. 2012

• Configure the Optical Media Repository, load a CD image, and use

it to install a new AIX partition

• Configure virtual Fibre channel devices using NPIV technology

• Configure virtual Ethernet adapters

• Configure a Shared Ethernet adapter

• Perform Virtual I/O Server maintenance operations

Student Notebook

V7.0

TOC

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Agenda

Welcome

Unit 1 - Virtual I/O Server Configuration

Exercise 1 - Virtual I/O Serve Configuration

Unit 2 - Virtual SCSI Configuration

Exercise 2 - Virtual SCSI Configuration

Unit 3 - File-backed Storage Devices

Exercise 3 - File-backed Storage Devices

Unit 4 - Virtual Fibre Channel Storage Devices

Exercise 4 - Virtual Fibre Channel Storage Devices

Unit 5 - Virtual Ethernet Adapter Configuration

Exercise 5 - Virtual Ethernet Adapter Configuration

Unit 6- Shared Ethernet Adapter

Exercise 6 - Shared Ethernet Adapter

Unit 7 - VIOS Maintenance

Exercise 7 - VIOS Maintenance

Wrap up / Evaluations

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

xiv VIOS © Copyright IBM Corp. 2012

Text highlighting

The following text highlighting conventions are used throughout this book:

Bold Identifies file names, file paths, directories, user names,

principals, menu paths and menu selections. Also identifies

graphical objects such as buttons, labels and icons that the

user selects.

Italics Identifies links to web sites, publication titles, is used where the

word or phrase is meant to stand out from the surrounding text,

and identifies parameters whose actual names or values are to

be supplied by the user.

Monospace Identifies attributes, variables, file listings, SMIT menus, code

examples and command output that you would see displayed

on a terminal, and messages from the system.

Monospace bold Identifies commands, subroutines, daemons, and text the user

would type.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty Unit 1. Virtual I/O Server Configuration

What this unit is about

This unit describes the concepts and features of the Virtual I/O Server

and virtual I/O devices. It explains why virtualized I/O is important for

scaling large server environments and details the installation and

configuration of a Virtual I/O Server partition in preparation for

configuring virtual devices.

What you should be able to do

After completing this unit, you should be able to:

• List the reasons for implementing virtual I/O

• Describe virtual I/O devices

• Describe the function of the Virtual I/O Server

• Describe the installation of the Virtual I/O Server product including

configuring the Virtual I/O Server software

How you will check your progress

Accountability:

• Checkpoint questions

• Machine exercises

References

IBM Power Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp

Virtual I/O Server product documentation from the IBM Power

Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp?t

opic=/p7hb1/iphb1kickoff.htm

IBM Support: Fix Central:

http://www.ibm.com/support/fixcentral/

IBM Support Portal:

http://www.ibm.com/support/entry/portal

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-2 VIOS © Copyright IBM Corp. 2012

IBM Support Portal: PowerVM Virtual I/O Server:

https://www.ibm.com/support/entry/myportal/Overview/Software/Other

_Software/PowerVM_Virtual_I~O_Server

The following IBM Redbooks documents can be found at:

http://www.redbooks.ibm.com

REDP4340 Integrated Virtual Ethernet Adapter Technical Overview

and Introduction

SG24-7940 IBM PowerVM Virtualization Introduction and

Configuration

SG24-7590 IBM PowerVM Virtualization Managing and Monitoring

SG24-6615 A Practical Guide for Resource Monitoring and Control

(RMC)

SG24-7825 PowerVM Migration from Physical to Virtual Storage

REDP4194 IBM System p Advanced POWER Virtualization

(PowerVM) Best Practices

This list of third party software products which are allowed to be

installed in the VIOS is available from IBM’s W3 site:

http://dbluedst.pok.ibm.com/DCF/isg/isgintra.nsf/all/T1010620

VIOS developerWorks link:

http://www.ibm.com/developerworks/wikis/display/virtualization/VIO

PowerHA support for NPIV link:

http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/FLA

SH10691

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 1-1. Unit objectives QV3441.0

Notes:

UNIX Software Service Enablement

Unit objectives

After completing this unit, you should be able to:

 List the reasons for implementing virtual I/O

 Describe virtual I/O devices

 Describe the function of the Virtual I/O Server

 Describe the installation of the Virtual I/O Server product

including configuring the Virtual I/O Server software

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-4 VIOS © Copyright IBM Corp. 2012

Figure 1-2. What is virtual I/O? QV3441.0

Notes:

Virtual I/O

Virtual I/O is the term used to describe the ability to share physical I/O resources

between partitions. When referring to the virtual devices available to clients from a

Virtual I/O Server, the physical resources being shared are adapter cards or devices

configured in the Virtual I/O Server partition.

AIX V5.3 and all the supported Linux versions (SLES 9, SLES 10, RHEL 3 Update 3,

and RHEL 4) and up support the ability to be a virtual client partition utilizing virtual I/O

from a Virtual I/O Server.

UNIX Software Service Enablement

What is virtual I/O?

 Virtual I/O allows the sharing of physical resources (adapters and

devices) between partitions on the same server

 Allows servers to scale to large numbers of partitions because of reduced

hardware requirements

 Most virtual devices depend on a Virtual I/O Server partition to host the

physical adapters

 Example 1: One disk can provide storage to one or more LPARs

 Example 2: One physical Ethernet adapter can be shared by

multiple partitions to access the external network

 Virtual I/O functionality is available on POWER5 and higher

processor-based systems

 Virtual I/O Server client partitions must be running a minimum of:

AIX V5.3, Linux SLES 9 or RHEL 3, or IBM i 6.1

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 1-3. Virtual devices and scalability QV3441.0

Notes:

System limitation example

As an example, the IBM Power 750 server supports up to 32 POWER7 processors. If

configured with the PowerVM Standard Edition (or above) feature, this system is

capable of running up to 320 partitions concurrently. However the machine supports a

maximum of 52 PCI-X adapter slots, so with a two slots per partition minimum, the

maximum number of partitions that could be created without the use of virtual I/O would

be 26.

UNIX Software Service Enablement

Virtual devices and scalability

 Virtual devices provide scalability

 POWER5 and POWER6 architectures support up to 254 partitions

 POWER7 architecture supports up to 1024 partitions

 Hardware supports more partitions than there are PCI slots to support

them:

 Slots can be the limiting factor

 For example, each partition requires at minimum:

 1 slot for accessing the boot disk (typically SCSI or Fibre Channel)

 1 slot for Ethernet adapter

 System example:

 A fully configured IBM Power 750 server:

 Supports up to 320 partitions

 Supports up to 50 PCI-X slots  Not enough for all 320 partitions without

using virtual devices

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-6 VIOS © Copyright IBM Corp. 2012

Figure 1-4. Benefits and considerations QV3441.0

Notes:

Benefits

When new partitions can be configured to use virtualized I/O resources, it allows them

to be configured in a timely manner, since no physical reconfiguration of the system

(that is, moving adapter cards and cables) is required. And when LPARs can share

adapters, this saves the customer money. The use of virtualized I/O facilitates server

consolidation because it permits multiple client partitions to reside on a single machine,

making efficient use of shared resources.

Considerations

Prevent single adapters or the Virtual I/O Server partition from becoming single points

of failure for multiple clients by using redundancy. More processing resources are

consumed for virtual devices. Configuring and managing virtual devices is more

complex than using native devices and the tools are different than those used in other

operating systems.

UNIX Software Service Enablement

Benefits and considerations

 Benefits of virtualized I/O

 Partitions can be created without requiring additional physical I/O resources

 Economical I/O model:

 Efficient utilization of resources through sharing

 Facilitates server consolidation

 Facilitates functions which require virtual I/O such as Live Partition Mobility

and Active Memory Sharing

 Requires the IBM PowerVM Enterprise Edition

 Allows client attachment to previously unsupported storage solutions

 Example: A storage solution supported by AIX, but not Linux or

IBM i, could be made available as virtual SCSI disks

 Considerations for virtualized I/O

 Plan and design the configuration for availability

 Assign adequate resources to the Virtual I/O Server

 System support and administration staff must be trained

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 1-5. Virtual devices overview (1 of 2) QV3441.0

Notes:

Special-purpose virtual devices

LPARs have two pre-configured virtual serial adapters to support virtual console

access. Do not remove these and there is no need to create additional virtual serial

adapters. As of POWER6 server firmware release 320, every VIOS partition will have at

least one Virtual Asynchronous Services Interface (VASI) adapter. These devices can

be seen in the VIOS operating system only.

Virtual devices that do NOT require a Virtual I/O Server

The virtual Ethernet adapter is supported on POWER5 and higher processor-based

server partitions running AIX V5.3 or higher or Linux. No actual Ethernet hardware is

required to implement this feature. The Integrated Virtual Ethernet (IVE) adapter is

available on most POWER6 and POWER7 processor-based servers. This adapter is

also known as the Host Ethernet Adapter (HEA). IVE logical ports are supported in

partitions running AIX V5.2 and higher and Linux.

UNIX Software Service Enablement

Virtual devices overview (1 of 2)

 Pre-configured special-purpose virtual devices:

Virtual serial adapter

 On all partitions

 For HMC to LPAR virtual console connection

VASI adapter

 On VIOS type partitions

 High level administrative access to VIOS

 Currently used in Live Partition Mobility and Active Memory Sharing

configurations

 Virtual devices that do NOT require a Virtual I/O Server:

Virtual Ethernet Adapter

 Partitions on the same system communicate without using physical Ethernet

adapters

Integrated Virtual Ethernet Adapter (POWER6 and POWER7)

 Ethernet virtualized in hardware (supports 16-32 LPARs)

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-8 VIOS © Copyright IBM Corp. 2012

Figure 1-6. Virtual devices overview (2 of 2) QV3441.0

Notes:

Virtual devices that DO require a Virtual I/O Server

Virtual SCSI devices are backed by physical devices on the Virtual I/O Server. Backing

storage is the term for these physical devices that will be used as virtual devices for

client partitions. Even though SCSI is the protocol used for the virtualization, the actual

backing storage devices do not need to be SCSI devices. See the Fix Central web site

for a list of current supported backing devices.

The Shared Ethernet adapter is a network bridge device that can connect the virtual

Ethernet traffic on a server to a physical Ethernet adapter and have this traffic bridge to

an external network. It operates at the layer 2 or data link layer of the OSI seven layer

network model.

The virtual Fibre Channel adapter capability allows client partitions to access storage

area network (SAN) devices using NPIV. NPIV is an industry standard technology used

in Fibre Channel networks. Each partition is identified by a unique World Wide Port

Name (WWPN) as if it had its own physical HBA.

UNIX Software Service Enablement

Virtual devices overview (2 of 2)

 Virtual devices that DO require a Virtual I/O Server:

Virtual SCSI

 Virtual I/O Server partition uses logical volumes, physical volumes, optical,

tape, or file devices to provide the backing storage for virtual SCSI devices

presented to client partitions

Shared Ethernet Adapter

 Layer 2 bridge function to connect internal virtual Ethernet with external

physical network

Virtual Fibre Channel Adapter

 Provides virtualization of Fibre Channel adapter ports (N_Ports) called N_Port

Identifier Virtualization (NPIV)

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 1-7. Virtual I/O Server overview QV3441.0

Notes:

Virtual I/O Server description

Virtual I/O Server partitions are not intended to run applications other than those which

directly support the VIOS functions. When configuring the VIOS partition, allocate all of

the physical adapters to it which will be used to provide virtual devices. Once installed

on disk, the VIOS image will only boot in a special type of partition that provides the

special Virtual I/O Server environment.

Client partitions may have a mix of physical adapters and virtual adapters, unless there

is some other requirement (such as Live Partition Mobility or Active Memory Sharing

configurations) which prohibits the use of native physical adapters in the client partition.

The graphic in the visual above shows different types of partitions some of which are

using the Virtual I/O Server to provide access to devices. All connections between the

Virtual I/O Server and its clients for access to devices is done through the Hypervisor.

UNIX Software Service Enablement

Virtual I/O Server overview

 Virtual I/O Server partition is allocated the physical I/O

adapters used to access devices that will be virtualized

 Client LPARs may

have a mix of

physical and

virtual devices

Hardware

Management

Console

(HMC)

POWER Hypervisor

Linux AIX6 AIX7

I/O

Storage Network

LAN, WAN, ...

I/O

Storage Network

AIX 5.3

I/O

Storage Network

I/O

Storage Network

AIX 5.3

Linux

AIX 6

Linux

AIX 7

Virtual I/O paths

Virtual I/O

Server

Partition

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-10 VIOS © Copyright IBM Corp. 2012

Figure 1-8. Virtual I/O Server QV3441.0

Notes:

VIOS software

The Virtual I/O Server product media is supplied when one of the PowerVM features is

purchased as part of a machine order or upgrade. The Virtual I/O Server is

implemented as a customized AIX mksysb image, however, the interface to the system

is abstracted using a secure shell-based Command Line Interface (CLI).

The Virtual I/O Server partition may be installed by activating the partition and booting

directly from the product media. The partition may also be installed from the HMC or

from a NIM server.

UNIX Software Service Enablement

Virtual I/O Server

 Component of PowerVM Express, Standard, and Enterprise Edition

system features

 Supplied as an AIX mksysb image on DVD media

 Installed from the HMC, the managed system, or a NIM server

 Installed as a customized AIX-based appliance

 Special command line interface obscures the complexity and changes of

underlying operating system commands

 Can only be run in Virtual I/O Server type partitions

 Error displayed if you try to boot the Virtual I/O Server image in an

AIX/Linux partition

 Contains Integrated Virtualization Manager (IVM) GUI functions for

non-HMC managed systems

 Maximum number of Virtual I/O Server partitions

 Configurations with more than 10 Virtual I/O Server partitions are not

supported

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 1-9. Virtual I/O Server support portal QV3441.0

Notes:

Support web site

The IBM Support Portal is a unified, customizable view of all technical support tools

and information for all IBM systems, software, and services. The IBM Support Portal is

located at: http://www.ibm.com/support

Select the products that you are interested in, such as PowerVM Virtual I/O Server and

search the technical areas of interest. This portal provides a variety of links including:

- Downloads (from the Fix Central web site)

- Documentation

- Troubleshooting tips

To go directly to the IBM Support Portal - PowerVM Virtual I/O Server page:

https://www.ibm.com/support/entry/myportal/Overview/Software/Other_Software/Power

VM_Virtual_I~O_Server

UNIX Software Service Enablement

Virtual I/O Server support portal

 Links to the latest support information:

 PowerVM Virtual I/O Server

Overview

 Downloads (Fix Central)

 Troubleshooting tips

 Documentation

 Forums

 Planning

 Installation

Location:

https://www.ibm.com/support/entry/myportal/Overview/Softwa

re/Other_Software/PowerVM_Virtual_I~O_Server

Location:

https://www.ibm.com/support/entry/myportal/Overview/Softwa

re/Other_Software/PowerVM_Virtual_I~O_Server

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-12 VIOS © Copyright IBM Corp. 2012

Figure 1-10. Creating a Virtual I/O Server partition QV3441.0

Notes:

Creating the Virtual I/O Server partition

A partition that is to be used as a Virtual I/O Server must be configured as a VIO Server

type partition. This task will only be available if the PowerVM feature has been

authorized for the system. To verify this authorization, go to the Capabilities tab in the

managed system’s properties sheet. If Virtual I/O Server Capable is set to True, then

the system has been authorized for this feature.

A minimum of 30 GB of disk storage is required to install the Virtual I/O Server v2.2

product. The partition may use dedicated or shared processors and 768 MB to 1 GB of

memory is typically the minimum amount or memory needed. However, if using VIOS

Shared Storage Groups, the minimum memory requirement is 4 GB.

Once installed on disk, the Virtual I/O Server image will only boot in a partition

configured with the correct environment type. It will not boot in partitions suitable for

other operating systems. Additionally, other operating systems will not boot in a partition

that is configured with the Virtual I/O Server environment.

UNIX Software Service Enablement

Creating a Virtual I/O Server partition

 Use the Create Logical Partition ->

VIO Server task to create a VIOS type

partition

 Indicate if this

VIO Server will

be used for Live

Partition Mobility

 Configure other resources as you would for any partition

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 1-11. Installing the Virtual I/O Server QV3441.0

Notes:

Virtual I/O Server installation

You must have purchased the PowerVM Express, Standard, or Enterprise Edition

feature to get the VIO Server media. Once you have the media, you can choose from

the different install methods, which are described further in the next couple of visuals.

All the resources required for the install are available on the base media.

•

UNIX Software Service Enablement

Installing the Virtual I/O Server

 To install the Virtual I/O Server, you need the media, a Virtual I/O

Server partition, and an install method:

 Install directly from media

 Install from the HMC CLI (installios)

 Install from an AIX NIM master

 The mksysb install image and other install resources are on the

media

 The VIOS installation media includes an expansion pack that delivers

additional VIO Server security functionality:

 Kerberos (Network Authentication Service for users and Client and Server

Applications)

 LDAP (Lightweight Directory Access Protocol client and server

functionality) and

 SNMP V3 (Simple Network Management Protocol Version 3)

 To migrate from VIOS V1.x to V2.x, use the migration DVD

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-14 VIOS © Copyright IBM Corp. 2012

Figure 1-12. Installing the Virtual I/O Server from media QV3441.0

Notes:

Installation from media

The Virtual I/O Server may be installed directly from the supplied product media. In this

case, the partition being used should contain an adapter slot that has a connected DVD

device. Insert the product media into the DVD drive, and then activate the partition.

When starting the partition, select the checkbox to open a terminal window. Be sure to

click the Advanced button and override the default boot mode with a selection of

Diagnostic with default boot list. This will cause the partition to boot from the DVD

device.

When the terminal window appears, you will be prompted to enter a number to select

the terminal as the console for the newly activated partition and then to select a

language. Next, at the Start Install Now with Default Settings prompt, press Enter to

install the default PowerVM Express Edition. Select Change/Show Installation

Settings and Install to change the installation and system settings.

Once the installation is complete, the partition will reboot.

UNIX Software Service Enablement

 Place Virtual I/O Server product media in system DVD drive

 Ensure slot connected to DVD drive is in partition and available:

 Activate the partition and select:

 Boot mode = Diagnostic with default boot list

 Open a terminal window

 Interact with presented menu choices to complete the installation

 Partition will reboot when installation is complete

Installing the Virtual I/O Server from media

IBM Power 720

example

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 1-13. Installing from the HMC QV3441.0

Notes:

Additional installation options

Place the VIOS product media into the optical media drive on the HMC. Login to the

HMC and invoke the installios command. Once the managed system and partition

have been selected, the command will prompt for profile (default) and image source

(/dev/cdrom) selections. It will then prompt for network information. Once the choices

have been made, the command will activate the partition and install the Virtual I/O

Server product.

Installing the VIOS software from a NIM Server is similar to installing the VIOS from the

HMC because it is also a network install using a NIM master. You have the option of

using installios from a NIM master just like from the HMC. In this case, the creation

of NIM objects to represent the install resources and client (the VIOS) is done

automatically. In a traditional NIM environment, you can simply create these objects and

perform mksysb bos_inst operations.

UNIX Software Service Enablement

Installing from the HMC

 Place VIO Server product media in HMC DVD drive

 Log in to HMC using SSH connection

 Invoke installios and follow menu prompts

 Will prompt you to select:

 Managed system that contains the partition you want to install

 Partition you wish to install

 Partition profile name

 Source of product images (/dev/cdrom)

 Intended IP address, subnet mask and gateway

 Ethernet media speed and duplex setting

 If you want to configure the clients network after installation

 Uses NIMOL (NIM on Linux) to install AIX mksysb image, and

therefore requires that the VIO Server partition has an interface

configured on the same LAN as the HMC

 VIOS partition will reboot once installation is complete

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-16 VIOS © Copyright IBM Corp. 2012

Figure 1-14. Command Line Interface (CLI) (1 of 2) QV3441.0

Notes:

The VIOS CLI

Although the Virtual I/O Server is currently implemented as a customized AIX image,

the system administrator does not use regular AIX commands to configure and maintain

the partition. The administration of the Virtual I/O Server can be performed using a

Command Line Interface (CLI), that encapsulates the underlying native operating

system commands used to perform particular tasks. This abstracts the AIX command

interface for those who are unfamiliar with AIX administration. Many of the VIOS

management tasks can also be completed from the HMC GUI, or the cfgassist

command

The CLI is implemented using a secure Korn shell. This means the administrator is

limited to the commands defined in the CLI. Since a secure shell is used, there are

limitations on the types of input/output redirection that are allowed.

UNIX Software Service Enablement

Command Line Interface (CLI) (1 of 2)

 Virtual I/O Server can be configured and administered using CLI:

 Device management (physical, virtual, LVM)

 Network configuration

 Software installation and update

 Security

 User management

 Installation of OEM software

 Maintenance tasks

 Alternatively, you can use cfgassist, or the HMC to configure the VIOS

and managed virtual devices

 VIOS CLI uses a restricted Korn shell:

 Limited set of commands available

 Cannot change directory, or set SHELL, ENV or PATH

 Cannot redirect output with >, >>, >| or <>

 Use the help command to see list of available commands

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 1-15. Command Line Interface (CLI) (2 of 2) QV3441.0

Notes:

The VIOS CLI

The oem_setup_env command provides a root shell for use when installing and

configuring software for use with complex storage solutions, for example, for adding

device drivers for IBM System Storage products. Once in the root shell, you can use

AIX commands and SMIT. Procedures performed in the padmin environment are

supported, however if something is configured from the root environment, then the

configuration might not be tested for the VIOS or supported.

Virtual I/O Server commands can be run from the HMC GUI. The example graphic in

the visual above shows the Operations -> Perform VIOS Command task. If you run

this task, a popup window appears and you can type a VIOS command in the input box.

In addition, we’ll see later in this course that there is another way to manage virtual

devices from the HMC.

UNIX Software Service Enablement

Command Line Interface (CLI) (2 of 2)

Use oem_setup_env command to access AIX root shell

Use exit command to return to padmin shell

 To run a VIOS CLI command while in the root shell, precede the

command with the /usr/ios/cli/ioscli command

 Example: # /usr/ios/cli/ioscli ioslevel

 When writing VIOS scripts to be run by padmin, precede each

command with the ioscli command

 VIOS commands can be run from the HMC:

 Example: viosvrcmd m MSname p VIOS1 c "ioslevel"

Or, use the HMC GUI:

Operations ->

Perform VIOS

Command

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-18 VIOS © Copyright IBM Corp. 2012

Figure 1-16. Configuring the Virtual I/O Server (1 of 2) QV3441.0

Notes:

Since the CLI is used to configure the Virtual I/O Server, the root user login is not

typically used. Instead, you should login to the partition using the user padmin. This is

the login for the Prime Administrator of the VIOS partition. The first time the padmin

user logs in, you will be prompted to set the password. Once you have entered a new

password, you will have to type "a" to accept the software maintenance terms and

conditions and then you’ll be logged in using the restricted Korn shell. Even though you

accepted the software maintenance terms and conditions, you still must accept the

software usage licenses by running the license -accept command.

UNIX Software Service Enablement

Configuring the Virtual I/O Server (1 of 2)

 Login to the VIOS as the user padmin, not root

 Will be prompted to set the password

 The first time you log in, you will be prompted to accept software

maintenance terms

Indicate by selecting the appropriate response below whether you

accept or decline the software maintenance terms and conditions.

Accept (a) | Decline (d) | View Terms (v) > a

 Then, you must accept the license for the Virtual I/O Server product:

$ license -accept

Use cfgassist

command to configure

initial basic settings

 Similar to SMIT

 Set date/timezone

 Configure networking

ato accept

onfig Assist for VIOS

ove cursor to desired item and press Enter.

et Date and TimeZone

hange Passwords

et System Security

IOS TCP/IP Configuration

nstall and Update Software

. .

hared Storage Pools

Electronic Service Agent

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 1-17. Configuring the Virtual I/O Server (2 of 2) QV3441.0

Notes:

Creating VIOS users

As of VIOS 2.2 you can create users with role-based access controls (RBAC). For

example, if the system administrator creates the role UserManagement with

authorization to access user management commands and assigns this role to a user,

that user can manage users on the system but has no further access rights. Reference:

http://pic.dhe.ibm.com/infocenter/powersys/v3r1m5/index.jsp?topic=/p7hb1/iphb1_vios

_using_rbac.htm

Set up redundancy for VIOS boot disk

After installing the VIOS and before setting up clients is the best time to mirror the

rootvg for the VIOS because you must reboot the VIOS as part of the procedure. The

-defer option to the mirrorios command can be used to avoid an auto boot when the

command completes. The VIOS will need to be rebooted before creating virtual

devices.

UNIX Software Service Enablement

Configuring the Virtual I/O Server (2 of 2)

 Consider creating additional users with mkuser command

 Other user management commands include lsuser, rmuser, passwd

 The cfgassist utility can also be used to change passwords

 Different user types are supported:

 Prime administrator (can only be one of this type: padmin)

 Service Representative (use -sr flag with mkuser)

 Development Engineer (use -de flag with mkuser)

 If boot disk mirroring is to be used, configure before supporting clients

 Extend rootvg to a second disk: $ extendvg -f rootvg hisk1

 Mirror rootvg: $ mirrorios hdisk1

 VIOS will reboot on command completion, or use -defer to reboot later

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-20 VIOS © Copyright IBM Corp. 2012

Figure 1-18. Configure networking QV3441.0

Notes:

Configuring network access

If the partition was installed from the HMC or a NIM server, then the network interface

may still be configured (depending on the options you selected when running the

installios or NIM commands).

Use the cfgassist VIOS TCP/IP Configuration option or the mktcpip command to

configure the first Ethernet interface. Use lstcpip to check the configuration.

It is probably unlikely that a VIOS will need more than one interface. However, to

configure additional Ethernet interfaces after the first, you must use the chdev

command. The mktcpip command is used for the first interface because it sets the

hostname as well as the other TCP/IP parameters. Here is an example chdev

command for the Virtual I/O Server which configures the interface en1, has an IP

address of 10.0.0.5, and has a netmask of 255.0.0.0:

$ chdev -dev en1 -attr netaddr=10.0.0.5 netmask=255.0.0.0 state=up

UNIX Software Service Enablement

Configure networking

Use cfgassist or mktcpip command to configure first network

interface (if not configured during installation from HMC or NIM

server)

 Syntax for mktcpip:

mktcpip -hostname HostName -inetaddr Address \

-interface Interface [-start] \

[-netmask SubnetMask] [-cabletype CableType]\

[-gateway Gateway] \

[-nsrvaddr NameServerAddress \

-nsrvdomain Domain] [-plen prefixLength]

 Check network configuration with the lstcpip command

 Useful flags: -interfaces, -state, -stored, -adapters,

-namesrv

 The hostname command is also available

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 1-19. Resource Monitoring and Control (RMC) QV3441.0

Notes:

RMC resource managers

The RMC subsystem is used by clients to monitor the state of system resources and to

send commands to resource managers. The RMC subsystem acts as a broker between

the client processes that use it and the resource manager processes that control

resources. Resource managers may be active or may be inoperative until they are

needed. This is not a complete list, but a typical list you’ll see in AIX is as follows:

- CSMAgentRM - For CSM client resource agent (nodes) operations

- ServiceRM - For service event operations

- DRM - For DLPAR operations

- DMSRM - For tracking status of partitions

- ERRM - For taking action in response to a system event

- AuditRM - For audit log operations

- HostRM - For host monitoring operations related to load and OS status

- LPRM - The least privileged RM can run a root command or script, locally or

remotely, on behalf of an authorized user

UNIX Software Service Enablement

Resource Monitoring and Control (RMC)

 RMC daemons

 Exist on the HMC, VIOS, and client partitions

 Used for communications for operations such as:

 DLPAR, reporting service events, and HMC screens for shutting down

partitions and virtual device configuration

 The daemons are Resource Managers which each manage a class of resources

 List resource managers:

HMC: monhmc -s rmc

 VIOS and AIX: lssrc -g rsct_rm

# lssrc -g rsct_rm

Subsystem Group PID Status

IBM.CSMAgentRM rsct_rm 7012576 active

IBM.ServiceRM rsct_rm 5177570 active

IBM.DRM rsct_rm 5636306 active

IBM.DMSRM rsct_rm 5701836 active

IBM.ERRM rsct_rm inoperative

IBM.AuditRM rsct_rm inoperative

IBM.HostRM rsct_rm inoperative

IBM.LPRM rsct_rm inoperative

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-22 VIOS © Copyright IBM Corp. 2012

Figure 1-20. RMC management QV3441.0

Notes:

Restarting RMC daemons

If you find that the following features do not work, try stopping and restarting the

daemons:

- DLPAR operations

- HMC GUI commands to shutdown partitions or to view virtual SCSI configurations

Reporting serviceable events to the HMC also relies on RMC connections, but the lack

of this function may not be immediately noticed by the administrator.

