RTI_Recording_Service_UsersManual RTI Recording Service Users Manual

User Manual: Pdf

Open the PDF directly: View PDF .
Page Count: 126

Download
Open PDF In Browser	View PDF

RTI Recording Service

User’s Manual
Version 5.2.3

Trademarks
Real-Time Innovations, RTI, NDDS, RTI Data Distribution Service, DataBus, Connext, Micro DDS, the RTI
logo, 1RTI and the phrase, “Your Systems. Working as one,” are registered trademarks, trademarks or
service marks of Real-Time Innovations, Inc. All other trademarks belong to their respective owners.

Copy and Use Restrictions
No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form
(including electronic, mechanical, photocopy, and facsimile) without the prior written permission of RealTime Innovations, Inc. The software described in this document is furnished under and subject to the RTI
software license agreement. The software may be used or copied only under the terms of the license
agreement.

Third-Party Copyright Notices
Portions of this product include software derived from Fnmatch, (c) 1989, 1993, 1994, The Regents of the
University of California. All rights reserved. The Regents and contributors provide this software "as is"
without warranty.

Technical Support
Real-Time Innovations, Inc.
232 E. Java Drive
Sunnyvale, CA 94089
Phone:
(408) 990-7444
Email:
support@rti.com
Website:
http://www.rti.com/support

Contents
1

Welcome to RTI Recording Service
1.1

Paths Mentioned in Documentation ................................................................................................ 1-2

Using Recording Console
2.1

Starting and Stopping the Console .................................................................................................. 2-1

2.2

Using the Information Panel ............................................................................................................. 2-2

2.3

Configuring Recording Console....................................................................................................... 2-2
2.3.1 Configuring From an External File ..................................................................................... 2-3

2.4

Recording Data ................................................................................................................................... 2-6
2.4.1 Using the Pause/Resume Button During Recording ....................................................... 2-7
2.4.2 Troubleshooting Recording Problems ................................................................................ 2-7

2.5

Replaying Data.................................................................................................................................... 2-7
2.5.1 Using the Play Button ........................................................................................................... 2-8
2.5.2 Using the Fast-Forward Button ........................................................................................... 2-8
2.5.3 Using the Pause/Resume Button ........................................................................................ 2-8
2.5.4 Advanced Configuration...................................................................................................... 2-8
2.5.5 Restricting the Time Range to be Replayed ....................................................................... 2-9

2.6

Viewing Recorded Topics .................................................................................................................. 2-9

2.7

Scheduling Recording and Replay Tasks ...................................................................................... 2-10

2.8

Troubleshooting .................................................................................................................................2-11

Using the Record Tool
3.1

Starting the Record Tool .................................................................................................................... 3-1

3.2

Stopping the Record Tool .................................................................................................................. 3-2

3.3

Format of the Recorded Data ............................................................................................................ 3-2
3.3.1 Discovery Data....................................................................................................................... 3-2
3.3.2 User Data ................................................................................................................................ 3-3

Configuring the Record Tool
4.1

How to Load the XML Configuration ............................................................................................. 4-1

4.2

General Format ................................................................................................................................... 4-2
4.2.1 Configuration File Syntax..................................................................................................... 4-3
4.2.2 Supported Data Types........................................................................................................... 4-3

4.3

General Properties for the Record Tool ........................................................................................... 4-6

iii

4.4

Remote Access Properties ................................................................................................................. 4-6
4.4.1 Enabling RTI Distributed Logger in the Record Tool ....................................................... 4-6

4.5

Database (Output File) Properties.................................................................................................... 4-8
4.5.1 Choosing Which SampleInfo and Discovery Fields to Record ..................................... 4-14

4.6

Domain Type Configuration ........................................................................................................... 4-22

4.7

Domain Properties............................................................................................................................ 4-24
4.7.1 Enabling Monitoring Library in the Record Tool............................................................ 4-26
4.7.2 Recording Large User Data Types..................................................................................... 4-26

4.8

TopicGroup Properties..................................................................................................................... 4-27
4.8.1 ‘Create Index’ Syntax .......................................................................................................... 4-30
4.8.2 Indexing and Performance in SQLite: Tips and Tricks................................................... 4-32

4.9

RecordGroup Properties .................................................................................................................. 4-34

4.10 Recording Service Integration with Extensible Types................................................................. 4-36
4.10.1 Selecting a Type Version For a Topic “T” In a Recording Domain ............................... 4-36
4.10.2 Recording Two Versions of a Type in Different Tables in Same Database .................. 4-40

Accessing the Record Tool from a Remote Location
5.1

Overview ............................................................................................................................................. 5-1

5.2

Establishing a Connection with the Record Tool ........................................................................... 5-2

5.3

Remote Control Messages ................................................................................................................. 5-4
5.3.1 Updating the Record Tool’s Partition QoS Policy............................................................. 5-5

5.4

Using the Example Remote-Access Application—Record Shell ................................................5-11
5.4.1 Record Shell’s Commands.................................................................................................. 5-12
5.4.2 Running Multiple Record Tools in the Same Domain.................................................... 5-15

Using the Replay Tool
6.1

Recording Data for Replay ................................................................................................................ 6-1

6.2

Starting the Replay Tool..................................................................................................................... 6-1

6.3

Stopping the Replay Tool .................................................................................................................. 6-2

6.4

Performance and Indexing ................................................................................................................ 6-3

Configuring the Replay Tool
7.1

How to Load Replay’s XML Configuration File ............................................................................ 7-1

7.2

General Format ................................................................................................................................... 7-2

7.3

General Properties for Replay........................................................................................................... 7-3

7.4

Database (Input File) Properties....................................................................................................... 7-4
7.4.1 Enabling Monitoring Library with Replay ........................................................................ 7-6

7.5

Session Properties ............................................................................................................................... 7-6

7.6

Replay Topic Properties ..................................................................................................................... 7-7
7.6.1 Type Selection......................................................................................................................... 7-8

7.7

Time Control Properties..................................................................................................................... 7-9

7.8

Remote Administration Properties .................................................................................................. 7-9
7.8.1 Enabling RTI Distributed Logger in the Replay Tool ......................................................7-11

7.9

Type Configuration .......................................................................................................................... 7-13

7.10 Recording Service Integration with Extensible Types................................................................. 7-14
7.10.1 Selecting the Type Version to use when Replaying a Topic........................................... 7-15
7.10.2 Replaying Topics with Different Type Versions Stored in Different Tables ................ 7-16

Accessing the Replay Tool from a Remote Location

Viewing Recorded Data

10 Converting and Exporting Recorded Data
10.1 Exporting Data .................................................................................................................................. 10-1
10.2 Deserializing Serialized Tables ....................................................................................................... 10-3
10.3 Handling Data Types ....................................................................................................................... 10-3
10.4 Selecting Output Files ...................................................................................................................... 10-4
10.5 Exporting Discovery Tables ............................................................................................................ 10-4
10.6 Filtering User Topic Tables .............................................................................................................. 10-4

11 Example Configuration Files
11.1 How to Record All Topics in a Single Domain..............................................................................11-1
11.2 How To Record a Subset of Data from Multiple Domains ..........................................................11-2
11.3 How To Record Data to Multiple Files ...........................................................................................11-3
11.4 How To Record Serialized Data.......................................................................................................11-3
11.5 How To Record Using Best-Effort Reliability ................................................................................11-4
11.6 How To Enable Remote Access .......................................................................................................11-4

Fields Available for Recording
A.1 User Topic Tables ............................................................................................................................... A-1
A.2 DCPSParticipant Table (Discovery) ................................................................................................ A-2
A.3 DCPSPublication Table (Discovery)................................................................................................ A-3
A.4 DCPSSubscription Table (Discovery) ............................................................................................. A-5

Chapter 1

Welcome to RTI Recording Service

RTI® Recording Service includes:

❏ Recording Console, a simple graphical user interface (GUI) for using the Record and
Replay tools. This interface significantly reduces Recording Service configuration time and
complexity, and does not require any programming. The Recording Console makes it easy
to use Recording Service for testing algorithms and other processing logic against prerecorded test data, conducting regression testing from 'golden' data inputs, or recording
live data from the field for post-mission analysis. See Chapter 2: Using Recording Console.

❏ Record, an RTI Connext™ DDS application that records both RTI Connext DDS discovery
and topic data. All recorded data is stored in one or more SQL database files. See
Chapter 3: Using the Record Tool.

❏ Replay, a tool that can ‘play back’ the recorded data. You even have the option of replaying the data with different data rates or QoS settings. See Chapter 6: Using the Replay
Tool.

❏ Convert, a utility that enables serialized or deserialized data recorded with Record to be
exported to CSV, HTML, SQL, or XML formats. See Chapter 10: Converting and Exporting Recorded Data.
Recording Features

❏ Records data from applications in multiple domains.
❏ Records entire Topics, or specific Topic fields, based on POSIX file-name matching
expressions.

❏ Records all data types except bit-fields.
❏ Records to multiple files with configurable file-size limits. Optionally overwrites the oldest file when the maximum number of files has been reached.

❏ Records the DDS SampleInfo structure and a timestamp for both discovery data and user
data.

❏ Records using either Best Effort or Reliable communications.
❏ Optionally can be configured to record from all partitions or from only specified partitions.

❏ Supports remote operation.
Replay Features

❏ Publishes data samples that were recorded in serialized format.
❏ Highly configurable—you can:

1-1

Paths Mentioned in Documentation

• Choose which serialized topics to replay
• Set the replay rate (faster or slower) or use the original rate
• Change the QoS of the publications
• Configure the QoS for the tool itself
• Dynamically control the replay (start, stop, pause) and single-step through the
data samples
This document assumes you have a basic understanding of DDS terms such as DomainParticipants, Publishers, DataWriters, Topics, and Quality of Service (QoS) policies. For an overview of
DDS terms, please see the RTI Connext DDS Core Libraries User’s Manual.