When starting the daemons you must first run the command with -A, then run it again

with -p. That is, you cannot run these flags at the same time.

UNIX Software Service Enablement

RMC management

 Wait up to 5-7 minutes after partition or HMC has booted

 The RMC daemons on the HMC will establish communications with the

daemons on the VIOS and client LPARs

If the /var file system is full when RMC daemons start, the RMC

configuration may be corrupted

 To stop the daemons on an AIX LPAR:

# /usr/sbin/rsct/bin/rmcctrl z

 To start the daemons on an AIX LPAR and enable remote client

(peer) communications:

# /usr/sbin/rsct/bin/rmcctrl A

# /usr/sbin/rsct/bin/rmcctrl p

Use oem_setup_env for the VIOS to run these commands

 To restart daemons on HMC, reboot the HMC

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 1-21. RMC reinitialization QV3441.0

Notes:

RMC reinitialization

Sometimes restarting the daemons is not enough to fix an RMC problem, particularly on

a cloned system. Systems which are cloned from other systems may have the old

system’s authorization information. To clear this, reinitialize the RMC system with the

recfgct command.

UNIX Software Service Enablement

RMC reinitialization

 If necessary, reinitialize RMC files with recfgct

 Fix a broken RSCT configuration or reinitialize RSCT files after cloning an

LPAR's operating system

 Use if the restart of the RMC daemons did not fix the problem and there

are no network issues

 New files are copied from templates

 Command path:

# /usr/sbin/rsct/install/bin/recfgct

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-24 VIOS © Copyright IBM Corp. 2012

Figure 1-22. Checkpoint (1 of 2) QV3441.0

Notes:

UNIX Software Service Enablement

Checkpoint (1 of 2)

1. Which one of the following types of virtual devices can be used without a Virtual I/O

Server?

A. Virtual Ethernet adapter

B. Virtual SCSI disks

C. Shared Ethernet adapter

D. Virtual Fibre Channel adapter

2. Which one of the following is NOT a function of virtual I/O?

A. Reduction or elimination of need for additional physical I/O adapters

B. Supports Live Partition Mobility operations

C. Enables partitions to take advantage of unused processing cycles

D. Allows clients to attach to storage that may be unsupported for Linux or IBM i

3. Which one of the following is NOT a valid way to install the Virtual I/O Server product:

A. Directly from product media

B. Install AIX first, then install the Virtual I/O Server filesets and updates

C. Install from the HMC with the installios command

D. Install from an AIX NIM master

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 1-23. Checkpoint (2 of 2) QV3441.0

Notes:

UNIX Software Service Enablement

Checkpoint (2 of 2)

4. True or False: When installing the Virtual I/O Server, it installs just like an

AIX mksysb image.

5. To use virtual SCSI devices, which system feature must be purchased?

A. Virtual I/O Server software

B. PowerVM Express, Standard, or Enterprise Edition

C. Live Partition Mobility

D. None of the above

6. What is the command in the Virtual I/O Server command line interface that will

show a list of all available Virtual I/O Server command line commands?

7. True or False: The Virtual I/O Server can be booted in any type of partition.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-26 VIOS © Copyright IBM Corp. 2012

Figure 1-24. Exercise 1 - VIOS configuration QV3441.0

Notes:

UNIX Software Service Enablement

Exercise 1 - VIOS configuration

 Activity:

 Configure a new Virtual I/O Server installation and investigate the

VIOS command line interface (CLI)

 Configuration

Domain

Nameserver IP

Address

Default Gateway

Netmask

VIOS IP

Address

VIOS Hostname

Fill in the network information for the VIOS partition:

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 1-25. Unit summary QV3441.0

Notes:

UNIX Software Service Enablement

Unit summary

 Virtual I/O is a way to share hardware resources providing support for

more logical partitions

 The Virtual I/O Server software:

 Is installed in a Virtual I/O Server type partition

 Is needed for the virtual SCSI, Shared Ethernet adapter, and virtual

Fibre Channel devices

 The Virtual I/O Server can export physical volumes, logical volumes,

optical drives, tape drives, and files to client partitions

 The Virtual I/O Server has a command line interface with many

commands for managing the VIOS and its devices

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

1-28 VIOS © Copyright IBM Corp. 2012

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty Unit 2. Virtual SCSI Configuration

What this unit is about

This unit details the configuration of virtual SCSI devices both from the

Virtual I/O Server command line and from the HMC Storage

Management graphical user interface. Creating virtual targets with

different types of backing storage is covered in detail.

What you should be able to do

After completing this unit, you should be able to:

• Create virtual SCSI target devices using different types of backing

storage:

- Physical volumes, logical volumes, and optical devices

• Configure storage pool devices

• Configure a client partition to use virtual SCSI devices

• Determine the details of an existing virtual SCSI configuration

How you will check your progress

Accountability:

• Checkpoint questions

• Machine exercises

References

IBM Power Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp

Virtual I/O Server product documentation from the IBM Power

Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp?t

opic=/p7hb1/iphb1kickoff.htm

IBM Redbooks documents can be found at:

http://www.redbooks.ibm.com

SG24-7940 IBM PowerVM Virtualization Introduction and

Configuration, an IBM Redbooks document

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-2 VIOS © Copyright IBM Corp. 2012

SG24-7590 IBM PowerVM Virtualization Managing and Monitoring, an

IBM Redbooks document

SG24-7825 PowerVM Migration from Physical to Virtual Storage, an

IBM Redbooks document

REDP4194 IBM System p Advanced POWER Virtualization

(PowerVM) Best Practices, an IBM Redbooks

document

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-1. Unit objectives QV3441.0

Notes:

UNIX Software Service Enablement

Unit objectives

After completing this unit, you should be able to:

 Create virtual SCSI target devices using different types of backing

storage:

 Physical volumes, logical volumes, and optical devices

 Configure storage pool devices

 Configure a client partition to use virtual SCSI devices

 Determine the details of an existing virtual SCSI configuration

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-4 VIOS © Copyright IBM Corp. 2012

Figure 2-2. Virtual SCSI overview (1 of 3) QV3441.0

Notes:

Virtual devices

The Virtual I/O Server supports virtual SCSI disks which are backed by various types of

physical devices. Regardless of how the virtual SCSI disk is implemented using various

types of backing storage, all standard SCSI conventional rules apply to the device; it will

behave as a standard SCSI compliant device. Once a virtual disk is assigned to a client

partition, the Virtual I/O Server must be available before the client partitions are able to

access it.

Regardless of the type of backing storage (physical volume, logical volume, file) being

used for a virtual SCSI disk, the client partition recognizes it as a generic SCSI disk

device. As such it can be used in the same way as a physical disk.

If you are using Virtual I/O Server V1.2 (or later) and have AIX V5.3 with maintenance

level 3 in the client partition (or later), then you can also configure virtual optical drives.

UNIX Software Service Enablement

Virtual SCSI overview (1 of 3)

 Implemented as client/server relationship

 Virtual I/O Server owns the physical resource

 Client partition sees generic SCSI devices

 Current virtual device types supported:

 Disk backed by physical volume

 Disk backed by logical volume

 Disk backed by file (as of VIOS V1.5)

 Optical media device backed by physical optical media device

(as of VIOS V1.2)

 Optical media device backed by file (as of VIOS V1.5)

 Tape backed by tape device (as of VIOS V2.1)

 Inter-partition communication provided by Hypervisor

 Client partition can boot from virtual SCSI devices

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-3. Virtual SCSI overview (2 of 3) QV3441.0

Notes:

Overview

The visual above illustrates how the POWER Hypervisor provides the communications

channels that permit client virtual SCSI adapters to communicate with server virtual

SCSI adapters. Virtual target devices (VTDs) are used to allow backing storage that is

physically owned by the Virtual I/O Server to be presented as a target on a virtual SCSI

bus. A client partition using a virtual SCSI bus can detect and use all virtual targets on

that bus.

Virtual SCSI itself does not have any maximums in terms of number of supported

devices. The Virtual I/O Server supports a maximum of 1024 virtual I/O slots per Virtual

I/O Server. A maximum of 256 virtual I/O slots can be assigned to a single client

partition.

UNIX Software Service Enablement

Virtual SCSI overview (2 of 3)

S S S

PHYS

Virtual I/O Server

Client

PHYS

PHYS Physical

Adapter

SVirtual SCSI

Server Adapter

CVirtual SCSI

Client Adapter

VTD Virtual Target

Device

VTDVTD VTD

Hypervisor

Physical

Storage

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-6 VIOS © Copyright IBM Corp. 2012

Figure 2-4. Virtual SCSI overview (3 of 3) QV3441.0

Notes:

Details

The visual above illustrates how multiple virtual target devices, and their associated

backing storage devices, can be presented on a single virtual SCSI bus. This can

simplify configurations by reducing the number of virtual adapters. Notice in the visual

above that the three backing storage devices map to the three virtual devices in the

client LPAR. The devices associate with the virtual SCSI server adapter (vhost#) can be

changed dynamically. The client partition just needs to run the cfgmgr command to see

new devices.

UNIX Software Service Enablement

Virtual SCSI overview (3 of 3)

SVirtual SCSI

Server Adapter CVirtual SCSI

Client Adapter

VTD Virtual Target

Device

Virtual I/O Server

Client

Hypervisor

Physical Storage

(physical volumes,

logical volumes, optical

media drive, tape, files)

hdisk5

lv_3

cd0

vtscsi0

vtscsi1

vtopt0

VTD

hdisk0

hdisk1

cd0

One virtual adapter can host multiple virtual target devices

vscsi0

vhost0

VTD

(Physical adapters are implied)

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-5. Creating virtual SCSI devices QV3441.0

Notes:

Procedure

The visual shows the steps required to make virtual SCSI devices available to a client

partition. After the virtual SCSI server adapter is created, the backing storage is

associated with the adapter by creating one or more virtual target devices (backing

storage may have to be configured also). Finally, the client partition can be activated

with the new adapter configured in its profile, or a client adapter can be dynamically

configured from the HMC with a DLPAR operation. DLPAR operations are available if

there is network connectivity between the HMC and partitions.

After adding devices with DLPAR, use the cfgmgr command in AIX or the cfgdev

command in the Virtual I/O Server to have the operating system discover the new

devices. You do not need to run a command for this in Linux.

The next set of visuals show how to create virtual devices on the command line,

followed by visuals showing how to create virtual devices from the HMC. In addition,

devices may be created with the VIOS cfgassist utility.

UNIX Software Service Enablement

Creating virtual SCSI devices

1. Create virtual SCSI server adapter in Virtual I/O server

 Allow any partition to connect or map to specific client and adapter ID

2. Create virtual SCSI client adapter in client partition

 Specify connection to SCSI server partition and adapter ID

 Modify server adapter to specify client if not done at creation

3. Activate Virtual I/O server partition and check adapter is available

 Or, if added with DLPAR, discover new devices with the cfgdev

command

4. Create storage resources to export for virtual SCSI disk

5. Create virtual target device (VTD) by associating storage with virtual

SCSI server adapter

6. Activate the client partition*

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-8 VIOS © Copyright IBM Corp. 2012

Figure 2-6. Step 1: Virtual SCSI server adapter (1 of 2) QV3441.0

Notes:

Virtual SCSI server adapter

The first step in the procedure to create virtual SCSI devices available to a client

partition is to use the HMC to add a virtual SCSI server adapter to the Virtual I/O Server

partition. Using the HMC interface, edit a profile for the Virtual I/O Server partition or

add the adapter dynamically (or both for a permanent configuration). On the Virtual

Adapters tab, click on Actions, then choose Create -> SCSI Adapter.

Maximum virtual adapters

On POWER6 processor-based servers, the default maximum number of virtual

adapters supported by default for a VIO Server partition is 20, instead of 10 as it is for

POWER5 and POWER7 servers. The first 11 virtual adapter IDs (0-10) are reserved on

a Virtual I/O Server partition on POWER6. On POWER5 and POWER7 servers, only

the first two adapter IDs (0-1) are reserved. For AIX/Linux type partitions, the default

maximum is still 10 virtual adapters and the first 2 virtual slots are reserved.

UNIX Software Service Enablement

Step 1: Virtual SCSI server adapter (1 of 2)

 On Virtual Adapters tab, use Actions -> Create -> SCSI Adapter

 Default maximum number of virtual adapters on the VIOS:

 POWER7 servers have a default of 10 and slots 0-1 are reserved

 POWER6 servers have a default of 20 and slots 0-10 are reserved

 Cannot change maximum setting dynamically

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-7. Step 1: Virtual SCSI server adapter (2 of 2) QV3441.0

Notes:

Virtual SCSI server adapter properties

For adapters which support client partitions that may be relocated with Live Partition

Mobility, do not make the adapter required.

You can select the Any client partition can connect option when the client partition is

not yet created. If you do this, it is important to return to this panel again to change this

setting once you have created the client partition and the virtual SCSI client adapter.

Make a note of the adapter ID because it will be needed when you create the virtual

SCSI client adapter.

If the client partition exists, select Only selected client partition can connect and

provide the client partition name and its virtual adapter ID. The client adapter will not

exist yet, but choose the adapter ID that you plan to use on the client.

Once you have selected the correct options, click on the OK button to finish the dialog.

This will return you to the partition profile properties dialog.

UNIX Software Service Enablement

Step 1: Virtual SCSI server adapter (2 of 2)

 Enter the adapter ID, whether the adapter is required, and the client

adapter that will connect

By default, Any client partition can

connect is specified, but it is better

to specify a particular client partition

and ID to be sure only that client can

use the devices

Note (or change) the adapter ID and

indicate if the adapter is required

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-10 VIOS © Copyright IBM Corp. 2012

Figure 2-8. Step 2: Virtual SCSI client adapter (1 of 2) QV3441.0

Notes:

Virtual SCSI client adapter

The second step in the procedure to make virtual SCSI devices available to a client

partition is to use the HMC to add a virtual SCSI client adapter to the client partition.

Using the HMC interface, navigate to the Virtual Adapters tab on the desired profile for

the client, just as you did for the Virtual I/O Server. Or, use a DLPAR operation to add

the adapter. Use the Actions menu to create a new virtual SCSI adapter.

Maximum virtual adapters

Note that the default maximum number of virtual adapters on an AIX/Linux LPAR is 10.

This default maximum value is the same on all Power Systems. On all AIX/Linux and

VIOS partitions, the first two virtual slots are reserved for virtual serial adapters which

are used for the virtual console communications.

UNIX Software Service Enablement

Step 2: Virtual SCSI client adapter (1 of 2)

 Create client adapter after you create the server adapter

 Default maximum for AIX/Linux partitions is 10

 Adapters IDs 0 and 1 are reserved

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-9. Step 2: Virtual SCSI client adapter (2 of 2) QV3441.0

Notes:

Create virtual SCSI client adapter dialog

The visual above shows the Create Virtual SCSI Adapter Properties dialog for a client

partition. Just as for the server adapter, you must specify a virtual adapter ID. This

should match the adapter ID you specified for the client when you created the server

adapter (if it was specified at that time). If this is a required device, for example, a boot

device, then check the This adapter is required for partition activation checkbox.

Next, select the Virtual I/O Server partition name from the drop-down list available for

the Server partition field, and enter the adapter ID being used for the virtual SCSI

server adapter. The server ID listed on this dialog must match the one you used when

creating the virtual server SCSI adapter!

UNIX Software Service Enablement

 Configure the adapter in the Properties window

 Use the System VIOS Info button to view and choose a virtual SCSI

server adapter configured on the VIOS

Step 2: Virtual SCSI client adapter (2 of 2)

Specify the client

adapter ID

Is this adapter

required?

Select the VIOS

partition and enter the

server adapter ID

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-12 VIOS © Copyright IBM Corp. 2012

Figure 2-10. Step 3: Verify virtual SCSI server adapter QV3441.0

Notes:

Details

Use the lsdev command with the -virtual option, as shown in the visual above, to list

the virtual devices in the partition. Virtual SCSI server adapters have a device name

that starts with vhost.

You can examine a virtual SCSI server adapter’s configuration using the lsmap

command. This is used to identify which virtual slot the device is using, along with

information about the target devices associated with the adapter. In the example shown

in the visual above, the virtual SCSI server adapter has been created, but there are no

virtual target devices associated with it yet. In the example, the device vhost0 is virtual

adapter ID 11 (C11 in the location code) of the Virtual I/O Server partition, which has

partition ID 1 (V1 in the location code). The associated client LPAR ID is also shown in

this output (ID in hex is 2), and is ‘0’ when a client has not yet been activated.

UNIX Software Service Enablement

 Activate the Virtual I/O Server partition

 Or run cfgdev if added dynamically

 Check availability of virtual SCSI server adapters

$ lsdev -virtual

name status description

vasi0 Available Virtual Asynchronous Services Interface (VASI)

vhost0 Available Virtual SCSI Server Adapter

vsa0 Available LPAR Virtual Serial Adapter

 Check mapping to virtual slot

$ lsmap -vadapter vhost0

SVSA Physloc Client Partition ID

----------- ----------------------------- ------------------

vhost0 U8203.E4A.10CD1F1-V1-C11 0x00000002

VTD NO VIRTUAL TARGET DEVICE FOUND

Step 3: Verify virtual SCSI server adapter

Physical Location Code: V1 = partition ID 1, C11 = adapter ID 11

Client Partition ID: is 0 when the client partition has not yet been activated

No backing storage

mapped yet

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-11. Step 4a: Create storage resources with LVM QV3441.0

Notes:

Backing storage resources overview

If you want to configure virtual SCSI targets backed by a physical volume, the optical

media device, or a tape device, you can make the VTDs and map them to the vhost

adapter. There is no additional setup needed. If you want to configure virtual SCSI

targets that are backed by logical volumes or files, then first you need to create the

volume groups, logical volumes, and files which will be used. It is best not to create

these in the VIOS rootvg volume group.

The LVM concepts on the Virtual I/O Server partition are identical to those on AIX. The

difference is that on the Virtual I/O Server, you use VIOS CLI commands to create the

volume groups and logical volumes.

UNIX Software Service Enablement

Step 4a: Create storage resources with LVM

 Physical volumes

 Entire physical volume can be exported as a single SCSI target

 With SAN storage, LUN creation occurs on the storage system or controller

 No additional setup required to create backing storage for physical volumes or

physical tape and optical media devices

 Logical volumes

 Use the CLI to create volume groups and logical volumes if you wish to have

finer granularity

 To create a volume group:

$ mkvg [ -f ][ -vg VolumeGroup ] PhysicalVolume ...

 To create a logical volume:

$ mklv [ -lv NewLogVolName ] VolumeGroup Size

[ PhysicalVolume ... ]

 Example creating a logical volume named clientA_lv1 that is 5 GB in size in

an existing VolGroupA volume group:

$ mklv -lv clientA_lv1 VolGroupA 5G

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-14 VIOS © Copyright IBM Corp. 2012

Figure 2-12. Step 4b: Create storage resources with storage pools (1 of 2) QV3441.0

Notes:

Using I/O storage pools

Storage pools are available in Virtual I/O Server V1.2 and later. Similar to volume

groups, logical volume storage pools are collections of one or more physical volumes.

Storage pools are created and managed using the following commands:

- The mksp command creates a storage pool.

- The chsp command changes the characteristics of a storage pool.

- The lssp command displays information about a storage pool. The lssp with the

-bd option will list out the backing devices and show the associations between the

virtual host adapters and virtual devices.

UNIX Software Service Enablement

Step 4b: Create storage resources with storage

pools (1 of 2)

Storage pools are similar in

concept to volume groups

 Created using one or more physical volumes

 Used to manage logical volume and file-based backing storage

 To create a logical volume storage pool, specify the name and one or

more physical volume names:

$ mksp [ -f ] StoragePool PhysicalVolume ...

 Example creating the clientsp1 logical volume pool with two physical

volumes:

$ mksp -f clientsp1 hdisk3 hdisk4

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-13. Step 4b: Create storage resources with storage pools (2 of 2) QV3441.0

Notes:

Backing devices in storage pools

Backing devices are created and deleted using the following commands:

- The mkbdsp command assigns storage from a storage pool to a virtual SCSI adapter

as a backing device.

- The rmbdsp command removes a backing device from a virtual SCSI server adapter

by removing its associated virtual target device. By default the backing device is

also removed and its storage returned to the storage pool.

Example command that creates a backing device but does not map it to a vhost

adapter:

$ mkbdsp -sp clientsp1 6G -bd clientA_lv2

Example command that assigns an existing backing device to a vhost adapter:

$ mkbdsp -sp clientsp1 -bd clientA_lv2 -vadapter vhost0

UNIX Software Service Enablement

Step 4b: Create storage resources with storage

pools (2 of 2)

 Create logical volumes and backing devices in one step:

$ mkbdsp [-sp StoragePool] Size [-bd BackingDevice]

-vadapter ServerVirtualSCSIAdapter [ -tn VTDName ]

 Example creating a logical volume named clientA_lv1 that is 5 GB in size in

the existing clientsp1 storage pool on an existing vadapter:

$ mkbdsp -sp clientsp1 5G -bd clientA_lv1 -vadapter vhost0

Creating logical volume "clientA_lv1" in storage pool

"clientsp1".

vtscsi3 Available

clientA_lv1

Command output showing new VTD name

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-16 VIOS © Copyright IBM Corp. 2012

Figure 2-14. List storage resource configuration (1 of 2) QV3441.0

Notes:

List storage resources

The visual above shows common usages of the lsvg, lspv, and lssp commands. You

can use lsvg and lspv which are similar to AIX LVM commands. Notice that the lssp

command shows similar information. The name LVPOOL refers to a logical volume

storage pool which is the same as the LVM volume group. It doesn’t matter if you use

LVM-style commands or storage pool commands to create, list, or remove storage

resources.

The lspv -free command shows physical volumes that are not part of a volume group

or storage pool, and are not mapped to a vhost adapter.

UNIX Software Service Enablement

List storage resource configuration (1 of 2)

 Show physical volumes that are not already mapped to a virtual host

adapter or part of a shared memory pool:

$ lspv -free

NAME PVID SIZE(megabytes)

hdisk5 00f65f4e2800f7db 70006

 Show all storage pools:

$ lssp (or use lsvg)

Pool Size(mb) Free(mb) Alloc Size(mb) BDs Type

rootvg 69888 46336 128 1 LVPOOL

clientsp1 139776 134656 64 1 LVPOOL

 Show physical volumes in a storage pool:

$ lsvg -pv clientsp1 (or use lssp -detail -sp clientsp1)

clientsp1:

PV_NAME PV STATE TOTAL PPs FREE PPs FREE DISTRIBUTION

hdisk3 active 1092 1012 219..138..218..218..219

hdisk4 active 1092 1092 219..218..218..218..219

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-15. List storage resource configuration (2 of 2) QV3441.0

Notes:

List storage resources

In the examples above, clientsp1 is a volume group (LVPOOL). clientA_lv1 is a logical

volume in the clientsp1 volume group.

To remove storage resources, use the rmlv, reducevg, and rmbdsp commands.

Example storage pool command to remove a logical volume and its virtual target

device:

$ rmbdsp -sp clientsp1 -bd clientA_lv1

UNIX Software Service Enablement

List storage resource configuration (2 of 2)

 Show logical volumes in a volume group/storage pool:

$ lsvg -lv clientsp1

clientsp1:

LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT

clientA_lv1 jfs 80 80 1 open/syncd N/A

 Show logical volumes on a disk:

$ lspv -lv hdisk3

hdisk3:

LV NAME LPs PPs DISTRIBUTION MOUNT POINT

clientA_lv1 320 320 00..255..65..00..00 N/A

 Show all backing devices in a storage pool:

$ lssp -bd -sp clientsp1

Name Size(mb) VTD SVSA

clientA_lv1 5120 vtscsi3 vhost0

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-18 VIOS © Copyright IBM Corp. 2012

Figure 2-16. Step 5: Create the virtual target device QV3441.0

Notes:

Mapping the storage devices to the server adapter

The virtual SCSI target devices shown in the visual above are mapped to a physical

volume (hdisk1), a logical volume (datavglpar1) and an optical device (cd0). To map a

tape device, simply map the rmt# device like the optical media device example above.

The lsdev -virtual output for the tape device will look like this (using the default VTD

name):

vttape0 Available Virtual Target Device - Tape

The examples above show commands which create the new vtscsi0, vtscsi1, and

vtopt0 devices. If you don’t want to use the next numerical vtscsi# device name, then

you can specify a device name with the use of the -dev DeviceName option.

Important: If you used the mkbdsp command to create a logical volume in a storage pool

as shown in the previous visual, then you do not have to use the mkvdev command to

create the VTD and associate it with the vhost adapter. This was done all in one step.

UNIX Software Service Enablement

Step 5: Create the virtual target device

 The virtual target device (VTD) maps the storage resource to the

virtual SCSI adapter:

$ mkvdev vdev TargetDevice vadapter VirtualServerAdapter

[ dev DeviceName ]

 Examples:

$ mkvdev vdev hdisk1 vadapter vhost0

vtscsi0 Available

$ mkvdev vdev datavglpar1 vadapter vhost0

vtscsi1 Available

$ mkvdev vdev cd0 vadapter vhost0

vtopt0 Available

 Check the target devices with lsdev command:

$ lsdev virtual

name status description

vasi0 Available Virtual Asynchronous Services Interface (VASI)

vhost0 Available Virtual SCSI Server Adapter

vsa0 Available LPAR Virtual Serial Adapter

vtscsi0 Available Virtual Target Device  Disk

vtscsi1 Available Virtual Target Device  Logical Volume

vtopt0 Available Virtual Target Device  Optical Media

The mkvdev step is not

needed if mkbdsp created

and mapped the device

The mkvdev step is not

needed if mkbdsp created

and mapped the device

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-17. Verify the mapping of virtual targets QV3441.0

Notes:

lsmap details

The lsmap command can be used to show which virtual target devices are associated

with a virtual SCSI server device. You can specify the -all flag to see mappings for all

virtual SCSI server adapters, or choose a particular vhost adapter as shown in the

visual above.

The client partition ID is listed as zero if the client partition has not yet been activated.

Otherwise, the partition ID is given (in hex). If a previously running client shuts down, its

LPAR ID will still show in the lsmap output.

UNIX Software Service Enablement

Verify the mapping of virtual targets

$ lsmap -vadapter vhost0

SVSA Physloc Client Partition ID

--------------- -------------------------------------- ------------------

vhost0 U8202.E4B.06BCC8P-V1-C11 0x00000002

VTD vtopt0

Status Available

LUN 0x8200000000000000

Backing device cd0

Physloc U78AA.001.WZSG3GX-P2-D9

Mirrored N/A

VTD vtscsi0

Status Available

LUN 0x8100000000000000

Backing device hdisk1

Physloc U78AA.001.WZSG3GX-P2-D5

Mirrored false

VTD vtscsi1

Status Available

LUN 0x8300000000000000

Backing device datavglpar1

Physloc

Mirrored N/A

Use lsmap to verify

mapping of virtual

targets

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-20 VIOS © Copyright IBM Corp. 2012



Figure 2-18. Step 6: Virtual SCSI disks (client view) QV3441.0

Notes:

Details

Once the virtual target devices have been created on the Virtual I/O Server partition,

you can activate the client partition that will use the resources. The devices show up

under SMS, so a partition can boot from a virtual SCSI disk if it has an operating system

image installed on it. The partition can also boot from a virtual optical drive.