1.1

Paths Mentioned in Documentation
The documentation refers to:

❏
This refers to the installation directory for Connext DDS.
The default installation paths are:
• Mac OS X systems:
/Applications/rti_connext_dds-version
• UNIX-based systems, non-root user:
/home/your user name/rti_connext_dds-version
• UNIX-based systems, root user:
/opt/rti_connext_dds-version
• Windows systems, user without Administrator privileges:
\rti_connext_dds-version
• Windows systems, user with Administrator privileges:
C:\Program Files\rti_connext_dds-version (for 64-bits machines) or
C:\Program Files (x86)\rti_connext_dds-version (for 32-bit machines)
You may also see $NDDSHOME or %NDDSHOME%, which refers to an environment
variable set to the installation path.
Wherever you see used in a path, replace it with your installation path.
Note for Windows Users: When using a command prompt to enter a command that
includes the path C:\Program Files (or any directory name that has a space), enclose the
path in quotation marks. For example:
“C:\Program Files\rti_connext_dds-version\bin\rtiddsgen”

or if you have defined the NDDSHOME environment variable:
“%NDDSHOME%\bin\rtiddsgen”

1-2

Paths Mentioned in Documentation

❏
Examples are copied into your home directory the first time you run RTI Launcher or any
script in /bin. This document refers to the location of these examples as
.
Wherever you see , replace it with the appropriate path.
By default, the examples are copied here:
• Mac OS X systems:
/Users/your user name/rti_workspace/version/examples
• UNIX-based systems:
/home/your user name/rti_workspace/version/examples
• Windows systems:
your Windows documents folder\rti_workspace\version\examples
Where 'your Windows documents folder' depends on your version of Windows.
For example, on Windows 7, the folder is C:\Users\your user name\Documents;
on Windows Server 2003, the folder is C:\Documents and Settings\your user
name\Documents.
You can specify a different location for the rti_workspace directory. You can also specify
that you do not want the examples copied to the workspace. See the RTI Connext DDS
Core Libraries Getting Started Guide.

1-3

Chapter 2

Using Recording Console

This chapter describes how to use Recording Console, which provides an easy way to record and
replay data.

2.1

Starting and Stopping the Console
Recording Console’s executable is in /bin.
The tool should always be run using the executable script (rtirecordingconsole.bat for Windows, rtirecordingconsole for Linux). For Mac OS X systems, Recording Console includes a .app
directory which includes a script to run the tool.
On Linux systems:

1. Open a command prompt and change to the /bin directory
2. Start the Console by entering:
> rtirecordingconsole
3. Set a Domain ID as described in Section 2.3.
On Mac OS X systems:

1. Go to the /bin directory.
2. Execute the RTI Recording Console.app application by double-clicking on it or using
Mac's open command.
3. If the Connext DDS bundle was installed under the Mac "Applications" directory, Mac's
LaunchPad will also show the RTI Recording Console application in the /bin directory.
Click on the application to run it.
On Windows systems:

1. From the Start menu, navigate to RTI Recording Service , and select Recording Console.
2. Set a Domain ID as described in Section 2.3.
Table 2.1

Console’s Controls

Record

Stop

Pause/Resume

Single Step
(Press Pause first)

Play

Fast Forward

2-1

Using the Information Panel

Figure 2.1

Console at Startup

File
Management
Display
area

Configure
Start here to set a
Domain ID

2.2

Topics
(see Section 2.6)

Schedule
(see Section 2.7)

Controls
(see Table 2.1)

Information
(see Section 2.2)

Using the Information Panel
To open the Information panel, click on the
in the Console’s lower-right corner. From this
panel, seen in Figure 2.2, you can access documentation, contact RTI Support, or view the most
recent log file.
Note: Log files are automatically deleted when the Console shuts down, unless errors were
reported.

Figure 2.2

Information Panel

The information panel provides links to documentation, RTI Support, and the Console’s log file.

2.3

Configuring Recording Console
Before you can use Recording Console to record or replay data, you must specify a domain ID in
the Configuration panels. The default domain ID is 0.

2-2

Configuring Recording Console

❏ To record data, specify the domain in which the data you want to record is being published.

❏ To replay data, specify the domain that you want to publish the previously recorded data
into.
By default, Recording Console records and replays data using the default DomainParticipant QoS
settings described in the Connext DDS documentation. If your Connext DDS applications use
default DomainParticipant QoS settings (including transport settings), you can record and
replay data ‘out of the box’—with no QoS changes.
If your system uses any DomainParticipant QoS settings that would be incompatible with the
default settings, you need to write a configuration file that can be used by Recording Console
when recording or replaying the data.
There are two ways to configure Recording Console’s recording and playback parameters:

❏ Specify values in the Configuration panels shown below.
❏ Specify a file and a configuration within that file. This method is described in
Section 2.3.1.

Click here
to open these
Configuration
Panels

2.3.1

Configuring From an External File
If you have a use case that is not covered by the default configuration generated by Recording
Console, you can use an external configuration file as the basis of the settings to record or replay.
Recording Console can load any configuration file that is supported by the Record or Replay tools.
These files are described in Chapter 4: Configuring the Record Tool and Chapter 7: Configuring
the Replay Tool.
Note: Not all configuration settings are taken from the selected external file; see Section 2.3.1.1
and Section 2.3.1.3.

2-3

Configuring Recording Console

2.3.1.1

Configuring Recording from an External File

To use an external configuration file for recording:
1. Press

to open the Recording and Playback Configuration panels.

2. In the Recording configuration panel, select Configured by file.
3. Select a configuration file to use for recording either by pressing the Open Folder button
or by dragging-and-dropping a file from your file explorer window into the
Recording configuration panel.
4. Select one of the configurations in the file by choosing from the drop-down listbox.
1. Select “Configured by file”

2. Select a
file

3. Select a configuration
Only parts of the configuration file are used, while other settings are taken from Recording Console:

❏ : The name attribute will be used for the launched service.
❏ : This will be used for the
administration domain ID.

❏ : Many of the settings from this element are used, except as noted
below:
• : The domain ID is used, but all other settings are replaced so that
they are compatible with Recording Console's settings.
• : This is replaced with the database file specified in Recording
Console.
Note: To stop using the configuration file and go back to the previous QoS settings, select
Configured without a File.
Return to prior configuration

2.3.1.2

Using Domain 99 in External File Configurations

By default, Recording Console monitors and controls the Record and Replay services on domain 99.

2-4

Configuring Recording Console

When configuring either the Record or Replay services from a file, if you happen to be using
domain 99 as your recording or playback domain, we recommend that you change the administration domain from 99 to a different domain ID, by either specifying it in the configuration file
or by editing the file, settings.ini, which should be located in your home directory:

❏ On Windows Systems:
C:\Users\\Documents\rti_workspace\\user_config\
recording_service\settings.ini

❏ On Linux and OS X systems:
~/rti_workspace//user_config/recording_service/settings.ini
Again, this recommendation only applies if the domain in your configuration file is domain 99.
Note: Except for the specific changes described in this document, avoid making other changes in
the settings.ini file.
2.3.1.3

Configuring Replay from an External File

To use an external configuration file for replay:
1. Press

to open the Recording and Playback Configuration panels.

2. In the Playback configuration panel, select Configured by file.
3. Press the Open Folder button

to select a configuration file for playback.

4. Select one of the configurations in the file by choosing from the drop-down listbox.
5. Select a QoS profile from the drop-down listbox.
1. Select “Configured by file”

2. Select a
file

3. Select a configuration
Here again, only parts of the configuration file are used, while other settings are taken from
Recording Console. These configuration elements from the file are used:

❏ : The name attribute will be used for the launched service.
❏ : This will be used for the administration domain ID. The rest of the element is replaced to ensure runtime compatibility with Recording Console.

❏ : Many of the settings will be used from this element, except for:
• Only the first from the first will be used.
• The will be replaced with the file specified in Recording Console.

2-5

Recording Data

Note: To stop using the configuration file and go back to the previous QoS settings, select Configure without a File.
Return to prior configuration

2.4

Recording Data
To record data:

1. Make sure you have set the domain ID in the Record Configuration panel (see
Section 2.3).
2. Choose where to save the recorded data. You can create a new file or choose an existing
one:
• To create a new file: Press the New Recording button
in the upper-right corner
and specify a file name and location for the new recording. Then click on Create File.
• To record over an existing file:
• Press the Open Folder button
that you want to record into.

in the upper-right corner and locate the file

• Another way to open a recording file is simply to drag the file from your file
explorer window and drop it into the long black rectangle at the top of the Console.
Note: If you specify an existing file, the file will be overwritten with new data. New
data is not appended to the end of the existing file contents.
3. Press the Record button

to start recording.

File size grows as data is recorded

Begin time
Stop

Pause/

End time

2-6

Replaying Data

2.4.1

Using the Pause/Resume Button During Recording
To pause recording, press the Pause/Resume button

To resume recording, either press Pause/Resume again or press Record

2.4.2

Troubleshooting Recording Problems
Problem—You pressed the Record button, but the recording file size stays at zero.
Solution—Make sure that:

❏ The Recording Domain ID matches the domain ID used by the source (the application
from which you want to record data).

❏ Data is coming in from the source, by using tools such as rtiddsspy (provided with Connext DDS in its /bin directory).

❏ You have access rights to create files in the directory where the recording file is to be created.

2.5

Replaying Data
You can use the Console to replay data that was recorded using the Console or the Record tool. You
may replay data recorded with an older version of Recording Service.
Note: If you recorded the data using the sqlite_ pragmas functionality described in Section 4.5,
the resulting database cannot be replayed by Recording Console; use the Replay tool instead (see
Chapter 6).
To replay data:

1. Make sure you have set the domain ID in the Playback Configuration panel (see
Section 2.3).
2. Press the Open Folder button
to be replayed, then click Open.

in the upper-right corner, locate the file whose data is

• Another way to open a recorded data file is simply to drag the file from your file
explorer window and drop it into the long black rectangle at the top of the Console.
When the file is loaded, you will see the time of the original recording:
Original recording time and date

3. Press Play

to begin replaying the data.

2-7

Replaying Data

The display will show you the elapsed time since the start of the replay.
Slider controls
replay speed

Elapsed time

Replay rate. For example.
x1 = original rate,
x2 = twice the original rate.

2.5.1

Using the Play Button
To begin replaying data, press Play

2.5.2

Using the Fast-Forward Button
The Fast-Forward
button will replay at a higher rate as long as the button is pressed.
“Higher rate” means the next highest option from the list of playback speed choices in the dropdown menu. For example, if you were replaying at normal speed, fast-forward will replay at x2
(double) speed. If you were replaying at double speed, fast-forward will replay at x10 speed.
If replay is paused when you press Fast-Forward
button, it will replay at the higher rate as
long as the button is pressed. When you let go of the button, replay will go back to being paused.
If replay is not paused when you press Fast-Forward
button, it will replay at the higher rate
as long as the button is pressed. When you let go of the button, replay will return to the previous
rate.

2.5.3

Using the Pause/Resume Button
To pause a replay, press the Pause/Resume button

To resume the replay at the same rate, either press Pause/Resume again or press Play

2.5.4

Advanced Configuration

2.5.4.1

Changing the Replay Rate

To increase the rate temporarily, keep the Fast-Forward Button pressed (see Section 2.5.2).
The vertical slider on the right controls the replay rate (up for faster, center for original speed,
lower for slower).

2-8

Viewing Recorded Topics

You can specify a default replay rate in the Playback configuration panel. Select a speed from the
drop-down list or type in your own value.
If you use the slider to change the playback speed, you can easily go back to default playback
rate by pressing the Play button again while replay is in progress.

2.5.4.2

Auto-Repeat

To automatically repeat the replay in a continuous loop, select the Auto repeat check-box.

2.5.5

Restricting the Time Range to be Replayed
You can limit the start and end time for replaying data while replay is stopped by dragging the
bars seen below:

Drag these bars inward to restrict the time range for replaying data

Note: This time-restriction feature cannot be used when using a configuration file.

2.6

Viewing Recorded Topics
While recording, or when you have loaded a pre-recorded file, you can use the Recorded Topics
panel to see the topics that have been recorded. For each topic, the table shows the topic name,
and (when the recording is not in progress) the first and last recorded samples of that topic.

2-9

Scheduling Recording and Replay Tasks

If playback is configured through Recording Console, the topic table enables you to select which
topics to replay.

Click here
to open the
Recorded
Topics Panel

Search for topics by name here
Searching for Topics

To assist in selecting multiple topics, use the search bar on the bottom of the topics panel. You
can narrow down the topics that are displayed based on a substring in the topic name.
Note: The search bar does not support regular expressions.
When the desired group of topics is displayed, you can use the select all/unselect all buttons on
the entire group. Only the selected Topics will be replayed.
To restore and see the list of all topics, remove (erase) the search string from the search box.
Topic selection in the table is only available when the Console is in Ready mode (not recording or
replaying data) and you have used the Playback Configuration panel (not a configuration file).

2.7

Scheduling Recording and Replay Tasks
To schedule recording or replay:
1. Press the Schedule button

2. Select the type of task (record, replay, or stop current operation).
3. Select the starting time and date.

2-10

Troubleshooting

4. Optionally, select an ending time for the activity.

Important Notes:

❏ Recording Console’s window must remain active for the scheduled operation to run. You
may minimize the window, but closing it will cancel the activity.

❏ When selecting a file in which to record, be aware that any data already in the file will be
erased.

2.8

Troubleshooting
If Recording Console hangs, issues a generic error message, or behaves in an unexpected manner,
here are some steps to resolve the problem:
1. Read the instructions again to ensure you understand the expected behavior.
2. It’s possible that the environment or configuration in which Recording Console is running
is causing the unexpected behavior. Recording Console generates a log file. To open the log
file, click on the info button ('i') in the lower right corner, then click on the link at the bottom of the info panel. Each line in the log begins with a severity level. Look for lines that
start with WARN or ERROR.
3. Contact RTI’s support team.
If you cannot resolve the issue based on the information in the log file, please contact us
so we can help you overcome the issue.
Before contacting support, please increase the logging verbosity and try to reproduce the
problem so you can provide more detailed information. To do so:
a. Close any running instances of Recording Console (otherwise, any change you make to
the settings file may be overwritten).
b. Open the console settings file in your home directory:
• rti_workspace/version/user_config/recording_service/settings.ini (rti_workspace
is described in Paths Mentioned in Documentation (Section 1.1))
Note: Except for the specific changes described in this document, avoid making other
changes in the settings.ini file.
c. Find the line with "loggingLevel=INFO" and replace the default value of "INFO" with
"TRACE".
d. Save the updated settings file.

2-11

Troubleshooting

e. Run Recording Console again until the issue you encountered before is reproduced.
f. Find the log file. The info panel has a link indicating the name of the current log file.
In general, log files are stored here: rti_workspace/version/user_config/RTI Recording Console /logs
Log files are named by date and time. Log files are only saved if there is an error. If no
errors are reported, the log file is deleted when Recording Console shuts down.
g. Save a copy of the log file in a safe location and close Recording Console. Review the log
file and remove any sensitive information that you would prefer not to expose.
h. Repeat the above steps to restore the logging level to the default "INFO" setting.
i. Contact RTI support. Click on the info button ('i') in the lower right corner of Recording Console to open the info panel, which contains links for contacting RTI support.
Provide a description of the problem, how to reproduce it, and a copy of the log file. If
possible, please include a screenshot that demonstrates the problem.
RTI's support team is very responsive, you should receive a reply within 24 hours.

2-12

Chapter 3

Using the Record Tool

Besides using Recording Console, you can also record data by using the Record tool. While the
Console provides a simple graphical user interface (GUI) for using the Record tool, you can also
run it directly, without using the Console. You may find this method of recording useful when
you want to tie its service into your own infrastructure or software or if you need to use its more
advanced features. For instance, perhaps you want to run them from your own script to record
periodically or to process the recorded data automatically.
See also:

❏ Chapter 4: Configuring the Record Tool
❏ Chapter 5: Accessing the Record Tool from a Remote Location

3.1

Starting the Record Tool
Open a command prompt1 and change to the /bin directory. Then enter:

❏ On Linux and Mac OS X systems:
> rtirecord -cfgFile -cfgName

❏ On Windows systems:
> rtirecord.bat -cfgFile -cfgName

To see a list of available arguments, enter rtirecord -help.
The Record tool is dynamically linked against the Connext DDS libraries. You should run the tool
from the scripts in the /bin directory—not from the executable files themselves.
The scripts set all the paths and variables needed for the tool to find the shared libraries and run
correctly.
To see which configurations are available, use the -listCfgs option:
$ rtirecord -listCfgs

To use your own configuration file:
$ rtirecord -cfgFile config-file.xml -cfgName my_record_cfg

Table 3.1 describes the command-line options and which ones are required.

1. On Windows systems: from the Start menu, select Accessories, Command Prompt.

3-1

Stopping the Record Tool

Chapter 4: Configuring the Record Tool describes the contents of the configuration file. Example
files are provided in the /recording_service directory.
Table 3.1

Record Tool’s Command-Line Options

Command-line Option

3.2

Description

-appName

Assigns an application name to the DomainParticipants created by the Record
tool. If not specified, the same name used in -cfgName will be used.

-cfgFile

Required.
Specifies the XML configuration file (path and filename).
In addition to the file provided using this command-line option, the Record
tool can load other XML files—see Section 4.1.

-cfgName

Required.
This name is used to find the matching tag in the configuration
file.

-help

Prints version information and list of command-line options.

-licenseFile

Specifies the license file (path and filename). Only applicable to licensed
versions of the Record tool.

-listCfgs

Lists the available configuration profiles.

-verbosity

Specifies what type of logging information should be printed.
0: silent (Connext DDS and the Record tool)
1: errors (Connext DDS and the Record tool) (default)
2: warnings (the Record tool only)
3: warnings (Connext DDS and the Record tool)
4: information (the Record tool only)
5: tracing (the Record tool only)
6: tracing (Connext DDS and the Record tool)
This property can also be set in the configuration file. However, this command-line option overrides the value specified in the configuration file.

Stopping the Record Tool
To stop the Record tool: Press Ctrl-C. The Record tool will close all files and perform a clean
shutdown.
You can also start, stop, and even reconfigure the Record tool remotely—see Chapter 5: Accessing the Record Tool from a Remote Location.

3.3

Format of the Recorded Data
This section describes the format of the recorded data. For information on viewing the data, see
Chapter 9: Viewing Recorded Data.

3.3.1

Discovery Data
The Record tool stores discovery-related data in these tables:

❏ DCPSParticipant — corresponds to the Participant Built-in Topic

3-2

Format of the Recorded Data

❏ DCPSPublication — corresponds to the Publication Built-in Topic
❏ DCPSSubscription — corresponds to the Subscription Built-in Topic
Please refer to the RTI Connext DDS C API Reference HTML documentation for the fields in each
of the corresponding builtin topics. (In the HTML documentation for the C API, select Modules,
RTI Connext DDS API Reference, Domain Module, Built-in Topics.)
Note: When using self-contained database files, the locator_filter column of the Publication
Built-in Topic Data will not be replicated. The Subscription Built-in Topic Data will also not be
replicated. This is done to minimize the overhead of replication when opening a new database
file.
The fields stored for each discovery table are listed in Appendix A. If no field selection settings
enabled, the Record tool will just create the DCPSPublication table with enough information for
the Replay tool to work. Fields can be added to the tables, see Choosing Which SampleInfo and
Discovery Fields to Record (Section 4.5.1). If fields are added to the DCPSParticipant and DCPSSubscription tables they will be created.

3.3.2

User Data

3.3.2.1

Deserialized Data

When the Record tool stores data in deserialized form, it creates a mapping from a Topic’s type to
a table. Each individual scalar is stored in a column named with the fully qualified name.
For example, the following will create a column, bar$x:
struct Bar {
long x;
};
struct Foo {
Bar bar;
};
3.3.2.2

Serialized Data

If the topic data is saved in serialized form, the user data table will contain the following columns:

❏ rti_serialized_sample: Raw binary data for the sample
❏ rti_serialized_length: Length of the raw data blob
❏ rti_serialized_endian: 1 — little endian, 0 — big endian
3.3.2.3

Sample Info and Metadata Fields

The Record tool creates a table called TopicName$RecordGroupName$Domain-Name for each
recorded topic (unless the shared_table property is true). The Record tool stores topic data as
specified in the subscription properties.
For each topic, the Record tool also stores the following extra columns, by default:

❏ The Sample Info’s reception timestamp (in nanoseconds since Jan. 1st, 1970).
❏ The Sample Info’s valid data boolean indicator.
❏ Table_prefix: A string representing “RecordGroupName$DomainName” (only if
is specified—see TopicGroup Properties (Section 4.8)).
The fields selected for recording can be modified using the tag in
both the Database (Output File) Properties (Section 4.5)) and RecordGroup Properties (Section

3-3

Format of the Recorded Data

4.9) in the Recorder configuration. The name and SQL type for all the fields that can be recorded
appear in Appendix A.