Virtual disks show up as physical volumes that can be configured just like normal

physical disks. The client partition cannot detect the type of backing storage device

being used for a virtual SCSI disk. That information can only be viewed on the Virtual

I/O Server partition, since it owns the physical resources.

If the client virtual adapter was added dynamically to the client LPAR, or if the Virtual I/O

Server added a new VTD to an existing vhost adapter used by the client, then the client

LPAR needs to run the cfgmgr command to discover the new devices.

UNIX Software Service Enablement

Step 6: Virtual SCSI disks (client view)

 Activate client partition or run cfgmgr:

 Virtual optical devices and bootable virtual disk devices will show up in SMS and

can be used for operating system install/boot

 Virtual disks (including file-backed virtual disks) appear as hdisks:

# lsdev -c disk

hdisk0 Available 05-08-00-3,0 16 Bit LVD SCSI Disk Drive

hdisk1 Available 05-08-00-4,0 16 Bit LVD SCSI Disk Drive

hdisk2 Available 05-08-00-5,0 16 Bit LVD SCSI Disk Drive

hdisk3 Available 05-08-00-8,0 16 Bit LVD SCSI Disk Drive

hdisk4 Available Virtual SCSI Disk Drive

hdisk5 Available Virtual SCSI Disk Drive

 Virtual optical devices (including file-backed virtual optical devices)

appear as cd#devices and virtual tape devices appear as rmt# devices:

# lsdev | grep cd

cd0 Available Virtual SCSI Optical Served by VIO Server

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-19. Manage Virtual devices with cfgassist . QV3441.0

Notes:

Details

The cfgassist utility on the VIOS can assist with creating backing devices and

assigning them to clients. The cfgassist utility is used in a similar way to the SMIT

menus in AIX.

UNIX Software Service Enablement

Manage Virtual devices with cfgassist

 The cfgassist utility can be used to manage volume groups, logical

volumes, storage pools, and virtual devices

 Use the Storage Management menu for setting up backing storage

 Use the Devices menu for creating virtual disks and mapping them to a

vhost adapter

Storage Management

Move cursor to desired item and press Enter.

olume Groups

Logical Volumes

Physical Volumes

Storage Pools

Virtual Media Repository

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-22 VIOS © Copyright IBM Corp. 2012

Figure 2-20. Check configuration from HMC (1 of 2) QV3441.0

Notes:

View virtual SCSI adapters and devices

The visual above shows how to select the Hardware Information -> Virtual I/O

Adapters -> SCSI task in the Systems Management application on the HMC. Be sure

to select the appropriate LPAR before running this task. For example, if you select a

client LPAR, you will only see information related to that client. If you select the Virtual

I/O Server partition, then you see all devices and clients for that VIOS.

UNIX Software Service Enablement

Check configuration from HMC (1 of 2)

 Look at the partition profile properties, or use the

Virtual I/O Adapters option on the HMC, under Hardware Information:

HMC command:

lshwres -m msname -r virtualio --rsubtype scsi

HMC command:

lshwres -m msname -r virtualio --rsubtype scsi

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-21. Check configuration from HMC (2 of 2) QV3441.0

Notes:

View virtual SCSI adapters and devices

The top example in the visual above is from a VIOS partition. The bottom example is

from a client partition. Networking must be configured between the LPARs and the HMC

for the HMC to be able to display information from that partition.

In the Virtual I/O Server example in the visual above, the vhost0 adapter shows three

backing storage devices, but only two remote backing storage devices. This is invalid

because there is always a one-to-one mapping between backing storage devices and

the devices seen in the client LPAR. The HMC GUI doesn’t give accurate information

for the remote backing devices which are optical media devices.

If there is backing device information but no client information then the client has either

not yet been configured, or is not running, or there is no network link between the

partition and the HMC.

UNIX Software Service Enablement

Check configuration from HMC (2 of 2)

 View virtual SCSI device configuration from the HMC:

 Virtual I/O Server example:

 Client example:

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-24 VIOS © Copyright IBM Corp. 2012

Figure 2-22. Discover backing storage example (1 of 2) QV3441.0

Notes:

Check the client configuration

This visual and the next show how to track the virtual disk on the client back to the

physical backing storage used on the Virtual I/O Server. First, start with what you know,

which is the device you want to investigate. In this example we will use hdisk0. The

lscfg command will tell you the device is a virtual device, the adapter ID number used

on the client (C2 in the example on the visual above), and the LUN of the device which

will be useful later.

Check the server configuration

Select the client partition in the HMC’s Systems Management LPAR table view, then run

the Hardware Information -> Virtual I/O Adapters -> SCSI task. This screen will show

the Virtual I/O Server’s partition name (useful if there is more than one) and the name of

the virtual SCSI server adapter (vhost) being used on the Virtual I/O Server. The

example above shows this to be vhost0.

UNIX Software Service Enablement

Discover backing storage example (1 of 2)

 A client partition sees hdisk0: What exactly is hdisk0?

 Check to see if it's a virtual device and note the LUN-ID on the adapter for

hdisk0:

 Check which Virtual I/O Server is being used and which vhost adapter it is using

on the Virtual I/O Server

 Select the client partition and run Hardware Information -> Virtual I/O

Adapters -> SCSI

# lscfg | grep hdisk0

hdisk0 U8202.E4B.06BCC8P-V2-C2-T1-L8100000000000000

Virtual SCSI Disk Drive

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-23. Discover backing storage example (2 of 2) QV3441.0

Notes:

Viewing the backing storage devices

The lsmap command output in the visual above shows that for vhost0, there are three

backing storage devices. Use the LUN ID from the client’s lscfg command output to

find the correct backing device. In this example, it is hdisk1.

UNIX Software Service Enablement

Discover backing storage example (2 of 2)

 View the device mapping on the Virtual I/O Server:

$ lsmap -vadapter vhost0

SVSA Physloc Client Partition ID

--------------- -------------------------------------------- ------------------

vhost0 U8202.E4B.06BCC8P-V1-C11 0x00000002

VTD vtopt0

Status Available

LUN 0x8300000000000000

Backing device cd0

Physloc U78AA.001.WZSGHVK-P2-D9

Mirrored N/A

VTD vtscsi0

Status Available

LUN 0x8100000000000000

Backing device hdisk1

Physloc U78AA.001.WZSGHVK-P2-D5

Mirrored false

VTD vtscsi2

Status Available

LUN 0x8200000000000000

Backing device LV1

Physloc

Mirrored N/A

hdisk1 is the backing storage

device for the clients virtual

hdisk0

hdisk1 is the backing storage

device for the clients virtual

hdisk0

This one has the

correct LUN ID

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-26 VIOS © Copyright IBM Corp. 2012

Figure 2-24. Removing virtual devices (1 of 2) QV3441.0

Notes:

Removing virtual devices

The visual above shows how to find and remove a virtual device in a client partition. In

the example above, the parent virtual client adapter is found and a recursive rmdev

command is used to remove the virtual adapter and the child devices.

UNIX Software Service Enablement

Removing virtual devices (1 of 2)

For client partitions, remove a virtual device like a physical device:

 Example where virtual SCSI adapter ID 4 controls the virtual device cd0:

 Deconfigure the device and its children:

# lsslot c slot | grep C4

U8203.E4A.10CD1F1-V2-C4 Virtual I/O Slot vscsi0

# rmdev l vscsi0 R

cd0 Defined

vscsi0 Defined

 Delete the devices from the ODM, rather than leaving them in the Defined

state:

# lsslot c slot | grep C4

U8203.E4A.10CD1F1-V2-C4 Virtual I/O Slot vscsi0

# rmdev l vscsi0 R d

cd0 deleted

vscsi0 deleted

Parent virtual

client adapter

Parent virtual

client adapter

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-25. Removing virtual devices (2 of 2) QV3441.0

Notes:

Removing devices on the Virtual I/O server

On the Virtual I/O Server, run the lsmap -vadapter vhost# (where vhost# is the

virtual host adapter name) to see the backing devices for that adapter and the virtual

target device names. You must remove all associations to backing devices first (all

virtual target adapters) before removing a virtual server adapter or you can add the

-recursive flag to remove all associated VTDs when you remove the adapter.

Notice in the examples above that either the rmdev or the rmvdev command can be

used to remove the VTD devices.

The default behavior for the rmbdsp command is to remove both the VTD and the

backing storage device. To retain the backing storage device, use the -savebd flag.

UNIX Software Service Enablement

Removing virtual devices (2 of 2)

 These VIOS commands are equivalent to remove only the VTD:

$ rmdev dev vtopt0

$ rmvdev vdev cd0

$ rmvdev vtd vtopt0

 Remove the VTD and the backing logical volume named LV1 (valid

only for logical volumes):

$ rmdev vdev LV1 rmlv

 Remove the vhost adapter and any associated VTDs:

$ rmvdev dev vhost1 recursive

vtopt0 deleted

vhost1 deleted

 These commands are equivalent and remove storage pool VTDs and

backing devices (unless savebd is used):

$ rmbdsp vtd VTDname [savebd]

$ rmbdsp sp spanme bd backing_device_name [savebd]

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-28 VIOS © Copyright IBM Corp. 2012

Figure 2-26. Managing devices from the HMC (1 of 2) QV3441.0

Notes:

Managing storage devices from the HMC

If you are running HMC V7 R3.4 service pack 2 or later, then you can also manage

storage devices from the HMC. Select the server name and run the tasks shown on the

visual above. The VIOS/SSP prompt means Virtual I/O Server or Shared Storage Pool

Storage Group (also known as a VIOS cluster).

Active Memory Sharing

Devices used for Active Memory Sharing paging devices will not be shown in queries.

UNIX Software Service Enablement

Managing devices from the HMC (1 of 2)

 Select the server on the HMC and use Configuration -> Virtual

Resources -> Virtual Storage Management to manage virtual

devices

 View and map backing storage devices to clients

 Relies on RMC for communication (over the network)

Discover

devices on

VIOS

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-27. Managing devices from the HMC (2 of 2) QV3441.0

Notes:

Create and assign backing storage devices from the HMC

Notice the tabs on the Virtual Storage Management screen. You can manage virtual

disks (logical volumes), storage pools (volume groups), physical volumes, optical

devices, and virtual Fibre Channel devices.

The Virtual Disks tab above shows that there is currently one logical volume created

and it is part of the rootvg volume group.

The Create virtual disk ... button allows you to create new logical volumes and assign

them to an LPAR in one step. It will do the mklv and mvdev commands (or the mkbdsp

command) in one step. The Modify assignment... button will allow you to change the

mapping of a device to a client.

The Actions menu has options to extend logical volumes and delete them.

UNIX Software Service Enablement

Managing devices from the HMC (2 of 2)

 Create virtual disks (logical volumes) and storage groups (volume

groups)

 Map devices

to clients

 Actions menu:

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-30 VIOS © Copyright IBM Corp. 2012

Figure 2-28. Create virtual disk example (1 of 2) QV3441.0

Notes:

Create a virtual disk

The window that is shown above appears when you click the Create virtual disk ...

button from the Virtual Disks tab. In the example, a logical volume named LV2 is being

created and assigned to LPAR2. The logical volume size will be 6 GB.

UNIX Software Service Enablement

Create virtual disk example (1 of 2)

 Create a logical volume on the VIOS and map it to a client partition

From Virtual Disks tab, click the Create virtual disk button

 This example configures the LV name, selects the volume group/storage

pool in which the LV will be created, sets the size, and assigns it to a

partition:

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-29. Create virtual disk example (2 of 2) QV3441.0

Notes:

Create virtual disk example

The visual above shows the results of creating the new logical volume shown on the

previous visual.

Deleting virtual disks

To delete virtual disks using the HMC GUI, select the virtual disk in the table then click

the Modify assignment ... button. Set the LPAR assignment to None and click OK.

Then, select the virtual disk again and choose Delete from the --- Select Action ---

menu in the header of the table.

UNIX Software Service Enablement

Create virtual disk example (2 of 2)

New LV

assigned to

LPAR2

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-32 VIOS © Copyright IBM Corp. 2012

Figure 2-30. Modify assignment QV3441.0

Notes:

Modify virtual disk assignments

The example in the visual above will assign the LV2 logical volume to an LPAR. If the

virtual disk already belongs to an LPAR, then you’ll see the warning as shown in the

visual above. This warning doesn’t appear if the virtual disk is unassigned.

Notice that if there are multiple virtual adapter connections between a particular LPAR

and the VIOS, then you can choose the appropriate one by adapter ID (slot number). In

the example above, LPAR1 has two virtual SCSI client adapters with adapter IDs 2 and

7. LPAR2 only has one virtual client SCSI adapter so no slot ID is shown.

UNIX Software Service Enablement

Modify assignment

 You can assign existing virtual disks (logical volumes) to LPARs

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-31. Creating storage pools / volume groups QV3441.0

Notes:

Manage storage pools

The visual above shows the Storage Pools tab and how to create a new storage pool.

The resulting window shows only free physical volumes such as you would see if you

ran the lspv -free command on the VIOS.

The example above shows a new storage pool named clientvg being created with one

physical volume (hdisk4).

UNIX Software Service Enablement

Creating storage pools / volume groups

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-34 VIOS © Copyright IBM Corp. 2012

Figure 2-32. Physical volumes (1 of 2) QV3441.0

Notes:

Physical volumes

The Physical Volumes tab allows you to map physical volumes to clients. To do this,

click on the Modify assignment... button. Notice that the Select Action pulldown menu

allows you to add or remove physical volumes to or from storage pools (volume

groups).

UNIX Software Service Enablement

Physical volumes (1 of 2)

Actions

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-33. Physical volumes (2 of 2) QV3441.0

Notes:

Modify physical volume assignments

The visual above shows what happens if you try to reassign a physical volume that is

already assigned to a running partition.

UNIX Software Service Enablement

Physical volumes (2 of 2)

 Assign, or reassign physical volumes to LPARs

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-36 VIOS © Copyright IBM Corp. 2012

Figure 2-34. Optical devices QV3441.0

Notes:

Assign the optical media drive

You can assign the optical media drive to partitions as well.

UNIX Software Service Enablement

Optical devices

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-35. Virtual SCSI storage and LVM QV3441.0

Notes:

Overview

Virtual storage can be manipulated on the client LPAR using the Logical Volume

Manager (LVM) just like an ordinary physical volume. Some performance

considerations from dedicated storage are still applicable when using virtual storage,

such as spreading hot logical volumes across multiple disks on multiple adapters so

that parallel access is possible. When the entire physical volume is served to the client,

the rules and procedures apply on the client side, as if the drive were native.

When a logical volume created on the server partition is served to the client partition,

the client partition thinks it has access to a single physical volume. The actual areas of

physical disk used to serve the virtual device to the client are decided on the server

partition when the logical volume is created, depending on the number of disks in the

volume group (on the server) and the inter-policy. If the logical volume on the server is

spread over multiple physical volumes, then reads and writes from the client may be

spread over those physical volumes.

UNIX Software Service Enablement

Virtual SCSI storage and LVM

 LVM strategies such as mirroring and multi-path I/O (MPIO) can be

set up on the client using virtual disks just like with physical, dedicated

disks:

 Caution: The virtual disk on the client may already be a logical volume on the

server

 Performance strategies used for physical, dedicated disk storage still

apply to virtual storage, such as spreading out hot logical volumes:

 Remember that virtual target disk devices can be backed by physical volumes,

logical volumes, or files so be sure you know what the physical backing storage

devices are and where they are located

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-38 VIOS © Copyright IBM Corp. 2012

Figure 2-36. Dynamically adding virtual SCSI storage (1 of 2) QV3441.0

Notes:

Maximum transfer

The maximum transfer attribute determines the size, in bytes, of the largest single data

transfer request that can be handled by a device. As devices are added to a vhost

adapter, the maximum transfer rate of the device is compared to the maximum transfer

rate for the adapter and the system uses the smaller of those values. For example, if

two disks are added to a vhost adapter, and one can handle 40000 bytes and the other

can handle 100000 bytes, then the maximum transfer rate is negotiated on the vhost

adapter to be 40000. If you add a device to a vhost adapter after the rate has been

negotiated and the new device has a lower max_transfer value than the vhost adapter,

then it cannot be added dynamically in the client. The client will have to reboot. If you

add a device to an existing vhost adapter and the new device has a higher

max_transfer value than the adapter, then it can be added dynamically in the client

however its max_transfer rate is lowered to the setting on the adapter.

Reference: http://www-304.ibm.com/support/docview.wss?uid=isg3T1010944

UNIX Software Service Enablement

Dynamically adding virtual SCSI storage (1 of 2)

 Two strategies when adding backing storage dynamically

 Add the backing device to an existing vhost adapter

 Add a new vhost adapter and add backing device to new adapter

 If the existing backing devices on the vhost adapter have a higher

maximum transfer rate, you have three choices:

 Add a new vhost adapter to add the backing device dynamically

 Reconfigure all of the devices onto a new vhost adapter

 Add the device to the existing adapter and reboot the client LPAR

 Message when max_transfer is too high for the new device:

$ mkvdev -vdev hdisk9 -vadapter vhost0

vtscsi4 Available

To see the newly added disk, client partition needs reboot.

 View max_transfer value:

$ lsdev -dev hdisk9 attr max_transfer

max_transfer 0x40000 Maximum TRANSFER Size True

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-37. Dynamically adding virtual SCSI storage (2 of 2) QV3441.0

Notes:

Discover physical volume

To discover on which physical volume a logical volume resides, run:

$ lslv -pv LV_name

LV_name:N/A

PV COPIES IN BAND DISTRIBUTION

hdisk0 001:000:000 100% 000:001:000:000:000

UNIX Software Service Enablement

Dynamically adding virtual SCSI storage (2 of 2)

 Best to group devices by max_transfer rate for optimal

performance because reducing the rate for a particular device may

slow it down

 If adding a logical volume, check the max_transfer rate of the

underlying physical volume

Example:

Existing vhost0 has three existing devices: cd0, hdisk1, LV1

cd0 max_transfer = 0x80000

hdisk1 (SAS disk) max_transfer = 0x100000

 Logical volume LV1 on hdisk4 (SAS disk) max_transfer = 0x100000

hdisk9 (SAN DS3500 disk) is then added with max_transfer = 0x40000

hdisk9 can be added successfully to vhost0, however it causes the message

and will not be added dynamically in the client

 Workaround that doesnt disturb existing devices: Dynamically add another vhost

adapter and add hdisk9 to the new vhost adapter

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-40 VIOS © Copyright IBM Corp. 2012

Figure 2-38. VIOS storage groups (1 of 2) QV3441.0

Notes:

VIOS storage groups

The first release of VIOS storage groups came out in December 2010 and was Virtual

I/O Server Version 2.2.0.11, Fix Pack 24, Service Pack 1. This release had limited

functionality as it only allowed one VIOS partition in the storage group. As of December

2011, Version 2.2.1.3 with iFix IV11852m13 allows up to 4 nodes in a storage group.

VIOS nodes in a storage group can continue to provide traditional virtual services such

as virtual SCSI, virtual Fibre Channel, and shared Ethernet adapter services separate

from the cluster.

See the IBM UNIX Software Service Enablement (USSE) course titled IBM PowerVM

Virtual I/O Server III: Configuring Shared Storage Pools (course code AHQV347) for

more information on this topic.

UNIX Software Service Enablement

VIOS storage groups (1 of 2)

 Up to four VIOS partitions can be in a storage group (also known as

a VIOS cluster)

 Each node in the cluster has access to a shared storage pool

 VIOS provides the same standard VSCSI target interface to clients

 Nodes in the cluster communicate via the network

Shared storage pool

 Set of physical volumes accessible by all VIOS nodes in the cluster

 Assign a large amount of SAN storage to the shared storage pool and

manage virtual disks from any of the VIOS nodes

 Backing storage devices are called logical units and are standard file-

backed VSCSI

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-39. VIOS storage groups (2 of 2) QV3441.0

Notes:

VIOS shared storage pool

A VIOS storage group, also known as a VIOS cluster, consists of one to four VIOS

partitions which share a single storage pool. Storage is provisioned from the shared

storage pool for clients and utilizes the virtual SCSI protocol. The VIOS storage group

can be managed from the VIOS CLI or from the cfgassist utility. The HMC can be also

used to create backing storage and map them to clients.

UNIX Software Service Enablement

VIOS storage groups (2 of 2)

VIOS nodes in a cluster see all LUNs and all logical units.

Storage could be IBM, EMC, Hitachi, SVC, and other SAN.

VIOS VIOS

Hypervisor

Client1

Client2

Client3

VIOS VIOS

Hypervisor

Client1

Client2

Client3

Shared Storage Pool

Up to 4 VIOS nodes are currently supported in a VIOS cluster

VIOS cluster nodes have global view into shared storage pool

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-42 VIOS © Copyright IBM Corp. 2012

Figure 2-40. Checkpoint (1 of 2) QV3441.0

Notes:

UNIX Software Service Enablement

Checkpoint (1 of 2)

1. Which one of the following statements about virtual SCSI devices is false?

A. The Virtual I/O Server partition owns the physical resources when virtual SCSI

devices are used.

B. To a client, a virtual SCSI device backed by a logical volume or an optical

device looks like an hdisk.

C. One virtual SCSI server adapter can have many associated backing devices.

D. The client partition cannot detect the type of backing storage device being used

for a virtual SCSI disk.

2. True or False: A partition may boot from virtual SCSI disk.

3. Virtual SCSI devices can be backed with which type of device(s)?

A. Logical volumes

B. Physical volumes

C. Optical media devices

D. All of the above

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-41. Checkpoint (2 of 2) QV3441.0

Notes:

UNIX Software Service Enablement

Checkpoint (2 of 2)

4. True or False: On the Virtual I/O Server, the mkvdev command can be used to

create virtual target devices that map physical devices to virtual SCSI server

adapters.

5. Which VIO Server command will show the mapping between Virtual SCSI server

adapters and backing storage devices?

A. lsdev

B. lsmap

C. lscfg

D. lssp

6. True or False: To remove the mapping between a physical device and virtual

SCSI server adapter, simply remove the virtual target device.

7. True or False: Virtual SCSI client LPARs can use mirroring on virtual SCSI disks.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-44 VIOS © Copyright IBM Corp. 2012

Figure 2-42. Exercise 2 - Virtual SCSI configuration QV3441.0

Notes:

Exercise 2

In the first two parts of this exercise, you will configure the boot disk for LPAR1 and

LPAR2. You’ll perform the procedure first on LPAR1 using commands on the VIOS CLI.

Then, you’ll perform this procedure for LPAR2 using the HMC GUI Virtual Storage

Management task.

In Part 3, you’ll create logical volumes and assign them to LPAR1, again using both the

VIOS CLI and the HMC GUI task.

In Part 4, you’ll check the configuration from the HMC CLI and from the HMC GUI.

If you have extra time you can perform the optional steps in Part 5. This will have you

assign the cd0 device on the VIOS to LPAR1 as a virtual cd device.

UNIX Software Service Enablement

Exercise 2 - Virtual SCSI configuration

 Activity:

 Configure a virtual SCSI disk using the VIOS CLI

 Configure a virtual SCSI disk using the HMC GUI

 Dynamically configure logical volumes as virtual disks

 Check the virtual device configuration

 Optional Lab

 Configure a virtual optical media device

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 2-43. Unit summary QV3441.0

Notes:

UNIX Software Service Enablement

Unit summary

 Virtual SCSI target devices can have different types of backing

storage

 Physical volume, logical volume, optical, file-backed, and tape devices

 Virtual SCSI devices are implemented with a client-server model

Virtual I/O Server

 Owns physical slots which contain physical adapters

 Contains virtual SCSI server adapter and backing storage devices

Client partition

 Contains virtual SCSI client adapter

 HMC is used to configure virtual adapters

Virtual adapter IDs are used to refer to appropriate virtual adapters

 As of HMC V7R3.4.0 SP2, you can use the HMC to manage backing

storage for virtual devices

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

2-46 VIOS © Copyright IBM Corp. 2012

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty Unit 3. File-backed Storage Devices

What this unit is about

The unit continues the discussion of virtual storage devices. Students

will learn how to create file-backed virtual disks and file-backed optical

media devices. The Virtual Media Repository is described as a way to

store media files and load them into client virtual optical media drives

where they can be used like physical media drives.

What you should be able to do

After completing this unit, you should be able to:

• Create file-backed storage pools

• Create file-backed storage devices for use as client virtual disks

• Create the Virtual Media Repository

• Create virtual optical media devices

• Load media into virtual optical media devices

• Install software using a virtual optical media device

How you will check your progress

Accountability:

• Checkpoint questions

• Machine exercises

References

IBM Power Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp

Virtual I/O Server product documentation from the IBM Power

Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp?t

opic=/p7hb1/iphb1kickoff.htm

IBM Redbooks documents can be found at:

http://www.redbooks.ibm.com

SG24-7940 IBM PowerVM Virtualization Introduction and

Configuration, an IBM Redbooks document

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

3-2 VIOS © Copyright IBM Corp. 2012

SG24-7590 IBM PowerVM Virtualization Managing and Monitoring, an

IBM Redbooks document

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 3-1. Unit objectives QV3441.0

Notes:

UNIX Software Service Enablement

Unit objectives

After completing this unit, you should be able to:

 Create file-backed storage pools

 Create file-backed storage devices for use as client virtual disks

 Create the Virtual Media Repository

 Create virtual optical media devices

 Load media into virtual optical media devices

 Install software using a virtual optical media device

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

3-4 VIOS © Copyright IBM Corp. 2012

Figure 3-2. File-backed virtual devices (1 of 2) QV3441.0

Notes:

File-backed virtual device types

There are two types of file-backed virtual devices. One is a disk device and the other is

an optical media device. In the case of the disk device, the file is created of a certain

size and it is viewed by the client as an hdisk. In the case of the optical media device, a

client can use the file-backed device just like a DVD-ROM or DVD-RAM drive to write to

and read from.

There are commands to load media images into the virtual optical media device (on the

Virtual I/O Server CLI). A Virtual Media Repository is created to store images which can

be loaded into a virtual optical media device for a client partition to access. If the media

is read-only then multiple clients can access the media simultaneously.

The ability to create file-backed devices is available in VIOS V1.5 and above.

There can only be one Virtual Media Repository per Virtual I/O Server partition.

UNIX Software Service Enablement

File-backed virtual devices (1 of 2)

 File-backed (FB) virtual device types:

 File-backed disk devices

 Files can be created and mapped to a vhost adapter which clients

see as hdisk# device

 File-backed optical media devices:

 Create a FB virtual optical media device that is mapped to a vhost

adapter which clients see as cd# device

 Create a Virtual Media Repository to store virtual DVD-ROM/RAM

media files

 Repository media files can be loaded into virtual optical media

device

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 3-3. File-backed virtual devices (2 of 2) QV3441.0

Notes:

FB virtual devices

A client LPAR can boot from media in a virtual optical device, or from a file-backed disk

device if it has an operating system installed.

File-backed disks reside in a file-backed storage pool which is a logical volume with a

file system in which the file-backed backing devices are created.

UNIX Software Service Enablement

File-backed virtual devices (2 of 2)

 FB virtual device characteristics:

 Bootable FB devices appear in SMS

 FB devices reside in FB storage pools

 FB storage pool implemented as a file system residing in a LV storage

pool

 Granularity as small as 1 MB or as large as parent file system

 Read-only virtual media can be shared by multiple clients

 FB device locations:

Disk = /var/vio/storagepools/<FBSP_Name>/<FB Disk>

 Virtual Media Repository = /var/vio/VMLibrary

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

3-6 VIOS © Copyright IBM Corp. 2012

Figure 3-4. Creating file-backed virtual disks QV3441.0

Notes:

Creating FB virtual disks

The graphic in the visual above shows the relationship between the parent logical

volume storage group (also known as a volume group), the FB disk storage pool, and

the FB disk virtual target devices. When creating FB disk storage, you must create a FB

disk storage pool to contain the files which will be used for the FB virtual disks.