3-4

Chapter 4

Configuring the Record Tool

When you start the Record tool, you may specify a configuration file in XML format (it is not
required). In that file, you can set properties that control what to record, how to record, and
where to save the recorded data. This chapter describes how to write a configuration file.

4.1

How to Load the XML Configuration
The Record tool loads its XML configuration from multiple locations. This section presents the
various approaches, listed in load order.
The first three locations only contain QoS Profiles and are inherited from Connext DDS (see the
chapter on Configuring QoS with XML in the RTI Connext DDS Core Libraries User's Manual).

❏ /resource/xml/NDDS_QOS_PROFILES.xml
This file contains the Connext DDS default QoS values; it is loaded automatically if it
exists. (First to be loaded.)

❏ File in NDDS_QOS_PROFILES
The files (or XML strings) separated by semicolons referenced in this environment variable are loaded automatically.

❏ /USER_QOS_PROFILES.xml
This file is loaded automatically if it exists. If the USER_QOS_PROFILES file is found and
there is a default profile specified in it, this default profile is automatically applied to the
QoS settings of the Recording Service entities.
The next locations are specific to Recording Service.

❏ /resource/xml/RTI_RECORDING_SERVICE.xml
This file contains the default configuration for the Record tool; it is loaded if it exists.
RTI_RECORDING_SERVICE.xml defines a configuration that records all topics on
domain 0.

❏ /USER_RECORDING_SERVICE.xml
This file is loaded automatically if it exists.

❏ File specified with the command-line option, -cfgFile (see Table 3.1 on page 3-2).
❏ File specified using the remote command ‘configure’

4-1

General Format

The configure command (see Table 3.1 on page 3-2) allows loading of an XML file
remotely. The file loaded using this command replaces the file loaded using the -cfgFile
command-line option. (Last to be loaded.)
You may use a combination of the above approaches.

4.2

General Format
The configuration file uses XML format. The XSD definitions followed by the configuration are
in the document resource/schema/rti_record.xsd. All the following main sections are optional; if
specified, they must be in this order:

❏ General Properties for the Record Tool (Section 4.3)
❏ Remote Access Properties (Section 4.4)—contained in the top-level tag,
❏ Database (Output File) Properties (Section 4.5)—contained in the top-level tag,

❏ Domain Type Configuration (Section 4.6)—contained in the top-level tag,

❏ Domain Properties (Section 4.7)—contained in the top-level tag,

❏ TopicGroup Properties (Section 4.8)—contained in the top-level tag,
❏ RecordGroup Properties (Section 4.9)—contained in the top-level tag,
Let’s look at a very basic configuration, just to get an idea of its contents. You will learn the
meaning of each line as you read the rest of this chapter.

simple_config.dat

RTIDDS_DESERIALIZEMODE_ALWAYS

4-2

General Format

domain0

All

Example configuration files are provided in the examples/record directory:

❏ simple_config.xml
With this configuration, the Record tool will record all fields from all topics in a specified
domain (domain ID 0).

❏ advanced_config.xml
With this configuration, the Record tool will record:
• The ‘x’ and ‘y‘’ fields from all Topics named Square in domains 0 and 1.
• The ‘color’ field from all Topics in domains 0 and 1.

❏ remote_shell.xml
This configuration file provides a configuration that can be used with the tutorial found
in the Recording Service Getting Started Guide to learn about how to modify the Record tool
while it is running.

4.2.1

Configuration File Syntax
Recording Service follows the same XML syntax rules as Connext DDS. Please see the RTI Connext
DDS Core Libraries User’s Manual for details.

4.2.2

Supported Data Types
As you will see in the following sections, each property that can appear in the configuration file
uses a specific data type. The Record tool converts between the value string in the XML file and
the specified type. Table 4.1 lists the supported types and the mappings used by the Record tool.

Table 4.1

Property Value Data Types

Type
char and octet
sequences and
arrays
DDS_Boolean

Format

Notes

Compact form
yes,1,true,on: TRUE
no,0,false,off: FALSE

These values are not case sensitive.

4-3

General Format

Table 4.1

Property Value Data Types

Type

Format

Notes
Enum values are not case sensitive.

DDS_Enum

A string

Legal values are those listed for the property in the
RTI Connext DDS C API Reference HTML
documentation.
A 32-bit signed integer.
You may include the following unit designations:
KB — 2^10
kB — 10^3

DDS_Long

-2147483648 - 2147483647
0x80000000 - 0x7fffffff

MB — 10^6
GB — 10^9
KiB — 2^10
MiB — 2^20
GiB — 2^30
For example, 100 kB is a legal value, meaning
100,000.
A 32-bit unsigned integer.
You may include the following unit designations:
KB — 2^10
kB — 10^3

DDS_UnsignedLong

0 - 4294967296
0 - 0xffffffff

MB — 10^6
GB — 10^9
KiB — 2^10
MiB — 2^20
GiB — 2^30
For example, 100 kB is a legal value, meaning
100,000.

4-4

General Format

Table 4.1

Property Value Data Types

Type

Format

Notes
Each field in each QoS policy structure has a
corresponding tag. The tag is the same as the field
name in the Connext DDS C API.

DDS_QosPolicy

See the RTI Connext DDS C
API Reference HTML documentation for the structure of
each QoS policy, and the RTI
Connext DDS Core Libraries
User’s Manual’s chapter on
Configuring QoS with XML.

For enumerations, the legal constants are those
defined for the Connext DDS C API.
For example:

DDS_TOPIC_PRESENTATION_QOS

rti

The above configuration will set (a) the Presentation
QoS
policy’s
access_scope
field
to
DDS_TOPIC_PRESENTATION_QOS and (b) the
Partition QoS policy’s name field to “rti”. (name is a
sequence of strings, which requires using the tag, also described in this table.)

FileSize

64 bit integer

You may include the following unit designations:
kB — 10^3
MB — 10^6
GB — 10^9
KB — 2^10
TB — 10^12
KiB — 2^10
MiB — 2^20
GiB — 2^30
TiB — 2^40
For example, 100 kB is a legal value, meaning
100,000.

String

UTF-8 character string

All leading and trailing spaces are ignored between
two tags.

4-5

General Properties for the Record Tool

4.3

General Properties for the Record Tool
Table 4.2 describes optional properties that control the Record tool’s main module.

Table 4.2

General Properties

Property
auto_start

replay_
compatibility

verbosity

4.4

Syntax

DDS_Boolean

DDS_Long

Description
Whether or not the Record tool should start recording data
when it is started. This option is mostly useful if the Record
tool is usually controlled remotely.
Default: True
If set to true, the Record tool will always preserve the fields
needed by the Replay tool to work with the recorded database, regardless of the field selection settings (see Table 4.4,
“Database Properties” and Table 4.15, “RecordGroup
Properties”) specified by the user.
Default: True
The verbosity is a bit-map that specifies what type of logging information should be printed.
The verbosity may be:
0: silent (Core Libraries and the Record tool)
1: errors (Core Libraries and the Record tool) (default)
2: warnings (the Record tool only)
3: warnings (Core Libraries and the Record tool)
4: information (the Record tool only)
5: tracing (the Record tool only)
6: tracing (Core Libraries and the Record tool)

Remote Access Properties
As you will see in Chapter 5: Accessing the Record Tool from a Remote Location, you can create
a Connext DDS application that can remotely control the Record tool.
By default, Remote Access is turned off in the Record tool for security reasons.
The Remote Access section of the configuration file is used to enable Remote Access and configure its behavior. A Remote Access section is not required in the configuration file.
The remote application can send commands to the Record tool that will:

❏ Start/stop recording.
❏ Shutdown the Record tool.
❏ Reconfigure the Record tool.
Table 4.3 describes the Remote Access properties. All Remote Access properties must be specified inside and tags. All remote access properties are
optional unless otherwise noted.

4.4.1

Enabling RTI Distributed Logger in the Record Tool
The Record tool provides integrated support for RTI Distributed Logger.

4-6

Remote Access Properties

Table 4.3

Remote Access Properties

Properties
under

accept_
broadcast_
commands

Syntax

DDS_Boolean

enabled

DDS_Boolean

datareader_qos

DDS_DataReaderQos

datawriter_qos

DDS_DataWriterQos

distributed_
logger

Distributed Logger
Properties

publish_
status_
period

DDS_Long

publisher_qos

DDS_PublisherQos

Description
Specifies if the Record tool will accept commands
that have been broadcast to any Record tool or only
accept commands addressed specifically to it. (See
Chapter 5: Accessing the Record Tool from a
Remote Location.)
Default: True
Enables or disables remote access to the Record tool
from another application. (See Chapter 5: Accessing
the Record Tool from a Remote Location.)
Default: False (remote access disabled)
Configures the QoS for the DataReader created by
the Record tool’s Remote Access module.
Default: default DataReader QoS settings.
Configures the QoS for the DataWriter created by
the Record tool’s Remote Access module.
Default: Default DataWriter QoS settings.
Configures RTI Distributed Logger.
See Enabling RTI Distributed Logger in the Record
Tool (Section 4.4.1).
Specifies, in seconds, the period between each status message sent by the Record tool.
Default: 1. Minimum value: 1.
Configures the QoS for the Publisher created by the
Record tool’s Remote Access module.
Default: default Publisher QoS settings.
Required if enabled is true.

remote_access_
domain

String

Specifies which domain the Record tool will use to
enable remote access. Only one domain can be
specified.
Note that this is a String, not a Domain ID. It is the
same String used in the
line.
Default: False

subscriber_qos

DDS_QosPolicy

Configures the QoS for the Subscriber created by
the Record tool’s Remote Access module.
Default: Default Subscriber QoS settings.

Distributed Logger is included in Connext DDS but it is not supported on all platforms; see the RTI
Connext DDS Core Libraries Platform Notes to see which platforms support Distributed Logger.
When you enable Distributed Logger, the Record tool will publish its log messages to Connext
DDS. Then you can use RTI Monitor1 to visualize the log message data. Since the data is pro-

1. RTI Monitor is a separate GUI application that can run on the same host as your application or on a different host.

4-7

Database (Output File) Properties

vided in a Connext DDS topic, you can also use rtiddsspy or even write your own visualization
tool.
To enable Distributed Logger, modify the Record tool’s XML configuration file. In the
section, add the tag as shown in the example below.

true
adminDomain

true

There are more configuration tags that you can use to control Distributed Logger’s behavior. For
example, you can specify a filter so that only certain types of log messages are published. For
details, see the Distributed Logger section of the RTI Connext DDS Core Libraries User’s Manual.

4.5

Database (Output File) Properties
Table 4.4 describes the Database properties. All database properties are optional except
database_name. All database properties must be specified within and
tags.
The Record tool stores data in a set of SQL database files. (Note, however, that you do not need to
install any database software to use the Record tool.)
[Note: Replaying data from a set of files is not supported. This holds true for both Recording Console and
the Replay tool. They can only replay data from one file at a time.]

A fileset is a named collection of file segments which belong to the same recording session. Each
of these file segments contains discovery and user-data, and the format is determined by SQLite.
The Record tool uses a fixed file-naming scheme:
name_set-number_segment-number
Where:

❏ name is the base filename for the fileset, specified in the configuration file with the
property.

❏ set-number is an integer identifying the fileset, specified in configuration file with the
property.

❏ segment-number is an integer identifying the file-segment within the fileset. The first segment number to use, and the maximum number of segments are specified in the configuration file with the and properties,
respectively.
For example: mydata_5_3 means this file belongs to fileset 5 and is the 3rd segment in that fileset.
The maximum size of a file segment, whether to overwrite existing files, and whether to overwrite the oldest file can all be set in the configuration file.

4-8

Database (Output File) Properties

Table 4.4

Database Properties

Properties
under

Syntax

ParticipantData fields

builtin_topic_
metadata_
PublicationData fields
fields

SubscriptionData fields

Description

Specifies which fields to include/ exclude in the
discovery tables.
There are settings for each of the three recorded
discovery tables described in Table 4.5.

Required.

database_
name

flush_period

String

The name of the fileset used to store recorded data.
The Record tool appends a set number and a segment number to the filename.
Default: undefined
Example:
myfile

Specifies how often (in seconds) to flush data to
disk.
Note: increasing this value causes the Record tool to
use additional memory.

DDS_Long

Default: 1 second
Minimum: 1 second

max_file_
segments

DDS_Long

Specifies how many file segments may be created.
Each time the max_file_size limit is reached for a
file segment, a new file is created if this number of
segments has not been exceeded.
Default: 1
Example:

100

max_file_size

FileSize

Specifies the maximum size for a file segment.
The Record tool records data to one or more files.
This property specifies the maximum file size. This
is not an absolute value, but a threshold value. As
soon as the threshold is exceeded, no more data is
written to file.
Default: 2 GB
Maximum: imposed by the operating system
Example:
1 GB

4-9

Database (Output File) Properties

Table 4.4

Database Properties

Properties
under

Syntax

Description
Specifies whether or not the Record tool should
delete all existing file segments in the fileset before
it starts recording.

overwrite

DDS_Boolean

This is useful if you want to reuse a data-file name
between recording sessions, but do not want to
keep any old data.
• True: If the file segments already exist, they
are deleted; otherwise, the file segments are
created as needed.
• False: If the file segments already exist, the
Record tool exits; otherwise, the file segments
are created as needed.
Example:
test
4
yes
In this case, the Record tool will delete test_0_0,
test_0_1, and test_0_2 before starting to record to
test_0_0.
Default:False
Specifies the path separator character that the
Record tool will use when creating table and
column names.
For instance, table names follow the
"TopicName$RecordGroupName$DomainName"
convention and fields in Topics use $ to navigate
hierarchical types, such as a$b$c..

path_
separator

DDS_Char

The dollar sign (‘$’) is used as the default path
separator instead of the more conventional comma
(',') because $ does not require quotes when used in
SQLite SQL statements.
For example, to use '#' as the path separator:
#
Notes:
• Using a character that can be included in a
field name, such as an "_" (underscore) may
lead to errors when you try to replay the file
with Replay if any field contains that character.
• This is not a required property, but when
added, it cannot be empty.

4-10

Database (Output File) Properties

Table 4.4

Database Properties

Properties
under

rollover

Syntax

DDS_Boolean

Description
Specifies whether or not the Record tool should
overwrite existing file segments in the fileset once
the file size limit (max_file_size) has been reached
for the last file segment.
• True: Overwrite existing file segments as
needed (starting with the first one).
• False: Stop recording data.
Default: False
Specifies the first segment to use in the fileset.

segment_
number

DDS_Long

If the segment number is >= 0, that is the first segment number in the fileset.
Default: -1. The next available segment number
will be used, starting at 0.
Note:The set number is determined first, then the
segment number.
Specifies the set number to use in the fileset.

set_number

DDS_Long

If set_number is >= 0, that specific fileset number is
used. In this case, the property takes
effect.
Default: -1. The next available set number will be
used, starting at 0.

4-11

Database (Output File) Properties

Table 4.4

Database Properties

Properties
under

Syntax

Description
Specifies a list of SQLite pragma statements to be
applied to the database.
There must be at least one '' object in the
list; there is no upper limit for the number of
pragmas that can be specified.
Pragmas are applied right after database creation
and before any table is created.
The pragma name is just the name of the SQLite
pragma (e.g., page_size) as defined by SQLite.

sqlite_
pragmas

pragma name
[plus assignment or
parameters]

...

The PRAGMA SQLite keyword should not be
specified, as the Record tool will do that
automatically.
Values or parameters can be provided (e.g.,
page_size = 4096 or
table_info(DCPSPublication)).
The output of the pragma statements is shown in
standard output.
See https://www.sqlite.org/pragma.html
Notes:
• One possible pragma is the database journal
mode. One of the journal modes is WAL
(write-ahead logging). This mode improves
concurrency when reading and writing to the
same database file from different processes or
threads. However, WAL doesn't work on network file systems.
• The order in which you specify the pragmas is
the order in which they will be applied.

user_topic_
metadata_
fields

field expression

...

field expression

...

Specifies lists of included and excluded Sample
Info or metadata fields in the user topic tables.
A field expression is a Posix fnmatch expression to
match field names as described in Appendix A
(e.g. "SampleInfo_*" will match all Sample Info
specific fields).
There is no upper limit to the number of
expressions to be specified.
The ‘included’ and ‘excluded’ fields are optional. If
both are defined, the ‘included’ expressions are
processed before the ‘excluded’ ones.
User topic field settings can also be expressed in
the Record Group settings (see Table 4.15,
“RecordGroup Properties”). If defined, Record
Group settings take precedence over general
database settings defined here.

4-12

Database (Output File) Properties

Table 4.5

Builtin Topic (Discovery) Field Selection Properties

Properties under

Syntax

Description

field expression

...
DCPSParticipant_
topic

field expression

...

Specifies lists of included and excluded fields in
the DCPSParticipant table.
A field expression is a POSIX fnmatch expression
to match field names as described in Appendix A
(e.g. "ParticipantData_*" will match all
DCPSParticipant specific fields).
There is no upper limit to the number of
expressions to be specified.
The ‘included’ and ‘excluded’ fields are optional.
If both are defined, the ‘included’ expressions are
processed before the ‘excluded’ ones.

field expression

...
DCPSPublication_
topic

field expression

...

Specifies lists of included and excluded fields in
the DCPSPublication table.
A field expression is a POSIX fnmatch expression
to match field names as described in Appendix A
(e.g. "PublicationData_*" will match all
DCPSPublication specific fields).
There is no upper limit to the number of
expressions to be specified.
The ‘included’ and ‘excluded’ fields are optional.
If both are defined, the ‘included’ expressions are
processed before the ‘excluded’ ones.

DCPSSubscription
_
topic

field expression

...

field expression

...

Specifies lists of included and excluded fields in
the DCPSSubscription table.
A field expression is a POSIX fnmatch expression
to match field names as described in Appendix A
(e.g. "SubscriptionData_*" will match all
DCPSSubscription specific fields).
There is no upper limit to the number of
expressions to be specified.
The ‘included’ and ‘excluded’ fields are optional.
If both are defined, the ‘included’ expressions are
processed before the ‘excluded’ ones.

4-13

Database (Output File) Properties

4.5.1

Choosing Which SampleInfo and Discovery Fields to Record
To reduce database size and increase flexibility, the Record tool provides a way to select which
Sample Info, discovery (DCPSPublication, DCPSSubscription and DCPSParticipant), and metadata fields should be recorded. An application may not be interested in recording all the metadata, SampleInfo, and discovery information.
Before this feature was introduced, the Record tool would record everything by default. Now, the
Record tool will only record the fields necessary for the Replay tool to work with the database.
These fields are:

❏ In the user table:
SampleInfo_valid_data and SampleInfo_reception_timestamp.