Step 3 in the visual above has two steps: create the backing device and map to the

vhost adapter. This can be accomplished in one mkbdsp command, or with two. That is,

you can perform these steps separately.

The graphic in the visual above shows that a volume group is the same as a logical

volume (LV) storage pool and it contains one or more physical volumes. An FB pool is

created inside an LV pool. The FB backing devices are contained in the FB pool.

UNIX Software Service Enablement

Creating file-backed virtual disks

 Procedure for creating file-backed virtual disks:

1. Create a volume group (mkvg) or logical volume storage pool

(mksp -f)

2. Create a FB storage pool (mksp -fb) inside LV storage pool

3. Create a backing device (file) in the pool and map to a vhost adapter

(mkbdsp)

4. The client associated with that vhost adapter sees new FB device as

an hdisk

Volume Group/LV Storage Pool - contains hdisk(s)

FB Storage Pool (contains FB virtual disks)

Backing dev Backing dev Backing dev

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 3-5. Create FB virtual disks example QV3441.0

Notes:

Create FB virtual disk example

The visual above shows an example command sequence to create a FB virtual disk.

To create the volume group or the logical volume storage pool use either the mkvg or

mksp command but not both. The result with either command is the same.

The second example command, which creates a new FB storage pool, uses the mksp

command with the -fb flag. If you use the -f flag, you will get a logical volume storage

pool and not a file backed storage pool. The FB storage pool in the example above will

be 10 GB in size. You can use g or G for gigabytes and m or M for megabytes in this

command.

The third example command creates a backing storage file (fb_disk1) and maps it to a

virtual host adapter (vhost3), creating the VTD (vtscsi3) to track this mapping. In this

example, the file is specified to be 30 MB in size.

UNIX Software Service Enablement

Create FB virtual disks example

 Create new volume group/logical volume storage pool

$ mkvg -vg newvg hdisk1 OR mksp -f newvg hdisk1

 Create new FB storage pool in the logical volume storage pool

$ mksp -fb fbpool -sp newvg -size 10g

fbpool

File system created successfully.

10444276 kilobytes total disk space.

New File System size is 20971520

 Create new file device with a certain size, create the VTD, and map to

vhost adapter

$ mkbdsp -sp fbpool 30m -bd fb_disk1 -vadapter vhost3

Creating file "fb_disk1" in storage pool "fbpool".

Assigning file "fb_disk1" as a backing device.

vtscsi3 Available

fb_disk1

New storage pool (newvg)

New FB storage pool

(fbpool)that is 10 GB

located in newvg

Create new 30 MB file called fb_disk1

Resulting VTD is named vtscsi3

and is mapped to vhost3

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

3-8 VIOS © Copyright IBM Corp. 2012

Figure 3-6. View mapping of file-backed virtual disk QV3441.0

Notes:

View the configuration

The example lsmap command in the visual above shows the new backing storage

device that was created with the commands on the previous visual. Notice the backing

storage device location. The file-backed backing device is called fb_disk1 and the FB

pool is named fbpool.

UNIX Software Service Enablement

View mapping of file-backed virtual disk

 View mapping of backing device

$ lsmap -vadapter vhost3

SVSA Physloc Client Partition

------------- ---------------------------------- ----------------

vhost3 U8202.E4B.105F4EP-V1-C3 0x00000000

VTD vtscsi3

Status Available

LUN 0x8100000000000000

Backing device /var/vio/storagepools/fbpool/fb_disk1

Physloc

Mirrored N/A

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 3-7. View storage pools and backing devices QV3441.0

Notes:

View storage pool information

The list storage pool (lssp) command will list all LVPOOLs and FBPOOLs. Volume

groups are LVPOOLS so they are listed too. Use the backing device (-bd) flag with the

FB storage pool name (-sp fbpool) to list all FB backing devices in the FBPOOL

named fbpool. The lssp -detail -sp LVPOOLname command will list the physical

volumes in the LVPOOL or volume group named LVPOOLname.

UNIX Software Service Enablement

View storage pools and backing devices

 View all pools:

$ lssp

Pool Size(mb) Free(mb) Alloc Size(mb) BDs Type

rootvg 139776 78080 256 1 LVPOOL

newvg 139904 129664 128 0 LVPOOL

fbpool 10199 10168 128 2 FBPOOL

 View backing devices in an FB pool:

$ lssp -bd -sp fbpool

Name Size(mb) VTD SVSA

fb_disk1 30 vtscsi5 vhost1

 View physical volumes in an LV pool:

$ lssp -detail -sp newvg

Name PVID Size(mb)

hdisk3 00f6bcc8abfac919 139904

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

3-10 VIOS © Copyright IBM Corp. 2012

Figure 3-8. FB virtual disks remove commands QV3441.0

Notes:

Remove and rename commands

The rmbdsp command disassociates a backing device from its virtual SCSI adapter by

removing the virtual target device (VTD), and also removes the backing device itself by

default. Add the -savebd flag to only remove the VTD and not the backing device itself.

The example rmbdsp command above will remove the VTD (for example vtscsi3) and

the backing device. An equivalent command to the rmbdsp example above is:

$ rmbdsp -vtd vtscs3

The rmsp command removes an FBPOOL.

The chsp command changes the characteristics of a storage pool such as its size. Or,

as in the example in the visual above, you can perform the equivalent of the AIX

reducevg command and remove the LVPOOL by removing its last physical volume.

The chbdsp command changes the characteristics of a backing device within a storage

pool (such as its name).

UNIX Software Service Enablement

FB virtual disks remove commands

 Remove a file backed disk named fb_disk1 which is in the file-

backed storage pool named fbpool

$ rmbdsp -sp fbpool -bd fb_disk1

 Remove a file storage pool named fbpool that does not contain any

file backed disks mapped to a vhost adapter

$ rmsp fbpool

 Remove a logical volume storage pool (volume group) named newvg

which has only one physical volume named hdisk1

$ chsp -rm -sp newvg hdisk1

 Rename the backing device cl_data which is in the LV storage pool

datasp to bank_data

$ chbdsp -sp datasp -bd cl_data -mv bank_data

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 3-9. File-backed virtual optical devices QV3441.0

Notes:

Procedure

The visual above shows the overall procedure for configuring and using file-backed

optical media device with media from the Virtual Media Repository. The repository is like

a jukebox which stores many different optical media images. You can copy in as many

media files as the volume group has space to store them.

The media images are copied into the repository from real media from the /dev/cd

device on the VIOS partition or from ISO formatted files. Media images can also be

empty files. For each client, create a virtual optical media device and map it to the

virtual SCSI server adapter (vhost#) for that client. Then, each time the client wishes to

use the optical device, a media file is loaded into the device. When the client has

finished the operation (such as loading software), the media can be unloaded.

Media which is copied in as read-only can be used by multiple clients.

UNIX Software Service Enablement

File-backed virtual optical devices

 Procedure for creating and using file-backed virtual optical media

devices:

1. For each VIOS, create one

Virtual Media Repository

2. For each different media image, copy

an image into the repository

3. For each client, create a virtual optical

media device and map it to the appropriate

vhost adapter

4. For each optical media operation, load

a media file from the repository into the

client's virtual optical media device

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

3-12 VIOS © Copyright IBM Corp. 2012

Figure 3-10. Create FB virtual optical device (1 of 2) QV3441.0

Notes:

Changing the size of the Virtual Media Repository

You can increase the size of the repository with the chrep -size # command. For

example: chrep -size 5g

To add additional disks to a storage pool (volume group), you can use the command

chsp -add -sp storage_pool PhysicalVolume(s), where storage_pool is the name

of the volume group/storage pool to extend.

Create FB virtual optical device example

The example in the visual above creates the AIXinstall media "disk" with data on it from

the specified *.iso file. You can also create a blank 4 GB "disk" named LPAR1_file with

this command: mkvopt -name LPAR1_file -size 4g

Or, you can create a disk with data copied directly from a real cd0 device. Example

command: mkvopt -name vopt1 -dev cd0 [-ro]

UNIX Software Service Enablement

Create FB virtual optical device (1 of 2)

 Create volume group/logical volume storage pool

$ mkvg -vg medrep hdisk4 OR mksp -f medrep hdisk1

 Create the Virtual Media Repository in a LV pool

$ mkrep -sp medrep -size 10G

Virtual Media Repository Created

Repository created within "VMLibrary_LV" logical volume

 Copy media (AIXinstall) into repository

 Media could be blank, loaded from cd# device, or a blank file

$ mkvopt -name AIXinstall -file dvd.product.iso ro

 Remove commands

 Remove repository: $ rmrep

 Remove media in the repository: $ rmvopt name media_name

This example creates a

10 GB repository

Copy this file

(from current directory)

This is the new name for

the media in the repository

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 3-11. Create FB virtual optical device (2 of 2) QV3441.0

Notes:

Create the VTD and map it to the vhost adapter

The mkvdev command shown in the visual above creates a VTD and associates it with

the vhost4 adapter. Now there is a device that a client can use. If the client is already

running and has its virtual client SCSI adapter already configured to point to the vhost4

adapter on the Virtual I/O Server, then it just needs to run the cfgmgr command to see

the new cd# device.

There’s one more step before the client can use a media disk. You have to virtually load

the disk into the VTD so the client can see the media contents. The command to load

this disk is loadopt. You would use the unloadopt command to unload the virtual

media disk. (Or, alternatively, you can use loadopt with -f (force) which will unload

whatever is currently loaded before loading the new media.) The unloadopt command

syntax is: unloadopt -vtd vtopt0 where vtopt0 is the VTD name.

UNIX Software Service Enablement

Create FB virtual optical device (2 of 2)

 View repository and its contents

$ lsrep

Size(mb) Free(mb) Parent Pool Parent Size Parent Free

10198 8429 medrep 69888 59648

Name File Size Optical Access

AIXinstall 1769 None ro

 Create FB virtual optical device and map to vhost adapter

$ mkvdev -fbo -vadapter vhost4

vtopt1 Available

 Load the image into the media device

$ loadopt -vtd vtopt1 -disk AIXinstall

 Use the unloadopt command to unload or f option to loadopt

$ unloadopt -vtd vtopt1

VTD name

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

3-14 VIOS © Copyright IBM Corp. 2012

Figure 3-12. View repository information (1 of 2) QV3441.0

Notes:

Introduction

The visual above shows commands that can be used from the Virtual I/O Server CLI to

view information about the repository. The lsrep command is the primary command to

view information about the repository. It will also show whether any media in the

repository is loaded. The example in the visual above shows that the AIXinstall media

is loaded in the vtopt1 and in the vtopt3 devices. This media is marked as read only

and can be loaded into multiple vtopt# devices. Notice that the db_new disk is not

loaded.

The lssp command shows all storage pools: the logical volume storage pools

(LVPOOL) and the FB storage pools (FBPOOL). The lsvg command output only shows

the volume groups / parent logical volume storage pools.

UNIX Software Service Enablement

View repository information (1 of 2)

 List the repository and media disks information:

$ lsrep

Size(mb) Free(mb) Parent Pool Parent Size Parent Free

10198 6099 medrep 139904 119424

Name File Size Optical Access

AIXinstall 2048 vtopt1 ro

AIXinstall 2048 vtopt3 ro

aixopt1 2048 vtopt2 rw

db_new 3 None rw

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 3-13. View repository information (2 of 2) QV3441.0

Notes:

The mount command

The visual above shows an example mount command partial output. The last two lines

show the location for an FBPOOL and the media repository.

The lsvopt command

While lsrep lists all images and whether they are loaded, the lsvopt command lists the

virtual optical media drives and shows whether they have loaded media. If the media

column states No Media then an image is not loaded.

UNIX Software Service Enablement

View repository information (2 of 2)

 The mount command shows FB storage pools and the media

repository:

$ mount

node mounted mounted over vfs date options

-------- --------------- --------------- ------ ------------ ------------



/dev/fbpool /var/vio/storagepools/fbpool jfs2 Apr 08 09:45 rw,log=INLINE

/dev/VMLibrary /var/vio/VMLibrary jfs2 Apr 08 09:57 rw,log=INLINE

 View load status of virtual optical devices:

$ lsvopt

VTD Media Size(mb)

vtopt1 AIXinstall 2048

vtopt2 aixopt1 2048

vtopt3 No Media n/a

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

3-16 VIOS © Copyright IBM Corp. 2012

Figure 3-14. View virtual optical drive mapping QV3441.0

Notes:

Virtual optical drive mapping

The lsmap command will indicate whether a virtual optical drive has loaded media. If

media is loaded, the output will also list the media name as the backing device.

Loading and unloading of media must take place on the VIOS CLI or in the cfgassist

utility. It cannot be done from the client. From the HMC, you can use the viosvrcmd CLI

command or the GUI Operations -> Perform VIOS Command task to run the VIOS

loadopt or unloadopt command.

UNIX Software Service Enablement

View virtual optical drive mapping

 The lsmap command will show loaded media:

lsmap -vadapter vhost2

SVSA Physloc Client Partition ID

--------------- ------------------------------- ------------------

vhost2 U8202.E4B.06BCC8P-V1-C17 0x00000002

VTD vtopt1

Status Available

LUN 0x8100000000000000

Backing device

Physloc

Mirrored N/A

$ loadopt -vtd vtopt1 -disk aixopt1

$ lsmap -vadapter vhost2

SVSA Physloc Client Partition ID

--------------- ------------------------------- ------------------

vhost2 U8202.E4B.06BCC8P-V1-C17 0x00000002

VTD vtopt1

Status Available

LUN 0x8100000000000000

Backing device /var/vio/VMLibrary/aixopt1

Physloc

Mirrored N/A

vtopt# device without a

loaded image does not

show a backing device

Loaded vtopt#

device shows

name of image

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 3-15. Manage repository with cfgassist utility QV3441.0

Notes:

Manage the repository and media files with the cfgassist utility

Use the Virtual Media Repository submenu to list, create, remove or change the size

of the repository. You can also add or remove repository image files, change their

names, and change their access permissions between read-only and read-write.

Use the File Backed Virtual Optical Devices submenu to manage the FB virtual

optical drives. These are known by their VTD names, such as vtopt0. You can list all of

the devices, list their contents (loaded images), and load and unload images.

UNIX Software Service Enablement

Manage repository with cfgassist utility

 The cfgassist utility can be used to manage the repository and

its media files

 To manage the repository use:

Storage Management ÆVirtual

Media Repository

 To manage media files use: Devices -> Virtual Devices -> File

Backed Virtual Optical Devices

List Contents of the Repository

Create the Repository

Change Characteristics of the Repository

Remove the Repository

Virtual Optical Media Disks

List All File Backed Virtual Optical Devices

List Contents of a File Backed Virtual Optical Device

Load a Virtual Optical Media Disk into Virtual Optical Device

Remove a Virtual Optical Media Disk from Virtual Optical Device

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

3-18 VIOS © Copyright IBM Corp. 2012

Figure 3-16. Viewing FB configuration from the HMC QV3441.0

Notes:

Viewing the configuration from the HMC

The graphic in the visual above shows two FB virtual devices, along with other types of

devices. This panel shows the information from both the VIOS perspective and from the

client LPAR’s perspective. It shows the VIOS vhost adapter names, the associated

backing devices, the remote (client in this case) partition names, adapter IDs, and the

device names on the client.

In the example above, the vhost1 adapter has a FB virtual disk and the vhost2 adapter

has a FB virtual optical media device with the aixopt1 image loaded.

The HMC panel shown in the visual above does not show cd devices in the remote

backing device column.

UNIX Software Service Enablement

Viewing FB configuration from the HMC

 View the FB virtual configuration from the HMC

 Select the VIOS partition and run Hardware Information ÆVirtual I/O

Adapters ÆSCSI task

Example output:

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 3-17. Manage FB devices with the HMC (1 of 3) QV3441.0

Notes:

Using the HMC to manage FB devices

You can also use the HMC Virtual Storage Management task to create and manage

FB disks and optical media devices. The above visual shows an example screen which

lists the storage pools on the VIOS partition. In the example, there are three volume

groups/logical volume storage pools and there is one file-backed storage pool. Recall

that you must create a FB storage pool inside a LV storage pool before being able to

create FB virtual disks.

Select Action menu

If you select the storage pool, the actions menu will allow you to see the properties of

the pool, and extend or reduce the pool.

UNIX Software Service Enablement

Manage FB devices with the HMC (1 of 3)

 You can manage file-backed devices from the Configuration ->

Virtual Resources -> Virtual Storage Management task

Storage Pools tab will create and show file-backed storage pools

 When you create a storage pool, the options are File-Based and Volume

Group-Based

Action Menu

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

3-20 VIOS © Copyright IBM Corp. 2012

Figure 3-18. Manage FB devices with the HMC (2 of 3) QV3441.0

Notes:

Virtual Disks tab

Recall that you can create LV-based backing storage with the Virtual Disks tab. You can

also create FB backing storage as well. If you select the FB disk, you can extend it,

reduce it, modify its assignment, or look at its properties.

UNIX Software Service Enablement

Manage FB devices with the HMC (2 of 3)

Virtual Disks tab manages file-backed disks

Action Menu

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 3-19. Manage FB devices with the HMC (3 of 3) QV3441.0

Notes:

Optical devices tab

You can manage virtual optical media devices from the HMC. At the top of the panel as

shown in the visual above, you can manage physical optical devices. For example, the

physical /dev/cd0 drive on the VIOS partition can be mapped to a vhost adapter. The

client sees a virtual SCSI optical media drive.

At the bottom of the panel, you can manage file backed virtual optical media. If there is

no Virtual Media Repository, you can create one. The example above shows a system

on which a repository already exists. Once the repository exists, you can extend it,

delete it, or manage the media images. Notice that the Actions menu lists the ability to

add media, modify the partition assignment, and delete media.

The Modify Assignment action menu option will do both the mkvdev to create the VTD

and load the media.

UNIX Software Service Enablement

Manage FB devices with the HMC (3 of 3)

Optical Devices tab manages optical devices and media repository

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

3-22 VIOS © Copyright IBM Corp. 2012

Figure 3-20. Adding image file from HMC GUI QV3441.0

Notes:

Adding media files to the repository

If you choose to create a blank media file an option to enter a size will appear.

When adding existing files from the VIOS file system, the file must be accessible from

/home/padmin directory and be owned by the padmin user.

UNIX Software Service Enablement

Adding image file from HMC GUI

 You can add an existing file, import from a physical device, or create

blank media using the HMC GUI

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 3-21. Checkpoint QV3441.0

Notes:

UNIX Software Service Enablement

Checkpoint

1. True or False: An LPAR can boot from a file-backed hdisk or optical

media device.

2. True or False: The operating system for an LPAR can be installed using

the media repository.

3. True or False: An FBPOOL is the same thing as a volume group.

4. The ____________________ command creates FB backing devices and

maps them to vhost adapters.

5. Before creating any file-backed virtual disks, what must you create first

inside the logical volume storage pool?

_____________________________________

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

3-24 VIOS © Copyright IBM Corp. 2012

Figure 3-22. Exercise 3 - File-backed storage devices QV3441.0

Notes:

UNIX Software Service Enablement

Exercise 3  File-backed storage devices

 Activity:

 Create a file-backed virtual disk

 Create a Virtual Media Repository and Store an iso File

 Install AIX Image From Virtual Optical Media Device

 Note:

 The root password on LPAR3 is passw0rd (0 = zero)

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 3-23. Unit summary QV3441.0

Notes:

UNIX Software Service Enablement

Unit summary

 There are two types of file-backed virtual SCSI devices

 VSCSI disks and virtual optical media devices

File-backed storage pools are created in logical volume storage pools

File-backed devices are stored in file-backed pools

A virtual media repository can be created to store virtual media

Virtual media can be loaded into virtual optical media devices and

used by clients like physical optical media devices

 Virtual media can even be used to install the operating system for a client LPAR

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

3-26 VIOS © Copyright IBM Corp. 2012

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty Unit 4. Virtual Fibre Channel Storage Devices

What this unit is about

The unit continues the discussion of virtual storage devices. Students

will learn about the implementation of virtual Fibre Channel adapters

and the concept of N-Port ID Virtualization (NPIV).

What you should be able to do

After completing this unit, you should be able to:

• Configure virtual Fibre Channel adapters

• Map a physical Fibre Channel port to a virtual Fibre Channel

adapter

• Describe benefits of using NPIV

• View N_Port login status

• Log in to all N_Ports

• Use the chsyscfg HMC command to alter the WWPNs in a

partition profile

• Tune the max_xfer_size and num_cmd_elems attributes of

virtual Fibre Channel adapters

How you will check your progress

Accountability:

• Checkpoint questions

• Machine exercises

References

IBM Power Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp

Virtual I/O Server product documentation from the IBM Power

Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp?t

opic=/p7hb1/iphb1kickoff.htm

IBM Techdoc titled AIX disk queue depth tuning for performance



available from:

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-2 VIOS © Copyright IBM Corp. 2012

http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/TD1

05745

The following IBM Redbooks documents can be found at:

http://www.redbooks.ibm.com

SG24-7940 IBM PowerVM Virtualization Introduction and

Configuration

SG24-7590 IBM PowerVM Virtualization Managing and Monitoring

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 4-1. Unit objectives QV3441.0

Notes:

UNIX Software Service Enablement

Unit objectives

After completing this unit, you should be able to:

 Configure virtual Fibre Channel adapters

 Map a physical Fibre Channel port to a virtual Fibre Channel adapter

 Describe benefits of using NPIV

 View N_Port login status

 Log in to all N_Ports

 Use the chsyscfg HMC command to alter the WWPNs in a partition

profile

 Tune the max_xfer_size and num_cmd_elems attributes of virtual

Fibre Channel adapters

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-4 VIOS © Copyright IBM Corp. 2012

Figure 4-2. Virtual Fibre Channel adapter (1 of 3) QV3441.0

Notes:

NPIV

This industry standard method allows multiple initiators to share a single physical port,

easing hardware requirements for Storage Area Networks. In the PowerVM case, the

VIOS partition will have the actual physical connection.

HBA

A host bus adapter (HBA) refers to any type of plug-in adapter card.

World wide port name (WWPN)

This is the unique identifier for a port on a Fibre Channel fabric. The fabric is a topology

where devices connect with each other via one or more network switches. Think of a

WWPN as being similar to a Media Access Control (MAC) address for Ethernet ports.

UNIX Software Service Enablement

Virtual Fibre Channel adapter (1 of 3)

 Virtual Fibre Channel adapters support the use of N-Port ID

Virtualization (NPIV)

An N_Port is a nodes Fibre Channel port, in this case it is the virtual

Fibre Channels port

 NPIV is a standardized method for virtualizing a physical N_Port

 Enables multiple virtualized N_Port IDs to be configured on a physical

adapter port

 Implementation requires NPIV support in the physical Fibre Channel

adapter and the first level SAN switch

 Enables LPARs to have virtual Fibre Channel Host Bus Adapters (HBAs),

each with a dedicated world wide port name (WWPN)

 This gives each virtual Fibre Channel HBA a unique SAN identity similar

to that of a physical adapter port

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 4-3. Virtual Fibre Channel adapter (2 of 3) QV3441.0

Notes:

Virtual Fibre Channel

With NPIV, the VIOS’s role is fundamentally different. The VIOS facilitates adapter

sharing only; there is no device level abstraction or emulation. Rather than a storage

virtualizer, the VIOS serving NPIV is a pass through, providing a Fibre Channel Protocol

(FCP) connection from the client to the SAN.

If you use VSCSI disks, you need to use the VIOS to create each VSCSI VTD and map

each to the vhost adapter for each client. With NPIV, the SAN can zone storage to the

client LPAR’s WWPN. You do not need to any VTDs on the VIOS in order for the client

to see the storage. This reduces the amount of VIOS management needed.

Just like virtual SCSI, when configuring the virtual Fibre Channel adapters you specify

the VIOS partition name and virtual adapter ID in the client configuration, and you

specify the client information when configuring the VIOS server adapter.

UNIX Software Service Enablement

Virtual Fibre Channel adapter (2 of 3)

With virtual SCSI, the storage is virtualized

 The client operating system cannot distinguish between different types of

backing storage (SAN, SAS, etc.) because the VIOS uses SCSI

emulation

With NPIV, the Fibre Channel port is virtualized

 The VIOS serving NPIV:

 Acts as a pass through, providing an FC protocol (FCP) connection from the

client to the SAN

 Does not see the LUNs zoned for the clients WWPN

 The result is that the client partition has a unique identity to the SAN and

storage can be zoned to it, just as if it had a dedicated physical HBA

 The client can see storage zoned for its WWPN and recognize different

types of physical storage

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-6 VIOS © Copyright IBM Corp. 2012

Figure 4-4. Virtual Fibre Channel adapter (3 of 3) QV3441.0

Notes:

Requirements

Not all Fibre Channel adapters and SAN switches support NPIV. NPIV-capable

switches present the virtual WWPN to other SAN switches and devices as if they

represent physical FC adapter endpoints. Additional SAN switches and disk/tape

devices don’t need to be NPIV aware.

The VIOS needs at least one 8 Gigabit PCI Express Dual Port Fibre Channel Adapter

(feature code 5735) which supports NPIV. The SAN switch not only needs to support

NPIV, but it may have to be configured to enable NPIV.

UNIX Software Service Enablement

Virtual Fibre Channel adapter (3 of 3)

 Requires:

 POWER6 (or above) processor-based servers

 The physical Fibre Channel adapter and the first level SAN switch must

support NPIV

 VIOS 2.1.0.10 (fixpack 20.1) or later

 Update to VIOS V2.1 then install this fixpack

 HMC V7R3.4.0 (Service Pack 0) or later

 Supported client operating systems:

 AIX version 6.1 Technology Level 2, or later

 AIX 5.3 Technology Level 9, or later

 IBM i version 6.1.1, or later

 SUSE Linux Enterprise Server 11, or later

 RHEL 5.4 (with an additional driver), or later

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 4-5. NPIV benefits QV3441.0

Notes:

NPIV benefits

There are many benefits to using NPIV versus VSCSI to access disks. Some are

related to increased functionality such as the ability for the client operating system to

see the real type of disk that is being used. Client level tools often need this for certain

functions to work.

Some benefits are related to performance such as the ability to load more sophisticated

PCMs in the client which provide load balancing or other benefits when using MPIO.

Some benefits are for ease of administration such as not having to document all of the

client endpoint devices on the VIOS to prevent accidental overwriting of data. With Live

Partition Mobility, any virtual SCSI disks used by a client partition must be accessible by

all VIOS partitions involved in a relocation. This means that disks which look free on

one VIOS could be in use by an LPAR somewhere on the network. This has to be

documented carefully.

UNIX Software Service Enablement

NPIV benefits

 NPIV benefits versus VSCSI:

 Client can see the real type of disk rather than a SCSI emulation

 Easier for client tools to take advantage of copy service functions on SAN to

create flash copies or snap shots of the LUN

 Client can install a load-balancing path control module (PCM) to improve

functionality and performance

 VIOS doesnt have to keep track of all the clients devices

 Especially helpful for Live Partition Mobility environments

 For dual VIOS environments you no longer have to configure the

reserve_policy attribute for disks on both VIOSes

 Manage disk queue_depth disk attribute on client only, rather than on

client and VIOS

 To add VSCSI devices, its a two step process to add the device first to

the VIOS, then map it to the vhost adapter, but NPIV storage simply

needs to be zoned to the WWPN of the client partition

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-8 VIOS © Copyright IBM Corp. 2012

Figure 4-6. NPIV considerations QV3441.0

Notes:

NPIV considerations

While the ability to load PCMs in the client LPAR for added functionality and

performance is a benefit, this is also software that has to be maintained. If there are a

hundred LPARs using a PCM or drivers that have to be updated, this can take some

system administration time. This is not an issue if the client LPARs are using the

standard AIX MPIO PCM.

Related to the maintenance of PCMs and drivers, if these must be updated and the

client LPAR is using an NPIV disk or disks for rootvg, then as part of the update

process, you may have to map a new VSCSI disk, move rootvg to the VSCSI disk,

perform the update, then move rootvg back to the NPIV disk. It’s also possible that the

LPAR would need to be rebooted. For this reason, one recommendation is to use a

VSCSI disk or disks for rootvg and use NPIV for all other disks.

UNIX Software Service Enablement

NPIV considerations

 Considerations

 MPIO PCMs must be installed and updated in every client LPAR that

wants to use them

 Best to use non-NPIV disks for rootvg to avoid difficulties when updating

MPIO disk drivers

 With NPIV disks, you would have to move rootvg temporarily to another type

of disk (vSCSI for example), do the update, move it back, and potentially

reboot

 A VSCSI benefit is to provide access to storage that is supported on the

VIOS but not in a particular clients operating system

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 4-7. NPIV overview QV3441.0

Notes:

NPIV configuration

There is a one-to-one mapping of virtual Fibre Channel adapters between the client

LPAR and the VIOS, just like VSCSI adapters. The VIOS maps a physical FC port to its

virtual Fibre Channel server adapter. Then the administrator in the client LPAR contacts

the SAN system administrator to map LUNs. The client does not have to contact the

VIOS administrator to add endpoint devices.

UNIX Software Service Enablement

NPIV overview

 Like VSCSI, virtual Fibre Channel with

NPIV is implemented with a physical

adapter and a virtual server adapter on

the VIOS, and a virtual client adapter

on the client

 The client LPAR's virtual Fibre

Channel client adapter is associated

with a virtual Fibre Channel server

adapter in a VIOS

 VIOS maps the virtual Fibre Channel

server adapter to a port on a physical

Fibre Channel adapter

 The VIOS does not see the endpoint

devices

fcs0

hdisk0

LUN on SAN

Disk A

vfchost0

VIOS

LPAR

VIOS acts as a pass-through for

client I/O over the FC adapter

fcs0

Virtual

N_Port

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-10 VIOS © Copyright IBM Corp. 2012

Figure 4-8. Add virtual Fibre Channel adapters (1 of 2) QV3441.0

Notes:

Adding a virtual Fibre Channel adapter

The visual above shows using the HMC GUI to create the virtual Fibre Channel server

adapter. Just like virtual SCSI adapters, it can be created in the Create LPAR wizard, in

a partition’s profile, or dynamically.

One difference for virtual FC adapters versus VSCSI adapters is that when creating the

server adapter, there is no choice to connect to any client. You must specify the client

and its client adapter ID.

It is best to fully configure the VIOS FC server adapter and map it to a physical port

before creating the virtual FC client adapter. If you create the client adapter first, you’ll

see a message in the HMC that the VIOS is not configured.

UNIX Software Service Enablement

Add virtual Fibre Channel adapters (1 of 2)

 Add the virtual Fibre Channel adapters in a client/server pair

 Dynamically or add to profile and activate

 Configure the VIOS server adapter and map the port before creating the

client adapter

Use Actions -> Create Virtual Adapter -> Fibre Channel Adapter

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 4-9. Add virtual Fibre Channel adapters (2 of 2) QV3441.0

Notes:

View added virtual Fibre Channel adapter

The virtual Fibre Channel adapter uses a virtual adapter slot like other virtual adapters

and its properties can be seen in the partition properties or the partition profile

properties.

On the VIOS, the virtual FC server adapter is named vfchost#. On the client, the virtual

FC client adapter is named just like a physical adapter port (fcs#).

UNIX Software Service Enablement

 Resulting client Fibre Channel server adapter configuration

 List the virtual FC server adapter in the VIOS:

$ lsdev -virtual | grep vfchost

vfchost0 Available Virtual FC Server Adapter

Add virtual Fibre Channel adapters (2 of 2)

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-12 VIOS © Copyright IBM Corp. 2012

Figure 4-10. View available N_Ports QV3441.0

Notes:

Check available ports on the VIOS

The lsnports command is run on the VIOS to see what physical FC ports are available

and NPIV capable. If the fabric column shows a 0, then this port is not available for

NPIV. Perhaps the switch connected to that port needs to be configured or does not

have NPIV capability. You can also check cables and hardware compatibility.

The lsnports output shown in the visual above shows that each port can support 64

virtual ports, so that a maximum of 64 client LPARs can share each physical port. The

output above also shows that this system has one virtual port.

UNIX Software Service Enablement

View available N_Ports

 List the available physical NPIV capable ports on the VIOS with the

lsnports command

$ lsnports

name physloc fabric tports aports swwpns awwpns

fcs0 U789C.001.DQDC383-P1-C1-T1 1 64 63 2048 2037

fcs1 U789C.001.DQDC383-P1-C1-T2 1 64 64 2048 2037

Fields:

Fabric: 1=NPIV ready; 0=not ready for NPIV

Tports: Total number of virtual ports

Aports: Number of available (unused) virtual ports

Swwpns: Total number of client WWPNs supported

Awwpns: Number of client WWPNs available

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 4-11. Map virtual Fibre Channel server adapter QV3441.0

Notes:

Unmapping

If no parameter is specified after the -fcp flag in the vfcmap command, the command

unmaps the virtual Fibre Channel adapter from the physical Fibre Channel port. For

example: $ vfcmap -vadapter vfchost0 -fcp

UNIX Software Service Enablement

Map virtual Fibre Channel server adapter

 Map a virtual FC server adapter to a port on the physical HBA with

the vfcmap command

 Example where vfchost0 is the virtual server adapter name and fcs0 is

the physical Fibre Channel port:

$ vfcmap vadapter vfchost0 fcp fcs0

 Or, use the Configuration -> Virtual Resources -> Virtual Storage

Management -> Virtual Fibre Channel tab to map FC server

adapter to an FC port in HMC GUI

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-14 VIOS © Copyright IBM Corp. 2012

Figure 4-12. View Fibre Channel mapping QV3441.0

Notes:

View NPIV virtual mapping

The example lsmap command output in the visual above shows an example when a

virtual Fibre Channel adapter is configured and the client LPAR is currently active. If the

client LPAR is not active, the output will be similar to this:

$ lsmap -all -npiv

Name Physloc ClntID ClntName ClntOS

------------- ---------------------------------- ------ -------------- -------

vfchost0 U8202.E4B.06BCC8P-V1-C16 2

Status:NOT_LOGGED_IN

FC name:fcs0 FC loc code:U78AA.001.WZSGHVK-P1-C4-T1

Ports logged in:0

Flags:4<NOT_LOGGED>

VFC client name: VFC client DRC:

UNIX Software Service Enablement

View Fibre Channel mapping

 Use the lsmap all npiv command for NPIV mapping

information

 This example reflects a configuration in which the client LPAR has

been configured and is active:

$ lsmap -all -npiv

Name Physloc ClntID ClntName CIntOS

========= ========================== ====== =========== ======

vfchost1 U8203.E4A.0666BF2-V1-C16 2 LPAR1 AIX

Status:LOGGED_IN

FC name:fcs0 FC loc code:U789C.001.DQDC383-P1-C1-T1

Ports logged in:1

Flags:a<LOGGED_IN,STRIP_MERGE>

VFC client name:fcs2 VFC client DRC:U8203.E4A.0666BF2-V2-C5-T1

Virtual server adapter location code

Physical adapter port name and location code

Clients virtual adapter name and location code

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 4-13. Zone LUNs on the SAN using client WWPNs QV3441.0

Notes:

WWPNs

The lscfg -vl fcs# AIX command output will show the active WWPN in the Network

Address field.

This HMC command will show WWPNS for the managed system MSname:

lshwres -m MSname -r virtualio --rsubtype fc --level lpar

lpar_name=LPAR1,lpar_id=2,slot_num=5,adapter_type=client,state=1,is_r

equired=0,remote_lpar_id=1,remote_lpar_name=VIOS,remote_slot_num=15,"

wwpns=c0507600417e0004,c0507600417e0005"

A SAN uses zones to provide access to the target storage based on WWPNs. Zoning

indicates which WWPNs are permitted to access each LUN. Use the SAN subsystem

management tool to zone the WWPNs that are created by the HMC for the virtual Fibre

Channel client adapter. This is the same procedure on the SAN that is required when

using physical Fibre Channel adapters.

UNIX Software Service Enablement

Zone LUNs on the SAN using client WWPNs

 The SAN uses zones to provide access to targets based on WWPNs

 See WWPNs on client's adapter properties or use the

lscfg vl fcs#command on the AIX client

 Include both WWPNs to support Live Partition Mobility operations

 Use only first WWPN if not using Live Partition Mobility

Go to the Virtual Adapters tab on

the client's partition properties.

Then, view the properties for the

virtual adapter.

WWPNs

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-16 VIOS © Copyright IBM Corp. 2012

Figure 4-14. Virtual Fibre Channel DLPAR operations QV3441.0

Notes:

DLPAR operations for virtual Fibre Channel adapters

WWPNs are generated based on the range of names available for use based on a

prefix in the vital product data on the managed system. This 6 digit prefix comes with

the purchase of the managed system and includes the ability to use 32,000 pairs of

WWPNs.

When adding a virtual Fibre Channel client adapter, a pair of WWPNs is generated and

assigned. If you add the adapter with a DLPAR operation be sure to save the current

configuration as shown in the visual above. If, instead, you add the Fibre Channel client

adapter to the profile, then the next time the LPAR is activated, a new different pair of

WWPNs will be generated and of course this will affect connectivity to any storage

assigned to the LPAR. The Operations -> Save Current Configuration task will give

the option to save the configuration as a new profile or to overwrite an existing profile.

UNIX Software Service Enablement

Virtual Fibre Channel DLPAR operations

 Each time a virtual Fibre Channel client adapter is created, the HMC

obtains a new, non-reusable, pair of WWPNs from the Hypervisor

 WWPNs obtained by DLPAR add operations for virtual Fibre Channel

client adapters that are not saved to a profile will be discarded if the

partition is shut down or if the adapters are removed

 For continued access to the storage, save the partition configuration

in a new profile so that the WWPNs are saved

 The Hypervisor has a limit of 32,000 pairs of WWPNs

 If you run out of WWPNs, you must obtain an activation code for an

additional set of 32,000 pairs

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 4-15. Activating all WWPNs to aid SAN mapping QV3441.0

Notes:

SAN mapping

For most data centers, there are many procedures to follow. For large data centers this

is very much the case. Often one of those procedures is that the SAN administrator will

zone only WWPNs that are active on the fabric, and thus can be seen. They may not

want to or be allowed to enter WWPNs manually. This causes a problem if the client

LPAR administrator needs both of the WWPNs for a virtual Fibre Channel adapter to

have access to all LUNs. There are two strategies to solve this situation. The first

strategy is to use the chnportlogin HMC command to activate (login to the SAN) both

WWPNs of each client virtual FC adapter in a specified client LPAR. If this strategy is

not available, then the client LPAR administrator can switch the WWPNs in the LPAR

profile. This is covered later in this unit.

The lsnportlogin and chnportlogin commands are available as of HMC V7R7.3.0

and VIOS 2.2.0.13 FixPack 24 SP03

UNIX Software Service Enablement

Activating all WWPNs to aid SAN mapping

 Only active WWPNs will be seen by a SAN subsystem

 Only the first WWPN listed in the adapter properties and shown by the

lscfg vl fcs#AIX command will be active by default

 The second WWPN for a virtual Fibre Channel adapter is for Live Partition

Mobility and is only active for a short time during relocation

 While SAN subsystems have the ability to add non-active WWPNs, some

data center policies do not allow this

 View login status of N_Ports with the lsnportlogin HMC

command

 The chnportlogin HMC command provides ability to log in to all

WWPNs so the SAN administrator can see them

 After the ports are mapped, the inactive ones can be logged out with

the chnportlogin HMC command

 Log out of the non-active ports before doing a Live Partition Mobility

relocation

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-18 VIOS © Copyright IBM Corp. 2012

Figure 4-16. Logging into N_Ports from the HMC CLI QV3441.0

Notes:

Use lsnportlogin to view current login status for N_Ports. Use chnportlogin to log

the ports in and out of the SAN fabric.

The first time lsnportlogin is run, the status will be 2 which means unknown. Simply

the system when this initial unknown status is listed.

The --filter lpar_names=LPARname is a required flag for the lsnportlogin

command.

The chnportlogin will log out ports (WWPNs) that are not really active. That is, it will

not log out ports (WWPNs) that are actively in use to access disks. You’ll notice in the

lsnportlogin output that the truly active ports will show that they are logged in by the

client. The non-active ports will show that they’re logged in by the VIOS.

UNIX Software Service Enablement

Logging into N_Ports from the HMC CLI

 The first time the lsnportlogin command is run, all ports

appear to be in the unknown state (status is 2)

 Log in then log out to see valid status

 Example log in command:

hscroot@hmc:~> chnportlogin -m john -p LPAR1 -o login

 Example log out command:

hscroot@hmc:~> chnportlogin -m john -p LPAR1 -o logout

 Example list command that shows one port logged in by the client:

hscroot@hmc:~> lsnportlogin -m john --filter lpar_names=LPAR1

lpar_name=LPAR1,lpar_id=2,profile_name=Install,slot_num=5,wwpn=c05076

042d6f0004,wwpn_status=1,logged_in=client

lpar_name=LPAR1,lpar_id=2,profile_name=Install,slot_num=5,wwpn=c05076

042d6f0005,wwpn_status=0

 Any ports that were inactive when the chnportlogin command was

run to login to the fabric will show that they have been logged in by

the VIOS that owns the physical adapter

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 4-17. Alternate procedure to activate WWPNs (1 of 2) QV3441.0

Notes:

Alternate procedure

The procedure on the visual above addresses the problem of the virtual FC adapter’s

WWPN not appearing active on the fabric and the LPAR is not yet loaded with an

operating system. This might happen if the LPAR will only have NPIV disks. If the LPAR

has an operating system and is booted, then the first WWPN would be active and the

second could be logged in with the chnportlogin command.

Following the procedure above will allow the SAN administrator to see the first WWPN

as active and logged in on the SAN fabric. If you need to also show the second WWPN

as active on the fabric once an operating system is installed in the client LPAR, use

chnportlogin or follow the procedure on the next visual.

UNIX Software Service Enablement

Alternate procedure to activate WWPNs (1 of 2)

 If you need to make the first WWPN active before installing an LPAR,

activate to Open Firmware (OF) mode and run the following

procedure:

 Type ioinfo at the first OF prompt:

0> ioinfo

 At the next prompt choose option 6. FCINFO

==> 6

 At the next prompt chose the FC adapter. It should be option 1.

==> 1

 Then choose 1. List Attached FC Devices

==> 1

 After you press ENTER the first WWPN will be logged in

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-20 VIOS © Copyright IBM Corp. 2012

Figure 4-18. Alternate procedure to activate WWPNs (2 of 2) QV3441.0

Notes:

Alternate procedure

The visual above addresses the problem of the second WWPN not appearing active on

the fabric and the system is running code before the chnportlogin and lsnportlogin

commands are available. The first WWPN is active on the fabric if the LPAR is activated

or if you follow the OF procedure on the last visual.

The overall process is to have the SAN administrator zone the first WWPN which can

be seen, then switch the WWPNs in the profile and reactivate the LPAR. Now the SAN

administrator can zone the second (now the first) WWPN. This can take some time if

there is a week or longer turnaround time for SAN requests in large data centers!

This procedure can also be useful for another purpose. If a client administrator has a

particular set of WWPNs already configured with LUNs and for some reason the virtual

FC adapter with those WWPNs is deleted or otherwise no longer available, the

administrator can use chsyscfg to replace the WWPNs with the original ones.

UNIX Software Service Enablement

Alternate procedure to activate WWPNs (2 of 2)

 If you need to make the second WWPN active, switch the WWPNs in

the profile and then activate LPAR

 Use the chsyscfg HMC command to switch WWPNs:

chsyscfg -r prof -m MSname -i "name=profile,lpar_name=LPARname,

\"virtual_fc_adapters=\"\"#/client/#/VIOS/#/ZZZ,ZZZ/1\"\"\""

 Format for the virtual_fc_adapters attribute:

Client slot ID /client/ VIOS ID / VIOS name / VIOS slot ID / WWPN1,WWPN2 / required

 Example command:

chsyscfg -r prof -m john -i

"name=Normal,lpar_name=LPAR1,\"virtual_fc_adapters=\"\"5/

client/1/VIOS/15/c05076042d6f0005,c05076042d6f0004/1\"\"\""

This command can be also used to configure a different set of WWPNs.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 4-19. Activating WWPNs from the HMC GUI (1 of 2) QV3441.0

Notes:

Instead of using the HMC CLI chnportlogin and lsnportlogin commands, you can

use the HMC GUI. Select the client LPAR and run the Properties task. Go to the Virtual

Adapters tab. Use the Actions -> Advanced task. This will popup a window that is

shown on the next visual.

UNIX Software Service Enablement

Activating WWPNs from the HMC GUI (1 of 2)

 Partition Properties -> Virtual Adapters tab

2) Run: Actions ->

Advanced -> Login

Logout Fibre Channel

1) Select the adapter

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-22 VIOS © Copyright IBM Corp. 2012

Figure 4-20. Activating WWPNs from the HMC GUI (2 of 2) QV3441.0

Notes:

The visual above shows three HMC GUI panels. The one at the top shows all of the

WWPNs available on the client. The login status is shown in the third column from the

left. In the example shown above, only one WWPN is currently logged in with a login

status of 1. Click the Login button to log all the ports into the fabric.

The middle panel shows the working window while the login (or logout) process is

running.

The third panel on the bottom shows a panel where all of the WWPNs are logged in.

Notice that the first WWPN is the truly active WWPN because the client is logged in.

The second WWPN shows that the VIOS is logged in. If you were to click the Logout

button at this point, the screen would update to look like the first panel again. Note that

client and vios are key words and not partition names.

UNIX Software Service Enablement

Activating WWPNs from the HMC GUI (2 of 2)

 Use the Login and Logout buttons

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 4-21. Monitoring and tuning virtual Fibre Channel adapters (1 of 2) QV3441.0

Notes:

Monitoring

Monitor the virtual Fibre Channel adapter and any attached disks just like you would

physical devices. The nmon and fcstat commands are mentioned on the visual above

as example commands to monitor FC adapter statistics.

max_xfer_size

There are various strategies to tune the max_xfer_size attribute on physical FC

adapters. That is beyond the scope of this course. Once the VIOS administrator has

determined the proper physical FC adapter max_xfer_size value, then you should

configure any virtual FC adapters using the same value.

The example commands on the visual above show how to check the attribute on the

VIOS and on the client AIX LPAR. Use the chdev command with the -P flag to change

the attribute in the ODM then reboot.

UNIX Software Service Enablement

Monitoring and tuning virtual Fibre Channel

adapters (1 of 2)

 Monitor virtual Fibre Channel activity like a physical adapter

Use nmon and fcstat

 Use the ^ character to get to the nmon FC adapter monitor screen

 On the VIOS, nmon must be run from root environment

 Tune the max_xfer_size attribute of the clients vFC adapter (fcs#) to

match the physical adapter

 Controls the maximum I/O size the adapter will handle

Example:

$ lsdev -dev fcs0 -attr max_xfer_size

value

0x200000

# lsattr -El fcs0 | grep max_xfer_size

max_xfer_size 0x100000 Maximum Transfer Size True

# chdev -l fcs0 -a max_xfer_size=0x200000 -P

fcs0 changed

On VIOS

On client

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-24 VIOS © Copyright IBM Corp. 2012

Figure 4-22. Monitoring and tuning virtual Fibre Channel adapters (2 of 2) QV3441.0

Notes:

num_cmd_elems

There are various strategies to tune the num_cmd_elems attribute on physical FC

adapters. That is beyond the scope of this course. Once the VIOS administrator has

determined the proper physical FC adapter num_cmd_elems value, then you should

configure any virtual FC adapters to an appropriate value. Similar to tuning the

queue_depth value for virtual disks, the appropriate value for num_cmd_elems varies

based on the configuration and workload. Recall that up to 64 virtual Fibre Channel

adapters can be connected to the same physical adapter port. If the physical adapter

can handle 200 commands, and all 64 virtual FC adapters send 200 commands all at

once, that is too many for the adapter to handle. If the physical adapter can handle

2048 commands and there is only one virtual FC adapter that is sending only 200

commands, then you may see improved performance with higher workloads by

increasing the virtual FC attribute value to 2048. The example commands above show

how to check the attribute on the VIOS and on the client AIX LPAR. Use the chdev

command with the -P flag to change the attribute in the ODM then reboot.

UNIX Software Service Enablement

Monitoring and tuning virtual Fibre Channel

adapters (2 of 2)

 Tune the num_cmd_elems attribute of the vFC adapter

 Number of commands the adapter can handle (command queue depth)

 Tune similar to the way queue_depth is tuned for virtual disks

 If the vFC adapter is the only adapter using a physical FC port, then make

the value on the virtual adapter the same as the physical adapter

 If there are multiple vFC adapters for one physical FC port, decide on a

dedicated or shared strategy

 Maximum value is 2048

Example:

$ lsdev -dev fcs0 -attr num_cmd_elems

value

500

# lsattr -El fcs0 | grep num_cmd_elems

num_cmd_elems 200 Maximum Number of COMMAND Elements True

# chdev -l fcs0 -a num_cmd_elems=500 -P

fcs0 changed

On VIOS

On client

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 4-23. Checkpoint (1 of 2) QV3441.0

Notes:

UNIX Software Service Enablement

Checkpoint (1 of 2)

1. True or False: When a client partition uses a virtual Fibre Channel

adapter to access its disks, the VIOS hosting the physical adapter does

not see the clients disks.

2. True or False: You must zone SAN LUNs to both the virtual FC adapter

WWPNs and the VIOSs physical adapter port WWPN.

3. When using virutal Fibre Channel instead of virtual SCSI, which of the

following is a false statement:

A. VIOS administrator doesnt have to keep documentation about all of the client

endpoint devices.

B. Clients can use other path control modules with added benefits such as load

balancing.

C. NPIV provides an emulated environment.

D. Clients see real type of disk so tools which are specific to a certain type of disk

will work.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-26 VIOS © Copyright IBM Corp. 2012

Figure 4-24. Checkpoint (2 of 2) QV3441.0

Notes:

UNIX Software Service Enablement

Checkpoint (2 of 2)

4. The name for the virtual FC client adapters port is:

A. vfchost#

B. F_Port

C. N_Port

D. VTD

5. Once a client has a virtual Fibre Channel adapter connection through the

VIOS which supports NPIV, it has a unique identifier on the SAN fabric

called a _________________________.

6. The name of the virtual FC server adapter device on the VIOS is

_______________________.

7. True or False: The chsyscfg HMC command workaround procedure will

alter the running configuration of an LPAR.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 4-25. Exercise 4 - Virtual Fibre Channel storage devices QV3441.0

Notes:

UNIX Software Service Enablement

Exercise 4  Virtual Fibre Channel storage

devices

 Activity:

 Configure virtual Fibre Channel NPIV disks

 View information about the NPIV configuration

 Use the HMC to log WWPNs into and out of the SAN fabric

 Monitor activity on an NPIV device

 Check virtual Fibre Channel tunables

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

4-28 VIOS © Copyright IBM Corp. 2012

Figure 4-26. Unit summary QV3441.0

Notes:

UNIX Software Service Enablement

Unit summary

Virtual Fibre Channel adapters are implemented with a client-server

virtual adapter model similar to VSCSI adapters

N-Port ID Virtualization (NPIV) allows a client LPAR to manage SAN

disk storage at the SAN storage manager rather than having to create

and map individual backing storage devices on the VIOS

 Virtual FC adapters are configured in pairs with one on the VIOS and

one on the client LPAR

 The client virtual FC adapter has two WWPNs and both should be

zoned for LUNs

 The HMC CLI or GUI can be used to cause all of the virtual FC

adapter WWPNs in a client LPAR to login to the SAN fabric

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty Unit 5. Virtual Ethernet Adapter Configuration

What this unit is about

This unit describes virtual Ethernet adapters, VLAN IDs, the Virtual

Network Management task, and virtual Ethernet switches.

What you should be able to do

After completing this unit, you should be able to:

• Describe how virtual Ethernet is implemented on POWER

processor-based systems

• Describe the IEE 802.1Q VLAN-aware Ethernet switch functionality

provided by the Hypervisor

• Configure communications between partitions using virtual

Ethernet adapters

• Describe the configuration of the Hypervisor virtual Ethernet

switches

How you will check your progress

Accountability:

• Checkpoint questions

• Machine exercises

References

IBM Power Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp

Virtual I/O Server product documentation from the IBM Power

Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp?t

opic=/p7hb1/iphb1kickoff.htm

IBM Redbooks documents can be found at:

http://www.redbooks.ibm.com

SG24-7940 IBM PowerVM Virtualization Introduction and

Configuration, an IBM Redbooks document

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

5-2 VIOS © Copyright IBM Corp. 2012

SG24-7590 IBM PowerVM Virtualization Managing and Monitoring, an

IBM Redbooks document

REDP4194 IBM System p Advanced POWER Virtualization

(PowerVM) Best Practices, an IBM Redbooks

document

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-1. Unit objectives QV3441.0

Notes:

UNIX Software Service Enablement

Unit objectives

After completing this unit, you should be able to:

 Describe how virtual Ethernet is implemented on POWER processor-

based systems

 Describe the IEEE 802.1Q VLAN-aware Ethernet switch functionality

provided by the Hypervisor

 Configure communications between partitions using virtual Ethernet

adapters

 Describe the configuration of the Hypervisor virtual Ethernet switches

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

5-4 VIOS © Copyright IBM Corp. 2012

Figure 5-2. Virtual Ethernet overview QV3441.0

Notes:

Virtual Ethernet overview

Virtual Ethernet enables inter-partition communication without the need for physical

network adapters assigned to each partition. This technology enables IP-based

communication between logical partitions on the same system.

The AIX virtual Ethernet device driver is limited to controlling 256 virtual Ethernet

devices.

There is no extra software or system feature that must be purchased to use virtual

Ethernet adapters. You do not need a Virtual I/O Server unless you wish to bridge the

internal virtual Ethernet traffic to an external network with a Shared Ethernet adapter.

UNIX Software Service Enablement

Virtual Ethernet overview

 Virtual Ethernet on POWER5 and higher processor-based systems:

 Enables inter-partition communication without the need for physical

network adapters assigned to each partition:

 Adapters are virtual and if communication is only between partitions on

same managed system, no physical Ethernet adapter is needed

 Partitions can run either AIX V5.3 and up, Linux, or IBM i 6.1 or above

 256 virtual Ethernet adapters supported per partition for AIX

 Support for large MTUs (up to 65390) on AIX

 Supports both IPv4 and IPv6 standards

 Implemented using the VLAN IEEE 802.1Q standard

 A VLAN ID number is used to designate a virtual network

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-3. POWER Hypervisor Ethernet switch QV3441.0

Notes:

Introduction

The POWER Hypervisor switch allows partitions to communicate with each other by

using virtual Ethernet adapters and assigning VLAN IDs (VIDs) that enable them to

share a common logical network.

The virtual Ethernet adapters are created and the VID assignments are performed

using the Hardware Management Console (HMC) or the Integrated Virtualization

Manager (IVM). The system requires virtual Ethernet adapters to be configured with a

default VID that will be used by the POWER Hypervisor to tag untagged packets. The

default VID is called the Port VID (PVID).

The POWER Hypervisor transmits packets by copying the packet directly from the

memory of the sender partition to the receive buffers of the receiver partition without

any intermediate buffering of the packet.