❏ In the DCPSPublication table:
PublicationData_topic_name,
PublicationData_type_name,
PublicationData_typecode and PublicationData_typecode_length.
The Record tool provides capabilities to add to, or to remove from, this set of fields. It also provides a special boolean flag called replay_compatibility (see Table 4.2, “General Properties”).
When this flag is enabled (the default), it preserves the above set of fields, regardless of which
fields the user chooses to exclude. This way very general regular expressions can be used without affecting compatibility with the Replay tool.
Table 4.4 describes the XML settings for selecting fields in the discovery tables (tag
). There are settings for each of the three discovery tables created by the Record tool (tags , and
). In any of the tables, if you decide to exclude all fields (for example:
* ), the table will not be created; that is, if a table would
end up with no columns, the Record tool won't create it.
The field selection settings also apply to user topic tables. The tag used to define those settings is
called ; it can be used inside to define general settings, and inside to define settings that apply only to a specific Record
Group. Record Group settings take precedence over general database settings. These settings are
described in Table 4.4, “Database Properties”and Table 4.15, “RecordGroup Properties”.
4.5.1.1

Field Selection Examples

This section shows some examples of how to configure field selection settings for the Record tool.
The following is a basic recording configuration used as the starting point for the examples. It is
based on the default configuration shipped with the product (see
the file
RTI_RECORDING_SERVICE.xml in the resource/xml directory):

true

true
domain0

true
WARNING

4-14

Database (Output File) Properties

rti_recorder_default.dat

0
RTIDDS_DESERIALIZEMODE_ALWAYS

domain0

AllTopics

By default, the Record tool will record only the fields that the Replay tool will need to work with
the recorded database. Furthermore, because the replay_compatibility flag is set to true, these
fields will be protected by the Record tool and will always be present when recording.
4.5.1.2

Example 1: Record Everything

The
following
configuration
uses
the
builtin_topic_metadata_fields
and
user_topic_metadata_field settings in the recorder_database section to record all possible fields
in all tables, including the discovery ones.

true

true
domain0

true
WARNING

rti_recorder_default.dat

4-15

Database (Output File) Properties

0
RTIDDS_DESERIALIZEMODE_ALWAYS

domain0

AllTopics

4-16

Database (Output File) Properties

4.5.1.3

Example 2: Record No Extra Fields

The following configuration will record only user data fields in user data tables. No Sample Info
or metadata fields will be recorded. Moreover, no discovery information is selected to be
recorded so no discovery tables will be created. The replay_compatibility flag must be disabled
for this configuration to work, because if enabled, it would lock the Replay tool’s compatibility
fields and they would be recorded.

false

true
domain0

true
WARNING

rti_recorder_default.dat

4-17

Database (Output File) Properties

0
RTIDDS_DESERIALIZEMODE_ALWAYS

domain0

AllTopics

4.5.1.4

Example 3: Overriding Database Settings with Record Group Settings

In the following configuration, two sets of settings are defined for user topic tables. The settings
in the general database section include the SampleInfo_source_timestamp field in the user
topic field selection settings (apart from the fields included by default). There is an extra Topic
Group called KeyedTopics, which the user will associate with keyed topics. There may be an
interest in recording the instance handle field (SampleInfo_instance_handle) for these topics.
This may be done by defining a new Record Group with different field selection settings, as
shown below.
One important aspect of the configuration is the use of the exemption expression in the AllTopicsExceptKeyed Topic Group. By using the exemption expression 'Keyed*' we make the two
available Topic Groups (AllTopicsExceptKeyed and 'KeyedTopics) disjoint, because AllTopicsExceptKeyed will never match the same topics as KeyedTopics. This is very important to be
sure that the right settings apply to the right topics.

true

true
domain0

true
WARNING

4-18

Database (Output File) Properties

rti_recorder_default.dat

SampleInfo_source_timestamp

0
RTIDDS_DESERIALIZEMODE_ALWAYS

Keyed*

domain0

AllTopicsExceptKeyed

domain0

KeyedTopics

4-19

Database (Output File) Properties

SampleInfo_instance_handle

4.5.1.5

Example 4: Recording Everything Except the Metadata Fields

The following example configuration shows how to record all fields except the metadata fields
(prefixed with Metadata_). The key for this is to first include everything with the '*' regular
expression and then exclude fields matching Metadata_*.

true

true
domain0

true
WARNING

rti_recorder_default.dat

Metadata_*

4-20

Database (Output File) Properties

Metadata_*

0
RTIDDS_DESERIALIZEMODE_ALWAYS

domain0

AllTopics

4-21

Domain Type Configuration

4.6

Domain Type Configuration
The top-level tag allows you to pass type configuration information to
the Record and Converter tools in the form of XML type-configuration files.
Table 4.6 describes the Domain Type Config properties. All Domain Type Config properties are

optional.
Table 4.6

Domain Type Configuration Properties

Properties under

Description

Domain Group Properties

domain_group

Table 4.7

Syntax

A list of type configuration elements
associated with specific Recorder
domain definitions (see Section 4.7).
These type configuration elements can
be repeated.
For more details in the contents of this
tag See Table 4.7, “Domain Group
Configuration Properties”.

Domain Group Configuration Properties

Properties
under

domain_filter

type_config

Syntax

Description

POSIX fn expression

A list of POSIX expressions that specify the names of
the Recorder domain definitions for which the type
definitions specified will apply. The element tag can
be repeated. The POSIX expressions have to match the
name attribute of any of the domain definitions as
defined in Section 4.7.

XML Properties

Specific XML type configuration properties for the
domains specified by the list of elements in the
domain_filter tag seen above.
See Table 4.8, “Type Config Properties”.

Table 4.8

Type Config Properties

Properties
under

xml

Syntax

XML Type Configuration
Properties

Description
Allows you to specify XML type-configuration files,
the path in which to find them, and other properties
related to the registration of the types with the
DomainParticipants.
See Table 4.9, “XML Type Configuration Properties”.

4-22

Domain Type Configuration

Table 4.9

XML Type Configuration Properties

Properties
under

Syntax

Description

file_group

File Group Properties

Allows you to specify XML file groups; each of these
files contain type definitions to be used by the Record
tool. The element tag can be repeated.
See Table 4.10, “File Group Properties”.

max_sequence

Integer

max_string

Integer

The default sequence size, in case there are
unbounded sequences in the type definitions specified
in the any of the files specified in any of the file
groups.
The default string size, in case there are unbounded
strings in the type definitions specified in any of the
files specified in any of the file groups.

path

Path

A list of the paths (relative or absolute) to be used
when searching for the XML type definition files. The
tag can be repeated.

register_
top_level

Boolean

Whether or not to register the top-level types found in
the type definitions with their canonical names. Do
this with any of the files defined in any of the file
groups. Default: TRUE.

Table 4.10 File Group Properties

Properties
under

Syntax

Description

file_name

File Name

A list of file name strings, specifying files containing the
type definitions. The element tag can be repeated.

max_sequence

Integer

max_string

Integer

The default sequence size in case there are unbounded
sequences in the type definitions specified in the any of
the files specified in this specific file group.
The default string size in case there are unbounded
strings in the type definitions specified in any of the files
specified in this specific file group .

4-23

Domain Properties

Table 4.10 File Group Properties

Properties
under

register_
top_level

Syntax

Boolean

Description
Whether or not to register the top-level types found in the
type definitions with their canonical names. Do this with
any of the files found in this specific file group. The value
of this setting overrides the value of the upper-level setting (see Table 4.9, “XML Type Configuration Properties”). Default: TRUE.

Type Registration

type

Properties

Specific type properties. A list of type registration properties used to define how a type found in the files has to be
registered by the Record tool in the DomainParticipants.
See Table 4.11, “Type Registration Properties”

Table 4.11 Type Registration Properties

Properties under

Syntax

register_top_level

Boolean

Description
Whether or not to register this type's canonical
name (as defined in tag type_name) with the
DomainParticipant. This name is registered
alongside any of the registration type names
defined in tag registered_type_name above.
Default: TRUE.

registered_
type_name

Registration
name string

When registering the type with the DomainParticipant, this setting defines a list of names to register the type with. The element tag can be
repeated.

topics

Topic name string

type_name

Type Name string

4.7

This is a list of regular POSIX fn-name expressions (every entry is equivalent to an
expression). A topic whose name matches any of
the expressions will be recorded using the type
definition designated by . If a topic
matches more than one expression in different
entries, the first one that was matched
will be used.
A string representing the canonical type name as
defined in the XML type definition for the type.

Domain Properties
Table 4.12 describes the Domain properties. All Domain properties are optional.

4-24

Domain Properties

Domain properties must be specified inside and tags. If
you want to use a RecordGroup (Section 4.9), you must assign a domain name with these tags,
even if you do not specify any domain properties (because the domain name is needed in the
RecordGroup’s domain_ref property).
Table 4.12 Domain Properties

Properties
under

Syntax

Description
Determines how topic data is stored in a database (serialized
or deserialized).
The following values are allowed:

deserialize_
mode

DDS_Enum

domain_id

DDS_Long

• RTIDDS_DESERIALIZEMODE_AUTOMATIC
Deserialize data if possible, otherwise store data in serialized format.
• RTIDDS_DESERIALIZEMODE_NEVER
Do not deserialize the data; store data in serialized format.
• RTIDDS_DESERIALIZEMODE_ALWAYS
Only store data if it can be deserialized first.
Default: RTIDDS_DESERIALIZEMODE_NEVER
See Recording Large User Data Types (Section 4.7.2).
Sets the domain ID.
Default: 0
Configures the DomainParticipant’s QoS policies.

participant_qos

DDS_QosPolicy

Default: default DomainParticipant QoS settings
See the RTI Connext DDS Core Libraries User’s Manual for
details. (See the chapter on Configuring QoS with XML.)
See the Notes: below for more information.

You may specify more than one Domain. Each one must have a unique name, with its own
and tags.
For example, the following creates a Domain named “mydomain” using domain ID 68. The data
will be recorded in serialized format. The DomainParticipant will use default QoS settings,
except for the Discovery QoS policy’s accept_unknown_peers field:

RTIDDS_DESERIALZEMODE_NEVER

false

Notes:

❏ If you do not set the property in the settings, the
Record tool will automatically build a participant name and set it using the prefix "RTI
Recorder: ". This is for compatibility with RTI Administration Console. If this property is

4-25

Domain Properties

changed, the Record tool won't override the property, but compatibility between the
Record tool and RTI Administration Console will be broken if the participant name is not
prefixed with "RTI Recorder: " (notice the space after the colon).

❏ Transports are configured through the Property QoS under the tag.
4.7.1

Enabling Monitoring Library in the Record Tool
RTI Monitoring Library enables Connext DDS applications to provide monitoring data. The monitoring data can be visualized with RTI Monitor, a separate GUI application that can run on the
same host as Monitoring Library or on a different host. Recording Service is statically linked to
Monitoring Library (you do not have to install it separately).
To enable monitoring in the Record tool, modify the participants’ QoS in the XML configuration
to include the rti.monitor.library property with a value of rtimonitoring. For example:

rti.monitor.library
rtimonitoring
false

See also: Enabling Monitoring Library with Replay (Section 7.4.1).