The Hypervisor supports 4094 VLANs (1 - 4094). VLAN IDs 0 and 4095 are reserved.

UNIX Software Service Enablement

POWER Hypervisor Ethernet switch

 The POWER Hypervisor provides the Ethernet switch functionality:

 Implements a software Ethernet switch

 Supports up to a maximum of 4094 networks (VLAN IDs)

 Copies packets between partitions (memory to memory)

 Consistent with VLAN IEEE 802.1Q standard

 The ports on this switch are configured via the HMC and are part of the

partitions configuration:

 Configuring a Port VLAN ID (PVID) is required on the Hypervisor

switch

 Can create up to 16 virtual Ethernet switches

 One by default

 Allows you to reuse VLAN IDs for separate virtual LANs

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

5-6 VIOS © Copyright IBM Corp. 2012

Figure 5-4. Virtual LAN (VLAN) overview QV3441.0

Notes:

VLAN Overview

Virtual LAN is a technology used for establishing virtual network segments on top of

physical switch devices. Multiple VLAN logical devices can be configured on a single

system. Each VLAN logical device constitutes an additional Ethernet adapter instance,

and can be used to configure the same types of Ethernet IP interfaces that are used

with physical Ethernet adapters.

VLAN support in AIX is based on the IEEE 802.1Q VLAN implementation. The IEEE

802.1Q VLAN is achieved by adding a VLAN ID tag to an Ethernet frame, and the

Ethernet switches restricting the frames to ports that are authorized to receive frames

with that VLAN ID. Switches also restrict broadcasts to the logical network by ensuring

that a broadcast packet is delivered to all ports which are configured to receive frames

with the VLAN ID that the broadcast frame was tagged with.

As VLAN ensures logical separation at layer 2, it is not possible to have an IP network

that spans multiple VLANs (different VIDs).

UNIX Software Service Enablement

Virtual LAN (VLAN) overview

 Virtual LAN (VLAN) establishes virtual network segments on top of

physical switch devices:

 Layer 2 connectivity restricted to members of the same VLAN:

 ISO OSI Model: structured seven layer networking model

 IP (and thus IP addresses and IP routing) is at Layer 3

 Virtual and dedicated network devices can be used simultaneously

for communication

 VLAN example with two VLANs:

Switch

Hosts

VLAN 1 VLAN 2

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-5. IEEE 802.1Q VLAN QV3441.0

Notes:

VLAN IEEE standard

The VID can be added to an Ethernet packet either by a VLAN aware host or by the

switch in the case of VLAN unaware hosts. Therefore ports on a VLAN-aware Ethernet

switch have to be configured with a PVID, and may be configured with additional VLAN

IDs.

For VLAN unaware hosts, the switch will tag all packets entering through that port with

the Port VLAN ID (PVID). It will also untag all packets tagged with the PVID that exit the

port for delivery to the VLAN unaware host.

Hosts that are VLAN aware insert and remove their own tags and can be members of

more than one VLAN. In this case, the switch does not have to insert and remove tags

for the host.

UNIX Software Service Enablement

IEEE 802.1Q VLAN

 Ethernet packets are tagged with their VLAN membership information

 Ethernet switch restricts delivery to members of that VLAN

 The VID can be added to an Ethernet packet either by:

 VLAN aware host

 The switch (in the case of VLAN unaware hosts)

 Recommendations

 Single network per adapter  use untagged port and PVID

 Multiple networks per adapter  use VID

 For POWER5 and higher processor-based systems:

 Virtual Ethernet adapter with only a PVID is configured just like physical Ethernet

adapter:

 The Hypervisor Ethernet switch (Hypervisor) uses the PVID tag for untagged

packets

 Virtual Ethernet adapter with multiple VIDs is configured as a VLAN device with

tagging

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

5-8 VIOS © Copyright IBM Corp. 2012

Figure 5-6. Virtual Ethernet configuration example (1 of 2) QV3441.0

Notes:

Simple configuration

In a simple configuration, virtual Ethernet adapters are defined only with PVIDs, and

LPARs configured with the same PVID can communicate with each other on the same

VLAN. In this scenario:

- The operating system running in such a partition is not aware of the VLANs.

- Only packets for the VLAN specified as PVID are received.

- Packets have their VLAN tag removed by the POWER Hypervisor before the

partitions receive them.

- Packets sent by these partitions have a VLAN tag attached for the VLAN specified

as PVID by the POWER Hypervisor.

When you create this type of adapter from the HMC, you do NOT check the IEEE

802.1Q checkbox, and only specify the adapter ID and PVID for the VLAN ID. This

configuration is demonstrated later in this unit.

UNIX Software Service Enablement

POWER Hypervisor (IEEE 802.1Q VLAN-aware switch)

LPAR 1

PVID 1

Switch Port

ent0

LPAR 2

ent0

PVID 10

Switch Port

LPAR 3

ent0

PVID 10

Switch Port

LPAR 4

ent0 ent1

PVID 10

Switch Port

PVID 1

Switch Port

PVID only example

Virtual Ethernet configuration example (1 of 2)

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-7. Virtual Ethernet configuration example (2 of 2) QV3441.0

Notes:

VLAN tagging

In the visual, all LPARs shown can communicate on VLAN ID 10, and all but LPAR2 can

communicate on VLAN ID 20. Examine the ways that VLAN access is provided for each

LPAR.

LPAR1 and LPAR3 define VLAN devices. These are ent1 and ent2 in LPAR1, and ent1

in LPAR3. Packets leaving these devices are tagged by the host. On the other hand,

LPAR2 and LPAR4 are VLAN-unaware hosts and use PVIDs to communicate on

VLANs 10 and 20. In this case, the POWER Hypervisor switch adds the VLAN tags to

packets when they enter the switch port from the virtual Ethernet adapter.

UNIX Software Service Enablement

POWER Hypervisor

LPAR 2

VLAN

Unaware

host

ent0

PVID 10

LPAR 4

VLAN

Unaware

host

ent0 ent1

PVID 10 PVID 20

ent1

Indicates VLAN device

ent0

LPAR 1

VLAN

Aware

host

PVID 1

VID 10, 20

VID 10 VID 20

ent0

LPAR 3

VLAN

Aware

host

ent1

PVID 10

VID 20

Multiple VLANs example

ent1 ent2

Virtual Ethernet configuration example (2 of 2)

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

5-10 VIOS © Copyright IBM Corp. 2012

Figure 5-8. Packets leaving virtual Ethernet hosts QV3441.0

Notes:

VLAN tagging example

In the visual, two scenarios are shown that indicate how VLAN tagging is handled when

a packet leaves through a virtual Ethernet adapter.

In the first scenario (labeled with the number ‘1’), configured VLAN adapters tag the

packets leaving through those devices. Tagged packets enter the switch port, and the

tags are left intact. The port will discard packets that are tagged with a VLAN ID that is

not configured on the port.

In the second scenario (labeled with the number ‘2’), virtual Ethernet adapters that are

only configured with a PVID do not tag packets. The PVID is added to packets at the

switch port before they are processed and forwarded to other ports on the switch.

UNIX Software Service Enablement

Packets leaving VLAN adapters are tagged with their VID

Packets leaving VLAN-unaware adapters are tagged with the ports PVID by

the Hypervisor

POWER Hypervisor

LPAR 1 LPAR 2

VLAN

Unaware

host

ent0

PVID 10

LPAR 4

VLAN

Unaware

host

ent0 ent1

PVID 10 PVID 20

ent0

LPAR 3

VLAN

Aware

host

ent1

PVID 10

VID 20

ent0

VLAN

Aware

host

PVID 1

VID 10, 20

VID

ent1 ent2

Packets leaving virtual Ethernet hosts

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-9. Packet delivered to virtual Ethernet hosts QV3441.0

Notes:

Packet delivery to virtual Ethernet hosts

The visual describes what happens when the POWER Hypervisor switch receives a

tagged packet for delivery to virtual Ethernet hosts. Each of the scenarios shown is

described below, where the packet being delivered is tagged with VID 20:

1. The switch port for the virtual Ethernet adapter in LPAR1 is configured to receive

packets on this VLAN. The tagged packet is delivered to the ent0 adapter, which

then presents the packet to the ent2 VLAN device.

2. The switch port for the adapter in LPAR2 is not configured for this VID, so the packet

is not delivered here.

3. The switch port for the adapter in LPAR3 is configured for VID 20, and it is delivered

to the ent0 adapter, which then presents the packet to the ent1 VLAN device.

4. The switch port for one of the adapters in LPAR4 is configured for VID 20, but since

it is the PVID, the tag is stripped before being delivered to ent1.

UNIX Software Service Enablement

MTagged packet is delivered to ent2

NPacket is NOT delivered (VID 20 not configured for this port)

OTagged packet is delivered to ent1

PTag is stripped by Hypervisor and untagged packet delivered to ent1

POWER Hypervisor

LPAR 2

VLAN

Unaware

host

ent0

LPAR 4

VLAN

Unaware

host

ent0 ent1

PVID 1

VID 10, 20

PVID 10 PVID 10

VID 20

PVID 10 PVID 20

VID 20 | IP 9.47.87.0

LPAR 1

ent0

VLAN

Aware

host

ent1 ent2

VID

LPAR 3

ent0

VLAN

Aware

host

ent1

VID

OPNM

Packet delivered to virtual Ethernet hosts

 Packet destined for VID

20 is delivered to virtual

Ethernet hosts via the

Hypervisor switch

 Tagged packets are only

delivered to ports

configured on that VLAN

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

5-12 VIOS © Copyright IBM Corp. 2012

Figure 5-10. Virtual Ethernet adapters QV3441.0

Notes:

MAC address

The MAC address that the managed system assigns to the virtual Ethernet adapter is

unique within the managed system. However, this MAC address is not universally

unique, and might not be unique within a network containing multiple managed

systems. Having multiple virtual Ethernet adapters with the same MAC address can

cause network problems. To resolve these problems, you can change the MAC address

assigned to a virtual Ethernet adapter using the chhwres command on the Hardware

Management Console (HMC) or by configuring an alternate MAC address in the

operating system. This may be necessary when an LPAR is moved to another server

using Live Partition Mobility if the current MAC address is the same as one on the

destination server.

UNIX Software Service Enablement

Virtual Ethernet adapters

 Virtual Ethernet adapters are required to access the Hypervisor

Ethernet switch

 Use HMC to create virtual adapters

 Created in a partitions profile or dynamically

 HMC generates MAC address

 Local bit to prevent conflict with physical Ethernet

 Based on machine serial number for uniqueness

 Access to 19 VLAN networks per adapter (plus the PVID)

VLAN 100 VLAN 200

VLAN 300

LPAR 1 LPAR 2 LPAR 3 LPAR 4 LPAR 5

POWER Hypervisor (switch)

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-11. Creating a virtual Ethernet adapter (1 of 6) QV3441.0

Notes:

Creating virtual Ethernet adapters

You can change the maximum number of virtual adapters allowed for the partition only

in the profile. The new value will be used at the next activation and is not dynamic.

The maximum virtual adapters value does not affect Hypervisor memory overhead,

however, the more virtual adapters you create, the more memory the Hypervisor will

require to support the active adapters (in other words, the number of active virtual I/O

adapters affects the amount of Hypervisor memory).

Virtual adapters can be created using the Dynamic Logical Partitioning menu option

on the HMC. After adding any devices dynamically, run cfgmgr in the AIX partition so

the operating system will discover the new devices.

If you only change the partition profile, you will need to re-activate the partition for the

changes to take effect.

UNIX Software Service Enablement

Creating a virtual Ethernet adapter (1 of 6)

 Configure adapters in a profile or using DLPAR

Currently configured

virtual adapters

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

5-14 VIOS © Copyright IBM Corp. 2012

Figure 5-12. Creating a virtual Ethernet adapter (2 of 6) QV3441.0

Notes:

Actions menu

Once the Virtual Adapters tab is visible, select Actions -> Create Virtual Adapter->

Ethernet Adapter. This will bring up the dialog shown in the next visual.

Notice the other actions listed in the menu. Configured adapters shown in the table on

this tab can be deleted by selecting them and choosing the Delete action. An adapter’s

configuration can be modified by selecting the adapter and choosing the Edit action. An

adapter’s properties can be displayed with the Properties action.

UNIX Software Service Enablement

Creating a virtual Ethernet adapter (2 of 6)

 Use the Actions menu to manage adapters

Actions -> Create Virtual Adapter-> Ethernet Adapter

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-13. Creating a virtual Ethernet adapter (3 of 6) QV3441.0

Notes:

Introduction

The adapter will be created in the lowest numbered available virtual slot (or the lowest

available adapter ID). The adapter ID can be changed if desired, within the limits of the

maximum number available for the partition. You should enter the Port VLAN ID that will

be used for this adapter. Each adapter can connect to a single VLAN when using only

the PVID.

The checkbox below the VLAN ID field is used to mark the adapter as required. While

virtual adapters are always available at activation, it is possible that the Hypervisor does

not have enough memory to activate all virtual I/O adapters configured in the profile.

The required setting indicates that this partition cannot start without this adapter, and

that the adapter cannot be removed with a DLPAR operation.

The View Virtual Network button can be selected to view Virtual Switch (VSwitch) and

VLAN configurations.

UNIX Software Service Enablement

Creating a virtual Ethernet adapter (3 of 6)

 Alter adapter ID and default VLAN ID (PVID) if necessary

 Default VLAN ID (PVID) is 1

Display current

system VLAN

configuration

Optional settings prior to HMC V7R730

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

5-16 VIOS © Copyright IBM Corp. 2012

Figure 5-14. Creating a virtual Ethernet adapter (4 of 6) QV3441.0

Notes:

View Virtual Networks

Selecting the View Virtual Network button from the previous visual brings up this

dialog box shown in the visual above. From this task, you can:

- View existing VSwitches and their associated VLANs

- View VLANs by their definition or by partition participation

UNIX Software Service Enablement

Creating a virtual Ethernet adapter (4 of 6)

 View virtual Ethernet configurations by clicking the View Virtual

Network  button

 Select a VLAN to view configuration details

Example where

VLAN 10 is

selected

Example where VLAN

1 is selected

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-15. Creating a virtual Ethernet adapter (5 of 6) QV3441.0

Notes:

Advanced tab

Beginning with POWER7 servers with firmware level 720 and HMC at V7R7.2.0, you

can use this tab to configure advanced options for your virtual Ethernet adapter. When

adding a virtual Ethernet adapter to a logical partition, you can configure a custom MAC

address for the adapter and specify restrictions for operating system-defined MAC

addresses.

You can prioritize the logical partition network traffic by specifying the value of IEEE

802.1Q priority level for each virtual Ethernet adapter in the Quality of Service setting.

You can disable the setting, or set priorities 0 through 7, 7 being the highest priority. You

can edit the QoS priorities using the dynamic logical partitioning operation when the

partition is running.

UNIX Software Service Enablement

Creating a virtual Ethernet adapter (5 of 6)

 The MAC address is automatically generated by the HMC

 On POWER7 servers, you can:

 Specify a custom MAC address

 Set Quality of Service

(QoS)

 Restrict MAC addresses

Beginning with

POWER7 servers, you

can use this tab to

configure advanced

options

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 5-16. Creating a virtual Ethernet adapter (6 of 6) QV3441.0

Notes:

View new adapter

Once the adapter has been created, it shows up in the Virtual Adapters panel of the

partition profile. For virtual Ethernet adapters, the adapter table shown in the visual

above includes the adapter ID, the adapter type, and whether it is required or not. In this

example, you can see that the virtual Ethernet adapter was created with adapter ID 5

and it is not required. The Adapter ID is clickable to view the adapter’s properties such

as what PVID is configured.

UNIX Software Service Enablement

Creating a virtual Ethernet adapter (6 of 6)

Newly created

adapter

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-17. View adapter and interfaces QV3441.0

Notes:

Virtual Ethernet adapters in AIX

Once the virtual Ethernet adapter has been added, you should activate the partition

using the newly modified profile. Or, if the adapter was added dynamically, run the

cfgmgr command to discover the new device. The virtual Ethernet adapter will be listed

as an available boot device if you enter the Software Management Services (SMS)

menus.

On boot or when running the cfgmgr command, AIX detects the presence of the newly

added virtual Ethernet adapters, and adds appropriate ODM device entries, for example

ent1 as shown in the visual above. Since the newly added device is an Ethernet

adapter, the configure method for the adapter also defines associated network

interfaces, for example, en1 and et1 as shown in the visual above.

The virtual adapter instance and associated interfaces do not have AIX location codes,

since they are virtual devices.

UNIX Software Service Enablement

View adapter and interfaces

 Activate the partition using the modified profile

 Adapter is visible to partition

# lsdev -c adapter

ent0 Available Logical Host Ethernet Port (lp-hea)

ent1 Available Virtual I/O Ethernet Adapter (l-lan)

lhea0 Available Logical Host Ethernet Adapter (l-hea)

vsa0 Available LPAR Virtual Serial Adapter

 Interfaces are available for configuration

 Virtual interfaces do not have an AIX location code

+--------------------------------------------------------------------------+

| Available Network Interfaces |

| |

| Move cursor to desired item and press Enter. |

| |

| en0 Standard Ethernet Network Interface |

| en1 Standard Ethernet Network Interface |

| et0 IEEE 802.3 Ethernet Network Interface |

| et1 IEEE 802.3 Ethernet Network Interface |

| |

| Esc+1=Help Esc+2=Refresh Esc+3=Cancel |

| Esc+8=Image Esc+0=Exit Enter=Do |

| /=Find n=Find Next |

+--------------------------------------------------------------------------+

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 5-18. Adding VLAN access (1 of 3) QV3441.0

Notes:

Configuring VLAN access

In order to access multiple VLANs, the adapter must be configured as an IEEE 802.1Q

compatible adapter. Select the checkbox for this property. You can then add additional

VLAN IDs one at a time, up to a maximum of 19. This allows a single virtual Ethernet

adapter to access 20 different VLANs.

When you perform this configuration, you are configuring the switch port for this adapter

on the POWER Hypervisor switch to access multiple VLANs.

Once a virtual Ethernet adapter has been successfully added to a running partition

(either by configuring the profile and activating the partition or by DLPAR operation), the

VLAN configuration cannot be changed on POWER 5 and POWER6 servers. To alter

the list of VLAN IDs, you must remove the existing adapter and add a new one. This

can be done dynamically, but it means you must unconfigure the interfaces and remove

the devices from the operating system to complete the operation. Alternatively, you can

add an additional adapter with the proper VLAN ID.

UNIX Software Service Enablement

Adding VLAN access (1 of 3)

 Check IEEE box, enter new VID, and click Add button

 Up to 19 additional VIDs are allowed

 As of POWER7, VIDs can be changed dynamically

 Workaround for POWER5 and POWER6: Use DLPAR to remove the

adapter and re-add with new VIDs if networking to the HMC is provided

by another adapter

Check IEEE box to

be VLAN aware Add new VIDs

one at a time

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-19. Adding VLAN access (2 of 3) QV3441.0

Notes:

Adding the VLAN device

Once you have modified the virtual Ethernet adapter information in the partition profile

properties on the HMC, you should activate the partition. When the partition boots, you

will need to add one VLAN device for each additional VLAN ID that was configured on

the virtual adapter.

The easiest way to add a VLAN device is using the smit command, with the addvlan

fastpath. This will start the dialog shown in the visual above. The menu will allow you to

select the virtual Ethernet adapter that the VLAN device will be associated with, and

also specify the VLAN Tag ID that will be added to packets that are sent out on the

interface configured on the VLAN device. Once you have made appropriate selections,

press the Enter key. This will create the VLAN device (entN), and its associated

interfaces (enN and etN).

UNIX Software Service Enablement

Adding VLAN access (2 of 3)

 Add VLAN devices in AIX

 Use SMIT with addvlan fastpath

 Select base virtual Ethernet adapter

 VLAN priority is optional and used for prioritization of Shared Ethernet

adapter client traffic

Add A VLAN

Type or select values in entry fields.

Press Enter AFTER making all desired changes.

[Entry Fields]

VLAN Base Adapter ent1

* VLAN Tag ID [2] +#

VLAN Priority [] +#

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 5-20. Adding VLAN access (3 of 3) QV3441.0

Notes:

VLAN device

The VLAN device appears as an Ethernet device (ent2 in the visual above), with an

associated pair of interfaces (en2 and et2 in the visual). The interfaces can be

configured just like the ones on a regular physical Ethernet adapter. All packets sent on

the interfaces associated with the VLAN device will be tagged with the VLAN ID.

UNIX Software Service Enablement

Adding VLAN access (3 of 3)

 VLAN device appears as Ethernet adapter (entN) with associated enN

and etNinterfaces

 Configure interfaces as needed for access to VLAN

 Virtual interfaces do not have an AIX location code

# lsdev -C | grep '^e'

en0 Available Standard Ethernet Network Interface

en1 Defined Standard Ethernet Network Interface

en2 Defined Standard Ethernet Network Interface

ent0 Available Logical Host Ethernet Port (lp-hea)

ent1 Available Virtual I/O Ethernet Adapter (l-lan)

ent2 Available VLAN

et0 Defined IEEE 802.3 Ethernet Network Interface

et1 Defined IEEE 802.3 Ethernet Network Interface

et2 Defined IEEE 802.3 Ethernet Network Interface

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-21. Virtual Ethernet switches QV3441.0

Notes:

Multiple Hypervisor virtual Ethernet switches

By default, there is one Hypervisor virtual Ethernet switch. You can create up to a

maximum of 16. Many customers will not need multiple switches, but it could be useful

in a server consolidation project when many hosts are moved to one server and they’re

already configured to use certain VLANs. To keep the communication separate

between partitions which are already configured for the same VIDs, you could put them

on different switches rather than changing the VIDs.

UNIX Software Service Enablement

Virtual Ethernet switches

 When using multiple Ethernet switches with an HMC-managed

server, VLAN IDs on different switches are unique

 That is, VLAN ID 100 on one switch is not the same as VLAN 100 on a

different switch

 Example showing two switches (SW1 and SW2):

 LPAR1 and LPAR5 can communicate with LPARs on SW1 with same VIDs

 LPAR2 and LPAR3 can communicate with LPARs on both switches with same

VIDs

 LPAR4 can communicate only with LPARs on SW2 with same VID

SW1 VID100 SW2 VID 100

SW1 VID 200

LPAR 1 LPAR 2 LPAR 3 LPAR 4 LPAR 5

POWER Hypervisor

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 5-22. Manage virtual switches (1 of 2) QV3441.0

Notes:

Configuring virtual switches (VSwitches)

The Action menu will let you add and remove a vswitch, or change its name.

UNIX Software Service Enablement

Manage virtual switches (1 of 2)

Configuration -> Virtual Resources -> Virtual Network

Management

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-23. Manage virtual switches (2 of 2) QV3441.0

Notes:

Configuring virtual switches

After you create a new vswitch, it will appear as an option whenever you create or

modify virtual Ethernet adapters.

UNIX Software Service Enablement

Manage virtual switches (2 of 2)

Action -> Create VSwitch

 Now, when configuring LPARs virtual Ethernet adapters, you can

choose which switch

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 5-24. View virtual LANs (1 of 2) QV3441.0

Notes:

View virtual network configuration

The Virtual LANs section of the Virtual Network Management screen shows the current

configuration for VLANs on the managed system. If there are VLANs in use, it will list

them in the VLANs view.

Notice the Create SEA ... button in the visual above. This allows you to create a Shared

Ethernet adapter using the HMC. This is covered in the next unit of this course.

The View by: pulldown options are VLANs or partitions. If you choose partitions, then

the chart shows a list of partitions and what VLAN IDs they are using. This is shown on

the next visual.

UNIX Software Service Enablement

View virtual LANs (1 of 2)

Configuration -> Virtual Resources -> Virtual Network

Management

VLANs view shows VIDs in use

 Select a VID to see which partitions are using that VID

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-25. View virtual LANs (2 of 2) QV3441.0

Notes:

View virtual network configuration by partition

The Partitions view shows the VLANs in use by partition name.

UNIX Software Service Enablement

View virtual LANs (2 of 2)

Partitions view shows VLANs in use by partition

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 5-26. Checkpoint (1 of 2) QV3441.0

Notes:

UNIX Software Service Enablement

Checkpoint (1 of 2)

1. Which one of the following is not part of the IEEE 802.1Q VLAN protocol?

A. Ethernet packets are tagged with a VLAN ID

B. Ethernet switch restricts delivery to members of VLAN

C. Ethernet packets are delivered to target host based on IP address first, then

VLAN ID

2. Which one of the following statements about the Hypervisor switch is false?

A. It is the software implementation of an Ethernet switch and is consistent with the

IEEE 802.1Q protocol

B. Adds the PVID tag to packets from VLAN-unaware hosts

C. Configuring the default VLAN ID is optional when creating virtual Ethernet

adapters

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-27. Checkpoint (2 of 2) QV3441.0

Notes:

UNIX Software Service Enablement

Checkpoint (2 of 2)

3. True or False: When using only the default VLAN ID ("PVID") for a virtual

Ethernet adapter, you must also configure the VLAN device in the operating

system.

4. True or False: The Hypervisor Ethernet switch strips the PVID from any packets

that are being delivered to a VLAN-unaware host.

5. What is the maximum number of Hypervisor virtual Ethernet switches that may be

created on a single server? __________

6. True or False: It is possible to set up network communications between partitions

on the same server without any Ethernet hardware in use.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 5-28. Exercise 5 - Virtual Ethernet adapter configuration QV3441.0

Notes:

UNIX Software Service Enablement

Exercise 5  Virtual Ethernet adapter

configuration

 Activity:

 Configure virtual Ethernet adapters

 Configuration:

LPAR1 LPAR2

Hypervisor

ent1 ent1

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 5-29. Unit summary QV3441.0

Notes:

UNIX Software Service Enablement

Unit summary

Virtual Ethernet on POWER5 and higher processor-based systems is

implemented using:

 The POWER Hypervisor as a software Ethernet switch

 Virtual Ethernet adapters in partition configuration

 VLAN IEEE 802.1Q standard

Virtual Ethernet adapters allow network communication between

partitions on the same managed system without using physical

Ethernet adapters

Virtual Ethernet adapters are created and the Hypervisor switch ports

are configured with the HMC, either in a partitions profile or by using

DLPAR

 For one virtual Ethernet adapter to communicate on multiple VLANs,

VLAN devices must also be configured in the operating system

running in the LPAR

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty Unit 6. Shared Ethernet Adapter Configuration

What this unit is about

This unit describes the implementation of the Shared Ethernet adapter

and how to configure client partitions and Virtual I/O Server partitions

to use this type of device. Both command line and HMC GUI

instructions are covered for creating the Shared Ethernet adapter.

What you should be able to do

After completing this unit, you should be able to:

• Configure Shared Ethernet adapters in the Virtual I/O Server

partition using the VIOS command line and the HMC GUI

• Configure a client partition to use a Shared Ethernet adapter

• Configure a Shared Ethernet adapter for threading or

non-threading operation

How you will check your progress

Accountability:

• Checkpoint questions

• Machine exercises

References

IBM Power Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp

Virtual I/O Server product documentation from the IBM Power

Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp?tI

BM Redbooks documents can be found at:



http://www.redbooks.ibm.com

REDP4340 Integrated Virtual Ethernet Adapter Technical Overview

and Introduction, an IBM Redbooks document

SG24-7940 IBM PowerVM Virtualization Introduction and

Configuration, an IBM Redbooks document

SG24-7590 IBM PowerVM Virtualization Managing and Monitoring, an

IBM Redbooks document

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

SG24-6615 A Practical Guide for Resource Monitoring and Control

(RMC), an IBM Redbooks document

REDP4194 IBM System p Advanced POWER Virtualization

(PowerVM) Best Practices, an IBM Redbooks

document

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 6-1. Unit objectives QV3441.0

Notes:

UNIX Software Service Enablement

Unit objectives

After completing this unit, you should be able to:

 Configure Shared Ethernet adapters in the Virtual I/O Server partition

using the VIOS command line and the HMC GUI

 Configure a client partition to use a Shared Ethernet adapter

 Configure a Shared Ethernet adapter for threading or non-threading

operation

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 6-2. Shared Ethernet adapter overview (1 of 2) QV3441.0

Notes:

Shared Ethernet adapter overview

A Shared Ethernet adapter (SEA) is a bridge between a physical Ethernet adapter, or

aggregation of physical adapters, and one or more virtual Ethernet adapters on the

Virtual I/O Server. A SEA enables client partitions on the virtual Ethernet to share

access to the physical Ethernet to communicate with hosts on external networks. The

SEA provides this access by connecting the internal virtual Ethernets with the external

networks.

When there are not enough physical slots to allocate a physical network adapter to

each partition, network access using virtual Ethernet and Shared Ethernet adapters in a

Virtual I/O Server is preferable to IP forwarding as it does not complicate the IP network

topology.

As the Shared Ethernet processes packets at layer 2 of the OSI stack, the originating

MAC address, along with any VLAN tag information, is preserved when the packet is

bridged.

UNIX Software Service Enablement

Shared Ethernet adapter overview (1 of 2)

 Shared Ethernet adapter bridges traffic between internal VLANs and

the external network

 Forwards frames at OSI Layer 2 and is transparent to IP layer

Virtual I/O Server Partition

Layer 2 Bridge (Shared Ethernet adapter)

Device Driver Device Driver Device Driver

Physical

Adapter

Virtual

Adapter

Virtual

Adapter

External

LANs 1-16 virtual adapters

which connect to

VLANs (and client

partitions)

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 6-3. Shared Ethernet adapter overview (2 of 2) QV3441.0

Notes:

Protocol support

Unicast, broadcast, and multicast is supported, so protocols that rely on broadcast or

multicast, such as Address Resolution Protocol (ARP), Dynamic Host Configuration

Protocol (DHCP), Boot Protocol (BOOTP), and Neighbor Discovery Protocol (NDP)

work across a Shared Ethernet adapter.

Benefits

The Shared Ethernet adapter allows client partitions to access a physical network

without dedicated physical Ethernet adapters. The clients share the bandwidth of the

physical adapter on the Virtual I/O Server. Also, while setting up a router to route

packets to external networks is a valid configuration, using bridging to perform this

function allows a simpler network topology, especially when servers are being migrated.

Existing network addresses can be retained with a bridging solution.

UNIX Software Service Enablement

Shared Ethernet adapter overview (2 of 2)

 Shared Ethernet adapter

 Originating virtual Ethernet MAC visible to outside systems

 Broadcast/Multicast, ARP, and NDP support

 Handles physical/virtual MTU mismatch:

 Fragments IPv4 packets, or generate ICMP to aid PMTU discovery

 Generates ICMPv6 for PMTU discovery for IPv6

 Shared Ethernet adapter benefits:

 Efficient use of physical Ethernet adapters when the bandwidth requirements of

client partitions is less than total bandwidth of an adapter

 Maximum quantity of clients limited only by adapter's bandwidth

 Simpler network topology

 Routing is an option to connect LPAR VLANs to external hosts, however

bridging is often the preferred choice because network topology and

addressing do not need to change

A Virtual I/O Server is required in order to configure a

Shared Ethernet adapter

A Virtual I/O Server is required in order to configure a

Shared Ethernet adapter

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 6-4. Configure virtual Ethernet adapter for SEA QV3441.0

Notes:

Access to external Network

If you want the internal network to be connected to the external physical network, then

the virtual Ethernet adapter that is used by the Virtual I/O Server should have the Use

this adapter for Ethernet bridging setting enabled. In HMC versions prior to HMC

V7R730, this checkbox was titled Access external network.

This setting causes the Hypervisor switch to send packets to the adapter when it

doesn't recognize the destination MAC address as internal. This setting should only be

used for the virtual Ethernet adapters that are part of a Shared Ethernet adapter setup

in the Virtual I/O Server. A virtual Ethernet adapter with this setting becomes the virtual

Ethernet trunk adapter for all the VLANs that it belongs to. Since there is one virtual

Ethernet adapter with this setting per VLAN (or two for failover), any overlap of the

VLAN memberships should be avoided between the virtual Ethernet trunk adapters.

UNIX Software Service Enablement

Configure virtual Ethernet adapter for SEA

 Virtual Ethernet adapter configured as part of a SEA in the VIOS should have

Use this adapter for Ethernet bridging attribute checked

 Causes the Hypervisor switch to send packets to this adapter when it doesn't recognize

the destination MAC address as being for an internal virtual Ethernet adapter

 Normally one per VLAN; use priority setting if there are two

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 6-5. Create Shared Ethernet adapter QV3441.0

Notes:

The mkvdev -sea command

The visual above shows a simple Shared Ethernet adapter scenario. The SEA ent2

uses physical adapter ent0 and virtual adapter ent1. The SEA setup requires the

administrator to specify a virtual Ethernet adapter on the VIOS that will be used to

receive any untagged packets from the physical adapter (-default). The PVID is also

specified for this virtual adapter (-defaultid). The PVID value specifies the tag that will

be added to any incoming untagged packets.

List SEA type devices

As of VIOS 2.1, there is a new -type flag for the lsdev command:

$ lsdev -type sea

name status description

ent2 Available Shared Ethernet Adapter

UNIX Software Service Enablement

Create Shared Ethernet adapter

 Command to create Shared Ethernet Adapter:

$ mkvdev sea ent0 vadapter ent1 default ent1 \

defaultid 99

 Creates device (ent2 in this case):

$ lsdev virtual

name status description

ent1 Available Virtual I/O Ethernet Adapter (l-lan)

ent2 Available Shared Ethernet Adapter

Physical Virtual

VLAN ID (PVID) 99

ent0

(Physical)

ent2 (SEA)

ent1 (Virtual)

VIOS

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 6-6. View SEA configuration QV3441.0

Notes:

Using lsmap to view the SEA configuration

The visual above shows the same Shared Ethernet adapter scenario as the last visual.

The output of the lsmap -net -all command in the visual above shows both the

virtual and physical location codes. ent1 in this example is the virtual Ethernet adapter

and it shows that it is in virtual slot 14. The physical adapter is ent0 which is in physical

PCI-X slot C2.

Using lsdev to view the SEA configuration

Another way to view the SEA configuration is with the VIOS CLI command:

$ lsdev -dev ent3 -attr

Fields displayed include the PVID in use, the virtual adapter(s), and the physical

adapter.

UNIX Software Service Enablement

View SEA configuration

 Command to view Shared Ethernet Adapter configuration:

$ lsmap -net -all

SVEA Physloc

------ --------------------------------------------

ent1 U9131.52A.062A87G-V1-C14-T1

SEA ent2

Backing device ent0

Physloc U787F.001.DPM166X-P1-C2-T1

PVID 99

ent0

(Physical)

ent2 (SEA)

ent1 (Virtual)

VIOS

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 6-7. SEA with multiple virtual adapters QV3441.0

Notes:

Multiple VLAN example

In the example in the visual above, the SEA (ent3) uses the physical adapter ent0 and

the virtual adapters ent1 and ent2. The physical adapter ent0 is connected to a

physical Ethernet switch which is VLAN capable and the port should be set up as a

tagged port with VID 200 and VID 300.

Both virtual Ethernet adapters ent1 and ent2 have the trunk attribute (Access external

network checkbox) using the HMC. Note, this configuration can also be achieved with

one virtual Ethernet adapter by creating an additional VLAN ID of 200 for the ent1

virtual adapter.

The Shared Ethernet adapter can only provide default VID connectivity for incoming

untagged packets to one of the virtual adapters and VLAN 99 was chosen in this

example. The PVID of 99 in this example is inconsequential only because a virtual

adapter cannot be created without a PVID and any other unused PVID value could

have been used here.

UNIX Software Service Enablement

 Command to create Shared Ethernet Adapter:

$ mkvdev sea ent0 vadapter ent1 ent2 \

default ent1 defaultid 99

 Device listing:

$ lsdev virtual

name status description

ent1 Available Virtual I/O Ethernet Adapter (l-lan)

ent2 Available Virtual I/O Ethernet Adapter (l-lan)

ent3 Available Shared Ethernet Adapter

SEA with multiple virtual adapters

ent3 (SEA)

ent2 (Virtual)

ent1 (Virtual)

ent0

(Physical)

VID 100, PVID 99

VID 200, PVID 98

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 6-8. SEA with multiple physical adapters QV3441.0

Notes:

Link aggregation

Link aggregation is a network port aggregation technology that allows several Ethernet

adapters to be aggregated together to form a single pseudo Ethernet device. For

example, ent0 and ent1 can be aggregated to ent3; interface en3 would then be

configured with an IP address, or ent3 can be used as the physical adapter for an SEA.

The system considers these aggregated adapters as one adapter.

The main benefit of link aggregation is the ability to use the network bandwidth of all of

the adapters. Also, if an adapter fails, the packets are automatically sent on the next

available adapter without disruption to existing user connections. The adapter is

automatically returned to service on the link aggregation when it recovers.

UNIX Software Service Enablement

SEA with multiple physical adapters

 Create Ethernet link aggregation device:

$ mkvdev lnagg ent0 ent1

ent3 shows up as:

$ lsdev adapter | grep ent3

ent3 Available EtherChannel / IEEE 802.3ad Link Aggregation

 Create Shared Ethernet adapter:

$ mkvdev sea ent3 vadapter ent2 default ent2 defaultid 99

ent2 (Virtual)

PVID 99

(Physical)

ent4 (SEA)

ent0

ent1

ent3

(Aggregate)

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 6-9. Create SEA with cfgassist tool QV3441.0

Notes:

Create SEA with cfgassist tool

Shared Ethernet adapters can be created with the cfgassist tool. You can specify

multiple virtual Ethernet adapters and use F4 (or ESC 4) to select devices.

UNIX Software Service Enablement

Create SEA with cfgassist tool

 The cfgassist tool can be used to create a SEA

Devices ÆShared Ethernet Adapter ÆAdd a Shared Ethernet Adapter

Add a Shared Ethernet Adapter

Type or select values in entry fields.

Press Enter AFTER making all desired changes.

[Entry Fields]

* Physical Ethernet Adapter [ent0] +

* Virtual Ethernet Adapters [ent1] +

* Default Virtual Ethernet Adapter [ent1] +

* Default ID [1] #

ATTRIBUTES []

Migrate the settings from physical interface to no +

the SEA interface?

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 6-10. Configuring the VIOS interface (1 of 2) QV3441.0

Notes:

Configuring the TCP/IP parameters with a Shared Ethernet adapter

The visual above illustrates the different devices that can exist when implementing a

single Shared Ethernet adapter. In this example, there are two interfaces on which the

TCP/IP options can be configured and there is no performance gain for either

configuration.

The first option is to configure the TCP/IP parameters such as the IP address on the

interface associated with the Shared Ethernet adapter (en3 in this example).

The second option is to configure a second, optional, virtual Ethernet adapter and

configure the TCP/IP parameters on its interface (en2 in this example).

You cannot configure the interface for the actual physical adapter associated with a

Shared Ethernet adapter (en0 in this example). You also cannot configure the interface

associated with the Shared Ethernet adapter’s virtual Ethernet adapter (en1 in this

example).

UNIX Software Service Enablement

Configuring the VIOS interface (1 of 2)

Virtual

I/O

Server

External

LANs

Virtual ent1 Virtual ent2

Hypervisor

Virtual Ethernet

traffic

Physical ent0

Shared Ethernet

adapter ent3

 Ethernet interface configuration options for the Virtual I/O Server:

 Configure IP address on interface associated with the Shared Ethernet adapter (en3 in

this example)

 Configure IP address on optional second virtual Ethernet adapter (configure the en2

interface this example)

If you have the SEA

interface configured, it

must be unconfigured to

alter the SEA

configuration.

Configuration on second

virtual Ethernet makes it

easier to change the

SEA configuration later

without affecting VIOS

traffic with other LPARs

on the same server.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 6-11. Configuring the VIOS interface (2 of 2) QV3441.0

Notes:

Migrating the interface configuration

If the physical adapter’s interface is already configured, you can migrate the settings to

the SEA’s interface when the SEA is created by adding the -migrate -auto flags.

Reverse the operation

Once the SEA is created and configured, if you want to remove it and restore the IP

configuration to the physical adapter, you can use the procedure on the visual above.

UNIX Software Service Enablement

Configuring the VIOS interface (2 of 2)

 The configuration on the physical adapter's interface can be

migrated automatically to the SEA device

 Add the



migrate



auto flags to the mkvdev



sea command

 This example creates an SEA and migrates the current configuration on en0 to

the interface on the resulting SEA device:

$ mkvdev



sea ent0



vadapter ent1



default ent1 \



defaultid 10



migrate



auto

 Example procedure to reverse the result of the above mkvdev

command:

 Detach the SEA interface (en2):

$ chdev dev en2 attr state=detach

 Remove the SEA (ent2) and its interfaces:

$ rmdev dev ent2; rmdev dev en2; rmdev dev et2

 Return the configuration on the physical adapter's interface (en0):

$ chdev dev en0 attr state=up restoreroute

Log into the

console to perform

this procedure

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 6-12. Using HEA logical ports for SEA QV3441.0

Notes:

Host Ethernet Adapter (HEA)

The HEA is also known as the Integrated Virtual Ethernet adapter and is available on

most POWER6 and some POWER7 processor-based servers. The HEA has 2 or 4

physical ports, depending on the model purchased. Multiple logical ports can share one

physical port providing direct network connectivity for LPARs. However, if the VIOS

partition uses an HEA logical port for an SEA configuration, the VIOS will use the entire

physical port so no other LPARs can be configured for the same physical port. Then,

you must list the VIOS partition as the promiscuous LPAR on the HEA physical port.

If LPARs are using HEA ports and you shut the partitions down, the system will still

consider the logical ports to be owned by the Not Activated LPARs. To clear this

situation, you can either reactivate the partitions without HEA ports or use this HMC

command where LPAR1 is the partition with the port and the managed system is MSname:

rsthwres -r hea -m MSname -p LPAR1

HEA ports can be configured in a link aggregation only with other HEA ports.

UNIX Software Service Enablement

Using HEA logical ports for SEA

 A Host Ethernet Adapter (HEA) logical port can be used as the

physical adapter for an SEA

 Requirements:

 No other LPARs can share the same HEA physical port

 The physical port must have the VIOS partition as its promiscuous LPAR before

SEA is configured

 Run Hardware Information -> Adapters -> Host Ethernet task

 Select Physical Port and click Configure button

Choose a physical port, then click the

Configure button to configure VIOS as

promiscuous LPAR

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 6-13. Troubleshooting SEA "cfgsea"error QV3441.0

Notes:

Cause 2: From the VIOS CLI, before you make the SEA, you can check to see if the

virtual Ethernet adapter has the Access external network checkbox checked with this

command. If it is, the output will show: Trunk Adapter: True

$ entstat -all ent# | grep Trunk

Cause 3: To check the checksum offload attribute on the virtual and physical adapters

use the following command substituting ent# for the proper device name:

Virtual and Physical: $ lsdev -dev ent# -attr chksum_offload

If the lsdev command output is blank, then the device does not support the checksum

attribute. To change the attribute, use chdev. Set the value to yes to enable and no to

disable. This example command sets the chksum_offload attribute to enabled for the

ent0 device:

$ chdev -dev ent# -attr chksum_offload=yes

This attribute is set to yes by default for virtual adapters. For physical adapters, it

depends on the type of adapter.

UNIX Software Service Enablement

Troubleshooting SEA "cfgsea" error

 Typical command:

$ mkvdev -sea ent0 -vadapter ent2 -default ent2 -defaultid 10

 Example configuration error:

Method error (/usr/lib/methods/cfgsea):

0514-040 Error initializing a device into the kernel.

 Check the following causes:

1. Is the physical or virtual interface already configured?

2. Does the virtual adapter have "Access external network" and "Trunk priority"

configured?

3. Is the checksum offload attribute mismatched between physical and virtual

devices?

4. If the physical device is an HEA logical port, is the VIOS listed as the

promiscuous LPAR for the physical port?

5. Is there enough free memory?

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 6-14. Shared Ethernet adapter with VLANs QV3441.0

Notes:

LPAR access to external VLANs

In the visual above, LPARs 1, 2, and 3 can participate on the VLANs that extend past

their managed system. The Hypervisor strips the VLAN ID tags from packets when

routing them out through the SEA if the VLAN ID matches the PVID of the virtual

Ethernet adapter on the VIOS which is in the SEA configuration. The VIOS must have

the checkbox labeled IEEE 802.1Q compatible adapter checked and additional VLAN

IDs that will be used for external VLANs listed.

UNIX Software Service Enablement

Shared Ethernet adapter with VLANs

 VLAN tags will be stripped from exiting packets when the VLAN ID

matches the PVID of the VIOS virtual Ethernet adapter configured

with the SEA

 Other VLAN ID tags will remain intact

 Provides support for VLAN-unaware networks

 In this example, the SEA will forward packets destined to the external

VLAN 100 and 200 with VLAN tags intact:

PVID=1

VLANs=

100,200

VLAN 100

VLAN 200

VIOS

Managed system

Ethernet Switch

External Hosts

LPAR1 LPAR2 LPAR3

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 6-15. Configure SEA using the HMC (1 of 2) QV3441.0

Notes:

Requirements

The following requirements need to be met to configure an SEA with the HMC:

- HMC V7 R3.4.2 or later

- VIOS has a physical network device or LHEA logical port that is already configured

with an IP address

- Virtual Ethernet adapter already exists on the VIOS and configured to access

external network

- Virtual I/O Server version 2.1.1.0 or later

- RMC connection between the HMC and the VIOS

• To check the RMC connection, execute the following command on the HMC:

# lspartition -all

UNIX Software Service Enablement

Configure SEA using the HMC (1 of 2)

 Run the Configuration ÆVirtual Resources ÆVirtual Network

Management task

 Select the VLAN ID to bridge to access the Create SEA  button

Select

one of the

VLAN IDs

In this example, the VIOS

virtual adapter has a

PVID of 1, and additional

VLAN IDs of 10 and 20

In this example, the VIOS

virtual adapter has a

PVID of 1, and additional

VLAN IDs of 10 and 20

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 6-16. Configure SEA using the HMC (2 of 2) QV3441.0

Notes:

Create SEA ... button

The visual above shows the result of selecting a VLAN ID and clicking the Create SEA

... button. If there are multiple VIOS type partitions, choose the VIOS by name. If there

are multiple physical adapters, choose which adapter to configure. You can also

configure SEA failover from this task.

UNIX Software Service Enablement

Configure SEA using the HMC (2 of 2)

 The HMC:

 Configures the SEA device for all VLAN IDs included on same virtual Ethernet

adapter

 Moves the TCP/IP configuration from the physical adapter to the SEA device (if

necessary)

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 6-17. View SEA configuration from HMC QV3441.0

Notes:

View SEA configuration from the SEA

There are two “Views” available on the Virtual Network Management screen: VLANs

and Partitions. The visual above shows both and shows the SEA configuration

information available in each.

Notice that in the Partitions view, the three VLANs (1, 10, 20) are all marked with an

asterisk showing that all three are bridged. In the example from the last two pages, only

VLAN 10 was selected when the SEA was created. Because the virtual Ethernet

adapter that was configured for VLAN 10 was also configured for VLAN 1 and VLAN 20,

all three VLANs are bridged.

UNIX Software Service Enablement

View SEA configuration from HMC

 Once an SEA is configured, it is visible in VLANs view

 In the Partitions view, the

VLAN ID will have an *

to indicate it's bridged

Notice all 3

VLAN IDs are

bridged

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 6-18. Removing the SEA using the HMC QV3441.0

Notes:

Remove SEA ... button

If the Remove SEA ... button is selected, it will give you instructions using the proper

device names to remove the SEA. If failover mode was configured, you can choose the

appropriate VIOS partition by name. You must manually perform these steps to remove

the SEA. If the SEA interface was configured with the IP configuration, then the last

chdev command (as shown in the above visual) will restore the IP configuration on the

physical adapter’s interface.

UNIX Software Service Enablement

Removing the SEA using the HMC

 The Remove SEA  button gives instructions:

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 6-19. Threading/non-threading (1 of 2) QV3441.0

Notes:

Threaded or non-threaded SEA device driver

The threaded model ensures that the SEA driver does not monopolize CPU resources

to the detriment of virtual SCSI and NPIV services. However, threading adds more

instruction path length, thus uses more CPU cycles. If the VIOS will only be providing

SEA services (and not virtual SCSI or NPIV services) then the SEA should be

configured with threading disabled in order to run in the most efficient mode.

Threading causes incoming packets to be queued to a buffer in memory. A special

kernel thread is dispatched to process the buffer, which uses more processing power

and allows the processor to be shared more evenly with virtual SCSI and NPIV. With

non-threading, the virtual Ethernet and SEA driver forward packets at the interrupt level

which is more efficient and is why throughput goes up.

The SEA threading attribute determines the implementation that will be used by the

SEA device driver, and is not related to simultaneous multi-threading.

UNIX Software Service Enablement

Threading/non-threading (1 of 2)

 Threading is a feature that ensures that CPU resources are shared

fairly when a Virtual I/O Server hosts a SEA in addition to hosting

virtual SCSI or virtual Fibre Channel (NPIV) connections

 Special kernel thread is used by SEA device driver for queuing SEA requests

 Like most device drivers, virtual Ethernet and Shared Ethernet drivers typically drive high

interrupt rates and are CPU intensive

 Without threading, network traffic will have a higher priority than virtual SCSI or

virtual Fibre Channel interrupts

 On a CPU constrained system, this will result in worse virtual SCSI and virtual

Fibre Channel performance

 With threading enabled, there is more consistent quality of service when CPU is

constrained, but at a lower overall network throughput and higher overall CPU

consumption (all the time)

 Threading is enabled by default for Shared Ethernet adapter:

 Disable threading when a Virtual I/O Server is not used for virtual SCSI or NPIV:

 Can use separate Virtual I/O Servers for Shared Ethernet adapter and virtual

SCSI/NPIV services

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 6-20. Threading/non-threading (2 of 2) QV3441.0

Notes:

Threaded versus non-threaded performance

The threading option of the Shared Ethernet adapter may result in performance

degradation for network traffic if CPU resources are constrained. The degradation for a

streaming workload is around 16-20% when the MTU size is 1500, and around 31-39%

when the MTU size is 9000. The threading option causes more performance

degradation at lower workloads due to individual threads being started for each packet.

At higher workload rates, like full duplex or the request/response workloads, the threads

can run longer without waiting and being re-dispatched.

Enabling and disabling threading

The thread option for the Shared Ethernet adapter can be enabled (the default) or

disabled on a per-adapter basis using the chdev command in the Virtual I/O Server

command line interface (CLI).

UNIX Software Service Enablement

Threading/non-threading (2 of 2)

 Threading increases overall CPU cycle usage:

 Performance gains are 16-39% with it disabled, depending on MTU size

 Threading can be enabled/disabled on a per Shared Ethernet

adapter basis on the Virtual I/O Server:

 See the current status:

$ lsdev dev ent3 attr



thread 1 Thread mode enabled (1) or disabled (0)



 To disable:

$ chdev dev ent3 attr thread=0

SEA device name

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 6-21. Monitoring SEA traffic QV3441.0

Notes:

Monitoring network traffic when using an SEA

If you want to see only the VIOS traffic, then use a tool that shows the configured

interfaces such as the topas or netstat command.

When using an SEA however, often you wish to see the total traffic for all clients coming

in and going out of the physical adapter. The entstat command can be used in this

case.

If you wish to monitor each SEA client’s traffic individually, perhaps to see which clients

have the most traffic, then the seastat command is the one to use. You must enable

advanced accounting which can potentially incur a small performance impact so you

can disable it after monitoring.

When viewing seastat output, the first interface on each client will be listed with its

hostname. Subsequent interfaces will be listed by MAC address only. The seastat

command is available in VIOS V1.5 and up.

UNIX Software Service Enablement

Monitoring SEA traffic

 Some network monitoring tools monitor the interface: netstat,

topas

 Others monitor the adapter: entstat, seastat

seastat provides statistics per MAC address of all SEA clients

 Must enable advanced accounting on the SEA first:

$ chdev -dev ent# -attr accounting=enabled

 To view reports, use -d followed by SEA device name:

$ seastat -d ent#

 Example entry from seastat output:

==========================================================

MAC: 7E:DF:45:0E:31:04

----------------------

VLAN: 10

VLAN Priority: None

IP: 9.47.87.82

Transmit Statistics: Receive Statistics:

-------------------- -------------------

Packets: 74 Packets: 112

Bytes: 14648 Bytes: 18425

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 6-22. Checkpoint QV3441.0

Notes:

UNIX Software Service Enablement

Checkpoint

1. True or False: To use Shared Ethernet adapters, you must have a Virtual I/O

Server partition.

2. True or False: The virtual adapters configured as part of a Shared Ethernet

adapter must have the Access external network option enabled.

3. True or False: Only one virtual Ethernet adapter per VLAN can be configured as

the SEA virtual adapter.

4. True or False: Shared Ethernet adapter threading is important on Virtual I/O

Servers that provide both virtual storage and Shared Ethernet adapter services.

5. Which of the following cannot be done from the HMC Virtual Network

Management task:

A. Create an SEA to connect multiple VLANs to an external network

B. Remove an SEA

C. When creating the SEA, move the TCP/IP configuration of the physical Ethernet

adapter to the SEA device

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 6-23. Exercise 6 - Shared Ethernet adapter QV3441.0

Notes:

UNIX Software Service Enablement

Exercise 6 - Shared Ethernet adapter

 Activity:

 Configure a Shared Ethernet adapter

 Configuration

 Three partitions:

 Two AIX partitions: LPAR2

and LPAR3

 One Virtual I/O Server: VIOS

 Physical Ethernet adapter is

an HEA logical port

LPAR1 VIOS LPAR2

Physical

Ethernet

Adapter

Hypervisor

SEA

Virtual

Physical Ethernet

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 6-24. Unit summary QV3441.0

Notes:

UNIX Software Service Enablement

Unit summary

Shared Ethernet adapter uses a physical Ethernet adapter of the

Virtual I/O Server to connect client partitions with virtual Ethernet

adapters to external networks

 Shared Ethernet adapters can only be created in the Virtual I/O

Server partition

 The virtual Ethernet adapters used in Shared Ethernet adapter

configurations must have the Access external network attribute

 The Virtual Network Management HMC task can be used to create a

Shared Ethernet adapter

 If a Virtual I/O Server provides Shared Ethernet adapter services but

not virtual SCSI services, then the threading attribute can be disabled

for better Shared Ethernet adapter performance

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty Unit 7. Virtual I/O Server Maintenance

What this unit is about

This unit describes common maintenance operations for Virtual I/O

Servers including backup and restore and updating the software level.

What you should be able to do

After completing this unit, you should be able to:

• Describe Virtual I/O Server backup tools

• Use viosbr to backup and restore virtual device mapping

• Describe the use of backupios and savevgstruct

• Describe how to update the VIOS software

How you will check your progress

Accountability:

• Checkpoint questions

• Machine exercises

References

IBM Power Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp

Virtual I/O Server product documentation from the IBM Power

Systems Hardware Information Center:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp?t

IBM Redbooks documents can be found at:

http://www.redbooks.ibm.com

REDP4340 Integrated Virtual Ethernet Adapter Technical Overview

and Introduction, an IBM Redbooks document

SG24-7940 IBM PowerVM Virtualization Introduction and

Configuration, an IBM Redbooks document

SG24-7590 IBM PowerVM Virtualization Managing and Monitoring, an

IBM Redbooks document

SG24-7430 AIX 6 Advanced Security Features Introduction and

Configuration, an IBM Redbooks document

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 7-1. Unit objectives QV3441.0

Notes:

UNIX Software Service Enablement

Unit objectives

After completing this unit, you should be able to:

 Describe Virtual I/O Server backup tools

Use viosbr to backup and restore virtual device mapping

 Describe the use of backupios and savevgstruct

 Describe how to update the VIOS software

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 7-2. Maintaining the Virtual I/O Server QV3441.0

Notes:

When should you backup the VIOS?