4.7.2

Recording Large User Data Types
When the Record tool records serialized user data, each primitive type in the topic’s data structure will have its own column in the table. The maximum number of columns is approximately
5,050.
Therefore, if you have a data-type that would require more than 5,050 columns, you must set the
deserialize_mode property to RTIDDS_DESERIALIZEMODE_NEVER. (Disregarding this limit
will cause recording to fail.)
Note: Each primitive type is considered a column. For example, the following would require 3,000
columns:
long Array[3000];

As another example, the following
y[0].x.a,y[0].x.b,y[1].x.a,y[1].x.b, etc.

would

require

separate

columns

struct X {
long a;
long b;
};
struct Y {
X x;
};
struct Z {
Y y[10];
}

4-26

for

TopicGroup Properties

4.8

TopicGroup Properties
A TopicGroup is an optional logical collection of Topics. If you are not going to have a
RecordGroup in the configuration file, you do not need a TopicGroup. (See Section 4.9.)
Table 4.13 describes the TopicGroup properties.
TopicGroup properties must be specified inside and
tags.
The following properties are required:

❏ field_expr
❏ shared_table
❏ topics
Table 4.13 TopicGroup Properties

Properties
under

auto_detect_
reliability

compact_char_
array

compact_char_
sequence

compact_
octet_array

compact_
octet_sequence

Syntax

DDS_Boolean

Description
If set to true, use the same reliability as the Publisher
of the matched Topic.
Default: False
Store array of char in a single column. The default
(true) saves the most space. While it is possible to
store individual elements in separate columns, it is
not recommended as the number of columns stored
can become very large.
Default: True
Store sequence of char in a single column. The
default (true) saves the most space. While it is possible to store individual elements in separate columns,
it is not recommended as the number of columns
stored can become very large.
Default: True
Store array of octet in a single column. The default
(true) saves the most space. While it is possible to
store individual elements in separate columns, it is
not recommended as the number of columns stored
can become very large.
Default: True
Store sequence of octet in a single column. The
default (true) saves the most space. While it is possible to store individual elements in separate columns,
it is not recommended as the number of columns
stored can become very large.
Default: True

4-27

TopicGroup Properties

Table 4.13 TopicGroup Properties

Properties
under

Syntax

Description
Specifies the QoS settings for all DataReaders created
for this TopicGroup.

datareader_qos

DDS_DataReaderQos

A DataReader is created for each discovered Topic
that matches topic_expr. All the DataReaders for the
TopicGroup will use the same set of QoS policies.
You can specify all of the QoS policies with the
datareader_qos property.
See the RTI Connext DDS Core Libraries User’s Manual
for more information. (See the chapter on
Configuring QoS with XML.)

exemption

field_expr

POSIX fn expressions

Specifies a comma-separated list of expressions that
should not be recorded.
Default: Nothing is exempt
Required.
A list of comma-separated POSIX expressions that
specify which fields in the Topics to record. (The
Topics are specified with , see Table 4.14,
“Topics Properties”.)
If set to ‘*’, everything is recorded.
This parameter is ignored when recording serialized
data.

include_
meta_columns

DDS_Boolean

In the database, every sample is stored alongside all
its sample information. If this property is set to
FALSE, the sample is stored in a serialized way, with
no sample information attached to it.
Setting this to FALSE (not the default) saves storage
space. When set to FALSE, less columns are created
in the SQLite database. The columns in this database
are often filled with repetitive data. So this option
can save space and execution time when these
requirements are critical.
Default value: True

index

User Data Table fields

Allows you to create an index for the recorded database based on the fields provided as parameters. See
‘Create Index’ Syntax (Section 4.8.1) for details.

4-28

TopicGroup Properties

Table 4.13 TopicGroup Properties

Properties
under

Syntax

shared_table

DDS_Boolean

Description
Required.
Specifies whether the tables of recorded data are
shared or exclusive.
The Record tool stores Topic data in tables; those
tables can be Shared or Exclusive. An Exclusive table
means that each Topic recorded in a RecordGroup is
stored in its own table. The name of the table follows
the convention:
TopicName$RecordGroupName$DomainName
(You can change the $ separator with the
path_separator database property described in
Table 4.4.)
Thus, two topics with the same name but from two
different TopicGroups are stored in separate tables.
A shared table means that Topics with the same
name are stored in the same table, regardless of
where it was recorded from. In this case the table has
an additional column, table_prefix, which stores the
table prefix in the form:
RecordGroupName$DomainName.
Default: False (exclusive)

topics

POSIX fn expression

Required.
Specifies a topic expression and any exemptions to
that expression. See Table 4.14, “Topics Properties”.

Table 4.14 Topics Properties

Properties
under

Syntax

Description
Required.
A comma-separated list of POSIX expressions that specify the
names of Topics to be included in the TopicGroup.
The syntax and semantics are the same as for Partition
matching.
Default: Null

topic_expr

POSIX fn expression

Note: Keep in mind that spaces are valid first characters in topic
names, thus they can affect the matching process. For example,
this will match both Triangle and Square topics (notice there is
no space before Square):
Triangle,Square
However the following will only match Triangle topics
(because there is a space before Square):
Triangle, Square

4-29

TopicGroup Properties

For example, the following creates a TopicGroup called AllTopics, which will include all discovered Topics. From those Topics, all fields will be recorded. This example does not specify the
optional datareader_qos property, so it will use default DataReader QoS settings:

This next example creates a TopicGroup called ColorsOfSquares that will only include Topics
named “Square.” For the recorded Topics, only the “color” field will be recorded. The DataReaders for the matching Topics will have default QoS settings, except that the Reliability QoS’s kind
will be DDS_RELIABLE_RELIABILITY_QOS:

Square

color

DDS_RELIABLE_RELIABILITY_QOS

The following example creates a TopicGroup called AllMinusCircleAndSquare that will include
all Topics except “Circle” and “Square.” For the recorded Topics, all fields will be recorded:

*
Circle, Square

Notes:

❏ Topics are never removed from a TopicGroup. The resources used to create DataReaders
for discovered Topics are not released if/when the Topics are deleted.

❏ The Record tool will ignore Topics published with a type that conflicts with an already
discovered type.

4.8.1

‘Create Index’ Syntax
SQLite indexes may improve performance on certain SQL queries. The Replay tool creates an
index based on the reception timestamp when opening the database, if it did not already exist
(see Performance and Indexing (Section 6.4)). The process of building the index, however, may
take some time for large tables. But if it is important for you to avoid spending this time, or if
indexing needs to happen during the recording, this is an easy way to create customized indexes
while still recording data into the database.
Note: While other applications accessing the database may benefit from online indexing, the
Record tool’s performance may drop because of it, so you should consider this when using
online indexing.

4-30

TopicGroup Properties

The Record tool's ability to create and store indexes for the User Data Table is controlled by the
property under (see Table 4.13, “TopicGroup Properties”). The
property expects one or more tags, which create the indexes.
Syntax

The can contain as many indexes as needed (for some notes about indexes and
performance see Indexing and Performance in SQLite: Tips and Tricks (Section 4.8.2)). Each of
those indexes can be built using one or more fields. The fields listed could be meta-data fields
(such as the reception timestamp used by the Replay tool) or user-data ones. In addition, you
may optionally add a prefix to the index. The general syntax is:

...

One_field
Another_field
...

...

The above configuration will create an index named OptionalPrefix$TableName, where $ is the
default path-separator (although it could be replaced using the property) and
TableName is the name of the user-data table (see Table 4.4, “Database Properties” for details).
There are some details to consider when creating an index:

❏ An index can only be created based on columns that actually exist in the user-data table.
So if some or all of the columns specified in the index configuration are excluded (see
Choosing Which SampleInfo and Discovery Fields to Record (Section 4.5.1)), they cannot
be used to build an index.

❏ If the index cannot be created, a warning is shown and the Record tool will continue its
execution.

❏ Two indexes cannot share the same prefix. If that happens, only the first one is created.
❏ If no prefix is provided, the string "index" and a ordinal number are added before the
table name. For instance, if two indexes are created for a given table, their names will be
index1$TableName and index2$TableName.
Examples:

To improve the Replay tool's initialization performance, we can create an index using the field
SampleInfo_reception_timestamp like this:

SampleInfo_reception_timestamp

Note that this is the same index that would be created by the Replay tool. (This index corresponds to the one created by the old database property , which is no longer supported.)
User-data fields can also be used to create indexes. To add a user-data field to an index configuration, you must use the fully qualified name of the field. This notation is similar to the notation
used to access data fields in filter expressions for content-filtered topics (see RTI Connext DDS
Core Libraries User’s Manual for more details). Fields inside container types are accessed using a
period character '.' (e.g., field1.nestedField2.nestedField3...). Fields in collections (arrays and
sequences) can be accessed using square brackets ('[]') notation (e.g., collectionField[1].nested-

4-31

TopicGroup Properties

Field, matrixField[0][0].field1). Only primitive fields can be used for indexing, because the
Record tool does not store non-primitive fields.
For example, consider a race scenario. We have a collection of ten runners and their order in the
race is recorded with the Record tool. The data types are defined using IDL as follows:
struct Runner {
int id; //@key
char * name; //@key
}
struct Race {
struct Runner runnerList[10];
TimeStamp currentTime;
}

If other applications are trying to access the recorded data in order to analyze it, we may want to
speed up data retrieval for these applications by indexing the above data in many ways. For
instance, we could build an index based on the ID and name of the first runner. We may also
need an index based on the current time and the reception timestamp. The XML configuration to
create these indexes is:

...

runnerList[0].id
runnerList[0].name

SampleInfo_reception_timestamp
currentTime

...

The above configuration creates two indexes:

❏ firstRunnerIndex$TableName, which indexes on the first runner’s id and name.
❏ index1$TableName, which indexes on the SampleInfo_reception_timestamp field and
user-data currentTime field.