To ensure that your backup of the VIOS accurately reflects your current running VIOS,

you should backup the VIOS each time that its configuration changes. The term

"backup" in this context means documenting anything that might have changed. For

example, you may need to backup the VIOS image after installing a fix pack, or you

may need to backup the virtual device configuration information after changing virtual

device mappings.

UNIX Software Service Enablement

Maintaining the Virtual I/O Server

 VIOS back up strategy considerations:

 What is your strategy for VIOS down time (planned or unplanned)?

 Do you have multiple VIOS partitions? If not, schedule client down time

for VIOS updates or restores

 How often does your virtual I/O configuration change?

 Do you keep:

 An updated copy of the LPAR configuration (profile data)?

 A current back up of your virtual device mappings?

 A documented change management procedure, updated system

configuration documents, and a test plan?

Best practice: Automatically back up your virtual I/O configuration

and have a plan for keeping the VIOS software updated

Best practice: Automatically back up your virtual I/O configuration

and have a plan for keeping the VIOS software updated

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 7-3. VIOS backup options QV3441.0

Notes:

Backing up profile data

Be sure to backup the LPAR profile information and document it as well, including all

virtual adapter configuration information such as VLAN IDs, adapter IDs, and SEA

failover control channel configuration. The HMC system plans or the lshwres command

are useful to create documentation.

Backing up disk structure data

The savevgstruct command backs up user-defined disk structures. This command

writes a backup of the structure of a named volume group (and therefore storage pool)

to the /home/ios/vgbackups directory. The savevgstruct command is automatically

called before the backup for all active non-rootvg volume groups or storage pools on a

VIOS when the backupios command is run. Because this command is called before the

backup starts, the volume group structures will be included in the system backup.

UNIX Software Service Enablement

VIOS backup options

 Backing up the entire VIOS configuration involves backing up:

 LPAR profile data (perform from the HMC)

 Virtual device mappings on the Virtual I/O Server (manually, viosbr, or

backupios)

 VIOS operating system image (backupios)

 Structure data for non-rootvg volume groups (backupios and savevgstruct)

 Client data is backed up from the client

Use viosbr (as of VIOS V2.1.2.0) to:

 Back up user-defined virtual device configuration

 Restore the configuration to the same VIOS partition

Use backupios to:

 Back up the VIOS operating system image (rootvg)

 Restore to same or different VIOS partition

 Save disk structures (runs savevgstruct)

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 7-4. Using backupios for backups QV3441.0

Notes:

The backupios command

The backupios command is available from the VIOS CLI. It can be used to backup the

VIOS rootvg volume group to create an installable image. The backup can be used to

reinstall the VIOS if the running image becomes corrupt or otherwise unusable.

Besides the -file and -cd flags, you can also specify a -tape flag followed by a tape

device name to create a bootable tape.

If you want backupios to create only the mksysb file, use the -mksysb flag with the

-file flag which specifies a filename, not a directory. Note that the mksysb backup of

the VIOS can be extracted from the tar file created in a full backup, so either method is

appropriate if the restoration method uses a NIM server.

By default, backupios will also save the structure information for any user-defined

volume groups, unless you specify the -nosvg flag.

UNIX Software Service Enablement

Using backupios for backups

backupios can backup VIOS rootvg to a DVD, tape, or file system:

 Can create nim_resources.tar (used for HMC install)

 Can create a mksysb (with the mksysb flag)

 For NFS mounted file system, allow root access and make sure hostname

resolution works both ways

 The nomedialib flag will exclude the Virtual Media Repository from the

backup

 Example that creates bootable media:

$ backupios -cd /dev/cd0 -udf -accept

 Example for a mounted NFS file system to create all resources NIM

needs to restore with installios [or create a mksysb]

$ backupios -file /mnt [-mksysb]

 Add the nomedialib flag to exclude the optical media repository if

it is stored in rootvg

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 7-5. Manual backup of VIOS device mappings QV3441.0

Notes:

Backing up additional information

See the PowerVM virtualization on IBM System p Managing and Monitoring Redbooks

document to see a table of commands you can execute and the information they

provide. This chapter of the Redbooks document also gives information about using

Tivoli Storage Manager to perform backups.

For complex configurations, don’t forget things such as SEA control channels (for SEA

failover), IP addresses to ping, MTU settings, whether threading is enabled for the SEA

devices, and other tuning changes that may have been made.

You can write a script that runs the commands shown on the visual above, then add that

script to the VIOS crontab file so that it runs regularly. For example, to run a script

called /home/padmin/backup every Saturday at 2:00 AM, run the crontab -e

command first to put you in edit mode. Then, type 0 2 * * 6 /home/padmin/backup

and save and exit the editor. The crontab -l command will list crontab entries. The

crontab -r command will delete crontab entries.

UNIX Software Service Enablement

Manual backup of VIOS device mappings

 Additional VIOS information can be saved in /home/padmin for back up by

backupios

 Capture lsmap all and lsmap net -all output to document the

virtual device configuration

 Example commands to capture other configuration information:

 Network settings

lstcpip -stored, netstat -state, netstat -routinfo,

netstat -routtable, lsdev -dev entX -attr,

optimizenet -list, and entstat -all entX

 All physical and logical volume SCSI devices

lspv, lsvg, and lsvg -lv VolumeGroup,

lsdev dev devicename -attr

 All physical and logical adapters

lsdev -type adapter

 Code levels, users and security

ioslevel, lssw, viosecure -firewall -view,

viosecure -view -nonint, motd, loginmsg, and lsuser

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 7-6. VIOS manual backup examples QV3441.0

Notes:

Backing up additional information

Document information about network settings, adapters, users, and security settings to

the /home/padmin directory by running each command in conjunction with the tee

command as follows:

command | tee /home/padmin/filename

Where:

- command is the command that produces the information you want to save

- filename is the name of the file to which you want to save the information

The /home/padmin directory is backed up automatically when using the backupios

command; therefore, it is a good location to collect configuration information prior to a

backup.

The viosvrcmd is an HMC CLI command to run VIOS commands. You can create a

remote HMC shell script that runs the backup commands on a regular basis.

UNIX Software Service Enablement

VIOS manual backup examples

 Run commands on VIOS CLI using:

$ command | tee backups/filename

 Commands can be run from a remote system

 Perhaps as part of a script run by cron

 Commands can be run from HMC CLI:

hscroot@hmc:~> viosvrcmd -m MSNAME -p VIOLPAR -c "VIO_COMMAND"

 Commands can be run from the HMC or from an HMC script:

# ssh hscroot@hmc ' viosvrcmd m MSNAME p VIOLPAR -c

"backupios -file /tmp/bkfile" '

Use real managed

server name

Use real VIOS LPAR name

or use --id with LPAR ID

"Use real VIOS CLI

command"

Notice use of single

and double quotes

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 7-7. Restore options with backupios . QV3441.0

Notes:

Restoring virtual device mappings

If you are restoring to new hardware, virtual devices cannot be automatically recovered

because the disk structures will not exist.

If the physical devices exist in the same location and structures such as logical volumes

are intact, the virtual devices such as virtual target SCSI and Shared Ethernet Adapters

will be recovered during the restoration.

UNIX Software Service Enablement

Restore options with backupios

 Restore of VIOS is like an install

 Use the mksysb or tar file from backup

 Procedure for total restore:

1. Restore HMC LPAR configurations

2. Restore VIOS operating system from backup

 If restoring to same LPAR, all virtual device configurations will be

restored

 If restoring to a different LPAR (location codes are different) you will

need to manually recreate virtual devices (mkvdev)

 If you need to restore the non-rootvg volume group structures on the

same LPAR, use restorevgstruct

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 7-8. Backup and restore with viosbr . QV3441.0

Notes:

The viosbr command

This command is new as of VIOS V2.1.2.0. Use it to backup or restore VIOS virtual

device attributes or to view the contents of a viosbr backup file.

Information backed up includes: device attributes for disks, optical devices, tape

devices, Ethernet adapters, Ethernet interfaces, logical HEAs and logical devices like

storage pools, file-backed storage pools, optical repositories, virtual block storage

devices (vbsd), virtual devices like Etherchannel, SEA, virtual server adapter, and

virtual server fibre-channel adapters.

The restore function uses the information in the backup file to restore the VIOS partition

to the same state as when the backup was taken. The restore operation sets the

attribute values for physical devices, imports logical devices and creates virtual devices

and corresponding mappings.

UNIX Software Service Enablement

Backup and restore with viosbr

Use viosbr command to backup virtual device data to recover a

VIOS after a re-installation or to the same installation

 Restore must be run on the same VIOS partition where the backup was done

viosbr is used to back up and restore devices and their mappings

 Includes device attributes, logical and virtual devices information

 Example definitions that are saved: storage pools, file-backed storage pools, the

virtual media repository, AMS paging space devices, SEA, VTDs, and link

aggregation devices

 The restore function restores the VIOS to the same state as when the

backup was taken

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 7-9. viosbr command examples QV3441.0

Notes:

Using viosbr

Backup files are located by default in /home/padmin/cfgbackups unless another

directory is specified. All the configuration information is saved in a compressed XML

file.

When multiple backups are created through the use of the -frequency flag, the backup

files are named file.xx.tar.gz where file is the argument to the -file flag and xx is an

incremental number from 01 to 10. For example, the result of the daily backup

command shown in the visual above are these filenames: mybackup.01.tar.gz,

mybackup.02.tar.gz, mybackup.03.tar.gz, mybackup.04.tar.gz, and

mybackup.05.tar.gz for the five most recent files. After reaching the given number of

files, the oldest backup file is deleted during the next backup cycle.

Daily backups occur at 00:00, weekly backups on Sunday at 00:00, and monthly

backups on the first day of the month at 00:01.

UNIX Software Service Enablement

viosbr command examples

 Backup to a file:

 Creates a gzip tar file

$ viosbr -backup -file /home/padmin/backups/mybackup

 Backup daily on the VIOS and keep the last five backup files:

Use daily, weekly, or monthly for frequency

$ viosbr -backup -file mybackup -frequency daily numfiles 5

 View contents:

$ viosbr -view -file mybackup.012909.tar.gz

 Restore:

$ viosbr -restore -file mybackup.012909.tar.gz

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 7-10. Updating the Virtual I/O Server (1 of 2) QV3441.0

Notes:

Updating the VIOS

You can use the updateios VIOS CLI command with the -dev flag to install a fixpack.

For example, the following command updates the VIOS product using the updates it

finds in the specified directory (/home/padmin/update in the example below):

$ updateios -f -dev /home/padmin/update -install -accept

UNIX Software Service Enablement

Updating the Virtual I/O Server (1 of 2)

Use updateios from the VIOS CLI or cfgassist to install one

or more filesets or install a fixpack:

Syntax:

updateios -dev Media [-f] [ -install ] [ -accept ]

updateios -commit | -reject [ -f ]

updateios -cleanup

updateios -remove { -file RemoveListFile | RemoveList }

Uses installp to update the filesets, but may install new base filesets

 Will install or update all filesets in the directory or on the device indicated with

dev (there is no way to specify a list of filesets)

 For example, if your update fileset is located in /home/padmin/update directory:

$ updateios install accept dev /home/padmin/update

 Unload all Optical Media Repository images before updating

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 7-11. Updating the Virtual I/O Server (2 of 2) QV3441.0

Notes:

Preparing for a VIOS update

Before updating the Virtual I/O Server it is recommended that you take the steps listed

in the visual.

UNIX Software Service Enablement

Updating the Virtual I/O Server (2 of 2)

 Also possible to update with a cust operation from a NIM master:

 Create an lpp_source with the updates

 Before updating:

 Perform a backup if a current backup is not available

 For a single VIOS partition, shut down its virtual I/O clients

 Unload all images in the virtual media repository

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 7-12. View installed levels QV3441.0

Notes:

Viewing the installed code level

The lssw command will list installed filesets, their versions, and their status. It gives the

same output as if you ran the lslpp -l AIX command and then the emgr -l

command.

The ioslevel command lists the installed Virtual I/O Server product version.

The oem_platform_level command only shows the underlying base AIX version such

as shown above and not the actual underlying AIX level.

UNIX Software Service Enablement

View installed levels

Use lssw command to view all installed filesets and fixes

 Similar to AIX lslpp l and emgr l commands

Use ioslevel to see current VIOS level

$ ioslevel

2.2.1.3

Use oem_platform_env to view base AIX level

$ oem_platform_level

AIX Version:

6.1.0.0

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 7-13. Sizing the Virtual I/O Server (1 of 2) QV3441.0

Notes:

Sizing the VIOS

Using virtual I/O devices uses more overall system processing resources than if

partitions use native devices because the VIOS partition itself requires processor

resources. It is important to monitor processor and memory utilization on the VIOS

partition since if it runs out of resources, all clients using virtual devices will be affected.

References

See the VIOS sizing document at:

http://publib.boulder.ibm.com/infocenter/powersys/v3r1m5/index.jsp?topic=/iphb1/iphb1

_vios_planning.htm

In the IBM Systems Hardware Information Center, search for “Virtual I/O Server

memory allocation” for more memory sizing information. Search for “Virtual SCSI sizing”

for more information on sizing processor resources for virtual SCSI.

UNIX Software Service Enablement

Sizing the Virtual I/O Server (1 of 2)

 The memory and CPU needs on the virtual client when using virtual

I/O are not higher than when using physical (native) I/O

 That is, size client as if it has all dedicated, physical devices

 Memory sizing for Virtual I/O Server is typically 1 GB

 Use 4 GB if using Shared Storage Pools

 Needs are minimal because no disk data caching is done on the Virtual I/O

Server for clients

 Possibly more memory is needed for servers with large amount of disks

(hundreds) or 10 Gbps Ethernet adapters

 Partitions with IVE ports should configure 103 MB more memory per logical port

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 7-14. Sizing the Virtual I/O Server (2 of 2) QV3441.0

Notes:

Processing sizing for VSCSI

Factors for processor sizing are the number of disks, the I/O operation block sizes, and

the expected workload (or the maximum I/O throughput for the disks). The IBM

Systems Hardware Information Center contains tables to help calculate the processing

resources needed. Search for “Virtual SCSI Sizing Considerations”.

Processor sizing for SEA

The IBM Systems Hardware Information Center contains tables of CPU utilization

percentages based on the various SEA factors which affect CPU use. Search for

“processor allocation.” For example: A simplex, streaming workload at 1500 MTU with

threading enabled will use a 80.6% of a POWER5 1.65 GHz processor core. All four

italicized words in that example are workload or configuration factors which will affect

the amount of processing resources used.

UNIX Software Service Enablement

Sizing the Virtual I/O Server (2 of 2)

 Processor sizing

 Dedicated versus shared processors:

 Servers with dedicated processors will see slightly better device

performance due to virtual processor latency with shared processors

 For shared processors, configure as uncapped with a weight value larger

than clients

 When calculating exact CPU requirements for disk access, plan for this amount

on both the VIOS and the client

 For example, if CLIENT1 sizing shows it needs 2.5 processing units for all

of its disk I/O needs, if these are backed by a VIOS plan for 2.5 also in

the VIOS partition

 CPU sizing for storage services (VSCSI and virtual Fibre Channel) depends on

the number and type of disks and the workload

 CPU sizing for SEA services depends on the type of workload and the

throughput needed

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 7-15. Monitoring Virtual I/O Server resources QV3441.0

Notes:

Monitoring resources on the VIOS

The topas command can monitor all of the key resource areas: CPU, memory, disk I/O,

and network I/O. In addition, for CPU and I/O usage statistics, you can use the viostat

command which is like the iostat AIX command. Use entstat for Shared Ethernet

adapter devices since topas monitors networking at the interface level. The seastat

command can be used to monitor SEA client traffic.

There are two key points when monitoring virtual devices. First, if you run out of

processing or memory resources on the VIOS then this affects all of the clients actively

using virtual devices. Careful monitoring and tuning of the VIOS partition is necessary.

Second, if you discover a performance issue, be sure to determine the exact physical

device. Then tune the physical device as you normally would in a non-virtualized

environment. For example, if a client sees performance problems on an hdisk, that

hdisk may be a logical volume on the VIOS that could be moved to a less-busy disk.

UNIX Software Service Enablement

Monitoring Virtual I/O Server resources

 Monitor memory, processors, physical disk controllers, physical disks,

and logical volumes on the Virtual I/O Server:

 If the Virtual I/O Server runs out of resources, it affects all clients actively using

virtual devices

 For device performance issues, determine the physical device and tune

 Virtual I/O Server CLI specific tools are available to monitor resources:

topas, viostat, entstat, vmstat, seastat, vasistat

viostat is the same as the iostat AIX command

 Enable advanced accounting on SEA device to run the seastat command:

$ chdev dev ent# attr accounting=enabled

 All AIX tools are accessible from the root shell (oem_setup_env) to

monitor resources:

sar, ps, lparstat, mpstat, smtctl, nmon, filemon, and so on

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 7-16. Checkpoint QV3441.0

UNIX Software Service Enablement

Checkpoint

1. Use the __________ backup tool to create a mksysb or NIM resources tar file.

2. Use the __________ backup tool to back up virtual device mappings.

3. The ______________ backup tool also runs the _________________ command

to back up non-rootvg volume group structures.

4. True or False: In general, client partitions need more processor resources for

virtual SCSI devices than for native (directly attached) disk devices.

5. True or False: When using shared processors, set the priority (uncapped weight

value) of the Virtual I/O Server partition equal to its client partitions.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Figure 7-17. Exercise 7 - Virtual I/O Server Maintenance QV3441.0

Notes:

UNIX Software Service Enablement

Exercise 7 - Virtual I/O Server Maintenance

 Activity:

 Perform a backup and restore of virtual device mappings

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

Uempty

Figure 7-18. Unit Summary QV3441.0

Notes:

UNIX Software Service Enablement

Unit Summary

 The backupios VIOS tool will back up the VIOS operating system

image including non-rootvg disk configuration

 The viosbr tool is used to back up and restore virtual device

mapping

 Use the updateios tool to keep the VIOS software up to date

Sizing the VIOS partition depends on the workload and devices used

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

AP Appendix A. Checkpoint solutions

Unit 1, "Virtual I/O Server Configuration"

Solutions for Figure 1-22, "Checkpoint (1 of 2)," on page 1-24

UNIX Software Service Enablement

Checkpoint solutions (1 of 2)

1. Which one of the following virtual devices can be used without a Virtual I/O Server?

A. Virtual Ethernet

B. Virtual SCSI

C. Shared Ethernet adapter

D. Virtual Fibre Channel adapter

The answer is A.

2. Which one of the following is NOT a function of virtual I/O?

A. Reduction or elimination of need for additional physical I/O adapters

B. Supports Live Partition Mobility operations

C. Enables partitions to take advantage of unused processing cycles

D. Allows clients to attach to storage that may be unsupported for Linux or IBM I

The answer is C.

3. Which one of the following is NOT a valid way to install the Virtual I/O Server product:

A. Directly from product media

B. Install AIX first, then install the Virtual I/O Server filesets and updates

C. Install from the HMC with the installios command

D. Install from an AIX NIM master

The answer is B.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Solutions for Figure 1-23, "Checkpoint (2 of 2)," on page 1-25

UNIX Software Service Enablement

Checkpoint solutions (2 of 2)

4. True or False: When installing the Virtual I/O Server, it installs just like an

AIX mksysb image. The answer is true.

5. To use virtual SCSI devices, which system feature must be purchased?

A. Virtual I/O Server Software

B. PowerVM Express, Standard, or Enterprise Edition

C. Live Partition Mobility

D. None of the above

The answer is B.

6. What is the command in the Virtual I/O Server command line interface that will

show a list of all available Virtual I/O Server command line commands? The

answer is the help command.

7. True or False: The Virtual I/O Server can be booted in any type of partition. The

answer is false. The Virtual I/O Server product can only be booted in a Virtual I/O

Server type partition.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

AP Unit 2, "Virtual SCSI Configuration"

Solutions for Figure 2-40, "Checkpoint (1 of 2)," on page 2-42

UNIX Software Service Enablement

Checkpoint solutions (1 of 2)

1. Which one of the following statements about virtual SCSI devices is false?

A. The Virtual I/O Server partition owns the physical resources when virtual SCSI

devices are used.

B. To a client, a virtual SCSI device backed by a logical volume or an optical

device looks like an hdisk.

C. One virtual SCSI server adapter can have many associated backing devices.

D. The client partition cannot detect the type of backing storage device being used

for a virtual SCSI disk.

The answer is B. An optical device looks like a cd# device and not like an hdisk.

2. True or False: A partition may boot from virtual SCSI disk. The answer is true.

3. Virtual SCSI devices can be backed with which type of device(s):

A. Logical volumes

B. Physical volumes

C. Optical media devices

D. All of the above

The answer is A, All of the above.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Solutions for Figure 2-41, "Checkpoint (2 of 2)," on page 2-43

UNIX Software Service Enablement

Checkpoint solutions (2 of 2)

4. True or False: On the Virtual I/O Server, the mkvdev command can be used to

create virtual target devices that map physical devices to virtual SCSI server

adapters. The answer is true.

5. Which VIO Server command will show the mapping between Virtual SCSI server

adapters and backing storage devices?

A. lsdev

B. lsmap

C. lscfg

D. lssp

The answer is B.

6. True or False: To remove the mapping between a physical device and virtual

SCSI server adapter, simply remove the virtual target device. The answer is true

but be sure to first remove the virtual device from the client LPAR.

7. True or False: Virtual SCSI client LPARs can use mirroring on virtual SCSI disks.

The answer is true.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

AP Unit 3, "File-backed Storage Devices"

Solutions for Figure 3-21, "Checkpoint," on page 3-23

UNIX Software Service Enablement

Checkpoint solutions

1. True or False: An LPAR can boot from a file-backed hdisk or optical

media device. The answer is true.

2. True or False: The operating system for an LPAR can be installed using

the media repository. The answer is true.

3. True or False: An FBPOOL is the same thing as a volume group. The

answer is false. An LVPOOL is the same thing as a volume group.

4. The ____________________ command creates FB backing devices and

maps them to vhost adapters. The answer is mkbdsp.

5. Before creating any file-backed virtual disks, what must you create first

inside the logical volume storage pool? The answer is an FBPOOL.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Unit 4, "Virtual Fibre Channel Storage Devices"

Solutions for Figure 4-23, "Checkpoint (1 of 2)," on page 4-25

UNIX Software Service Enablement

Checkpoint solutions (1 of 2)

1. True or False: When a client partition uses a virtual Fibre Channel

adapter to access its disks, the VIOS hosting the physical adapter does

not see the clients disks. The answer is true.

2. True or False: You must zone SAN LUNs to both the virtual FC adapter

WWPNs and the VIOSs physical adapter port WWPN. The answer is

false.

3. When using virutal Fibre Channel instead of virtual SCSI, which of the

following is a false statement:

A. VIOS administrator doesnt have to keep documentation about all of the client

endpoint devices.

B. Clients can use other path control modules with added benefits such as load

balancing.

C. NPIV provides an emulated environment.

D. Clients see real type of disk so tools which are specific to a certain type of disk

will work.

The answer is C. It does not provide an emulated environment.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

AP Solutions for Figure 4-24, "Checkpoint (2 of 2)," on page 4-26

UNIX Software Service Enablement

Checkpoint solutions (2 of 2)

4. The name for the virtual FC client adapters port is:

A. vfchost#

B. F_Port

C. N_Port

D. VTD

The answer is C, N_Port.

5. Once a client has a virtual Fibre Channel adapter connection through the

VIOS which supports NPIV, it has a unique identifier on the SAN fabric

called a _________. The answer is WWPN.

6. The name of the virtual FC server adapter device on the VIOS is

_______________________. The answer is vfchost#.

7. True or False: The chsyscfg HMC command workaround procedure will

alter the running configuration of an LPAR. The answer is false. It alters

the profile and then the LPAR has to be activated.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Unit 5, "Virtual Ethernet Adapter Configuration"

Solutions for Figure 5-26, "Checkpoint (1 of 2)," on page 5-28

UNIX Software Service Enablement

Checkpoint solutions (1 of 2)

1. Which one of the following is not part of the IEEE 802.1Q VLAN protocol?

A. Ethernet packets are tagged with a VLAN ID

B. Ethernet switch restricts delivery to members of VLAN

C. Ethernet packets are delivered to target host based on IP address first, then

VLAN ID

The answer is that statement C is false.

2. Which one of the following statements is false about the Hypervisor switch?

A. It is the software implementation of an Ethernet switch and is consistent with the

IEEE 802.1Q protocol

B. Will add the PVID tag to packets from VLAN-unaware hosts

C. Configuring the default VLAN ID is optional when creating virtual Ethernet

adapters

The answer is that statement C is false.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

AP Solutions for Figure 5-27, "Checkpoint (2 of 2)," on page 5-29

UNIX Software Service Enablement

Checkpoint solutions (2 of 2)

3. True or False: When using only the default VLAN ID ("PVID") for a virtual

Ethernet adapter, you must also configure the VLAN device in the operating

system. The answer is false. To use PVIDs you do not need to create a VLAN

device in the operating system.

4. True or False: The Hypervisor Ethernet switch strips the PVID from any packets

that are being delivered to a VLAN-unaware host. The answer is true.

5. What is the maximum number of Hypervisor virtual Ethernet switches that may be

created on a single server? __The answer is 16__

6. True or False: It is possible to set up network communications between partitions

on the same server without any Ethernet hardware in use. The answer is true.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

Unit 6, "Shared Ethernet Adapter Configuration"

Solutions for Figure 6-22, "Checkpoint," on page 6-24

UNIX Software Service Enablement

Checkpoint solutions

1. True or False: To use Shared Ethernet adapters, you must have a Virtual I/O

Server partition. The answer is true.

2. True or False: The virtual adapters configured as part of a Shared Ethernet

adapter must have the Access external network option enabled. The answer is

true.

3. True or False: Only one virtual Ethernet adapter per VLAN can be configured as

the SEA virtual adapter. The answer is false. You can have two: one primary and

another as a failover.

4. True or False: Shared Ethernet adapter threading is important on Virtual I/O

Servers that provide both virtual storage and Shared Ethernet adapter services.

The answer is true.

5. Which of the following cannot be done from the HMC Virtual Network

Management task:

A. Create an SEA to connect multiple VLANs to an external network

B. Delete an SEA (The HMC gives you the list of commands to run.)

C. When creating the SEA, move the TCP/IP configuration of the physical Ethernet

adapter to the SEA device

The answer is B.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

AP Unit 7, "Virtual I/O Server Maintenance"

Solutions for Figure 7-16, "Checkpoint," on page 7-17

UNIX Software Service Enablement

Checkpoint solutions

1. Use the _backupios_ backup tool to create a mksysb or NIM resources tar file.

The answer is backupios.

2. Use the _viosbr_ backup tool to back up virtual device mappings. The answer

is viosbr.

3. The _backupios_ backup tool also runs the _savevgstruct_ command to

back up non-rootvg volume group structures. The answer is savevgstruct.

4. True or False: In general, client partitions need more processor resources for

virtual SCSI devices than for native (directly attached) disk devices. The answer

is false. For clients, there is no difference in processing resources for virtual

versus native disks.

5. True or False: When using shared processors, set the priority (uncapped weight

value) of the Virtual I/O Server partition equal to its client partitions. The answer

is false. Use a higher weight value for the server.

Student Notebook

Course materials may not be reproduced in whole or in part

without the prior written permission of IBM.

V7.0

backpg Back page

LSCAS Main Book File Qv3441 Student Guide

Navigation menu

Versions of this User Manual:

Views

Navigation