4.8.2

Indexing and Performance in SQLite: Tips and Tricks
Online indexing can affect the permanence of the insertions the Record tool makes in the database, because of the extra work to maintain the indexes. However, in situations where more than
one application besides the Record tool will be accessing the database, online indexing may
make a difference, especially for applications retrieving data from the database.
It is important that indexes and the applications using them are as efficient as possible. Here are
some tips and tricks about indexing and efficient usage of indexes in SQLite.

4.8.2.1

Small Size Data Fields Work Best as Index Columns

The fewer bytes used in every index entry, the better for performance. SQLite stores indexes as
B-Trees. Every tree node uses a page of the database file. If the index does not fit in one page,
overflow pages are created. Keeping the index in one page may make a difference in terms of
performance. Therefore, large BLOBS, compact octet or char arrays, etc. should be avoided as

4-32

TopicGroup Properties

index columns. If you need to index on one of these types, consider building a hash key or similar compact representation and use that for indexing instead. The space overhead may be worth
the performance improvement.
4.8.2.2

Use Multi-Column Indexes where Possible

SQLite can use indexes that are created by indexing more than one column. However, SQLite
only allows a single index per table within a simple SQL statement. For UPDATE and DELETE
statements, this translates into only allowing a single index, since these statements can only
work with one table at a time.
Suppose we have a table with columns C 1, C 2, ..., C n and we want to access the table with a
SELECT statement with columns C 1, C 2, ..., C i, i < n, in the WHERE clause, such as "... WHERE
C 1 = value 1 AND C 2 = value 2 ... AND C i = value i.” SQLite will be able to use the index provided for each of the terms in the WHERE clause. However, if we have individual indexes for
each column, SQLite can only access one of the indexes for the SELECT statement and will only
benefit from the use of indexes for one of the search terms.
4.8.2.3

Avoid Excessive Indexes

Each index added to the database has a performance cost:

❏ Every additional index takes up some additional space in the database.
❏ INSERT and UPDATE statements have to modify both the original table and every index
associated with it. The performance of INSERT and UPDATE decreases linearly with the
number of indexes.

❏ Compilation of the SQL statements takes longer when there are more indexes for the
compiler to choose from.

❏ More indexes to choose from, bigger the chance the optimizer selects a suboptimal index
for a query.
Avoid redundant indexes where possible. If for a certain table we have index 1 indexing on columns (x,y) and index 2 indexing on columns (x,y,z) then index 1 is redundant and can be eliminated.
4.8.2.4

Use Indexes to Improve Performance of ORDER BY Clauses

To evaluate statements containing ORDER BY classes, SQLite first puts all the results of the
SELECT statement in a temporary table, then sorts the temporary table according to the ORDER
BY clause columns. This can be time-consuming. Whenever possible, SQLite tries to avoid storing and then sorting the result set; it does this by using an index that makes the results of the
SELECT statement appear in order. To force this to happen, provide the table with an index that
indexes all the columns in the ORDER BY clause.
For example, suppose we have a table with columns (x,y,z,a) and we want to execute the following query:
SELECT a FROM table ORDER BY x,y,z

Then providing an index in columns (x,y,z) will make SQLite avoid using a temporary table and
sorting all the result set.
The index can also be used to satisfy a WHERE clause and an ORDER BY clause in the same
statement. In the example above, suppose we want to execute this statement:
SELECT a FROM table WHERE x=value ORDER BY y,z

The above index on columns (x,y,z) will still be used by SQLite both to speed up the WHERE
clause and to avoid sorting the whole result set in the ORDER BY statement.

4-33

RecordGroup Properties

Nevertheless, care must be taken so that the columns in the index are all used in the SQL statement. For example, consider the following statement:
SELECT a FROM table WHERE x=value ORDER BY z

It does not use "y", which creates a gap in the index terms of our index. This will make the index
be used to speed up the WHERE clause, but won't avoid storing and sorting the whole result set
afterwards.
4.8.2.5

Terms Used in Inequality Expressions should be Placed Last in Index Column Lists

SQLite allows at most one index column to be used in statements containing inequality expressions (such as '<', '>', '<=' or '>='). Columns that are constrained by inequalities should be placed
as the right-most term of the index. For example, suppose if we have a table with columns
(x,y,z,a), an index in columns (x,y,z) and the following SQL statement:
SELECT a FROM table WHERE x = value x AND y < value y AND z = value z

The index does not help optimize term "z", because "y" is constrained by an inequality.
SQLite allows up to two inequality expressions for the same column, as long as they provide
upper and lower bounds for the column. So this statement will effectively use the index as the
inequality expressions for "y" provide an upper and lower bound for the term:
SELECT a FROM table WHERE x = value x AND y < value y1 AND y > value y2
4.8.2.6

Use Indexed Column Names in WHERE Clauses, not Expressions

SQLite uses indexes for a column when the column name appears isolated in statements, not in
more complex expressions. For example, if we have a table with column "x" and an index in "x",
the index will be used if the SQL statement includes "x" but not expressions based on "x", such as
these:
... WHERE x - value = 0
... WHERE x * 1 = ?
... WHERE +x = value

All the WHERE clauses above will cause a full-table scan and SQLite will not use the index to
optimize the query.

4.9

RecordGroup Properties
A RecordGroup is a binding between a TopicGroup and a Domain. It controls which TopicGroup members are recorded for each Domain. Any Topic that is part of a TopicGroup in the
RecordGroup is recorded from the specified Domains.
RecordGroups are optional. If you do not create one, the Record tool will not record any data. This
is useful if you want to start the Record tool in “stand-by” mode—then you can use remote
access (see Section 4.10) to switch to a different configuration file (one that does have a
RecordGroup) and start recording.
Table 4.15 describes the RecordGroup properties. Notice that domain_ref and topic_ref are
required.
RecordGroup properties must be specified inside and tags. The name that you assign (“String”) will be used in the table name(s) in the
database (output) file(s).

4-34

RecordGroup Properties

Table 4.15 RecordGroup Properties

Properties under

Syntax

domain_ref

StringSequence

subscriber_qos

DDS_SubscriberQos

topic_ref

StringSequence

field expression

...

user_topic_
metadata_fields

field expression

...

Description
Required.
Specifies a sequence of references to domains.
Default: Null
Configures the Subscriber used by the
RecordGroup.
Default: default Subscriber QoS settings
See the RTI Connext DDS Core Libraries User’s
Manual for details. (See the chapter on Configuring QoS with XML.)
Required.
Specifies a sequence of references to
TopicGroups.
Default: Null
Specifies lists of included and excluded Sample
Info or metadata fields in the user topic tables.
A field expression is a POSIX fnmatch
expression to match field names as described
in Appendix A (e.g. "SampleInfo_*" will match
all Sample Info specific fields). There is no
upper limit in the number of
expressions to be specified.
The ‘included’ and ‘excluded’ fields are
optional. If both are defined, the ‘included’
expressions are processed before the ‘excluded’
ones.
User topic field settings can also be expressed
in the Database settings (see Table 4.4).
However, Record Group settings take
precedence over general database settings
defined there.

For example, the following creates a RecordGroup called RecordAll, which will include all
members of TopicGroup All that are discovered on Domain MyDomain. This example does not
specify the optional subscriber_qos property, so it will use default Subscriber QoS settings:

AllTopics

MyDomain

Note: A RecordGroup can refer to multiple domains and multiple TopicGroups. However, a
RecordGroup will only record one of each matching Topic from a Domain. If multiple matches
occur, only the first one will be recorded. (If you need to record the same Topic from the same
domain using different QoS policies, you should use different TopicGroups and RecordGroups.)

4-35

Recording Service Integration with Extensible Types

4.10

Recording Service Integration with Extensible Types
Recording Service includes partial support for the "Extensible and Dynamic Topic Types for
DDS" specification from the Object Management Group (OMG)1. This section assumes that you
are familiar with Extensible Types and you have read the RTI Connext DDS Core Libraries Getting
Started Guide Addendum for Extensible Types.
This support allows systems to define data types in a more flexible way, and to evolve data
types over time without giving up portability, interoperability, or the expressiveness of the DDS
type system.
With Extensible Types, different type definitions for the same type name may be discovered by
the Record tool during execution. The tool’s default behavior is to register the first type definition (type code) found for a type.
You can learn more in the RTI Connext DDS Core Libraries Getting Started Guide Addendum for
Extensible Types.

❏ Recording Service can subscribe to topics associated with optional, final, mutable, and
extensible types.

❏ A topic “T” within a recording service domain () can be associated with at
most one version of a type. To record more than one version, you will have to use multiple recording service domains.

❏ Users can pre-register a set of types in a recording service domain by providing an XML
description of the types. The XML description supports structure inheritance and mutability (described in the RTI Connext DDS Core Libraries Getting Started Guide Addendum for
Extensible Types).

❏ The TypeConsistencyEnforcementQosPolicy can be specified on a per-topic-group
() basis, in the same way as other QoS policies. Conversions between
extensible and mutable types are allowed if the TypeConsistencyEnforcementQosPolicy’s kind is set to ALLOW_TYPE_COERCION.

4.10.1

Selecting a Type Version For a Topic “T” In a Recording Domain
A topic “T” within a recording service domain () can be associated with at most one
version of a type. By default, Recording Service will use the type associated with the first discovered DataWriter on topic “T”. This makes the selection of the type non-deterministic since it
depends on the order in which DataWriters are discovered. To resolve this problem, you can
provide the type in the XML configuration using the . For example:

domain0

ShapeType.xml

1. http://www.omg.org/spec/DDS-XTypes/

4-36

Recording Service Integration with Extensible Types

false
ShapeTypeExtended

ShapeType

Circle
Square
Triangle

With the above type configuration, Recording Service will register the type "ShapeTypeExtended"
with the DomainParticipant associated with “domain0”. When any of the topics in the
list is matched ("Circle", "Square" or "Triangle"), Recording Service will use "ShapeTypeExtended"
as the type for recording it. The name used for registering the type depends on the
flag. If this flag is set to true, the type name used for the registration will be
the canonical type name (in the example above, it would be "ShapeTypeExtended"). However, if
this flag is set to false, at least one name has to be provided for registration of the type using the
XML setting. The first name in this list will be the one used by Recording Service.
When Recording Service discovers the first DataWriter on topic “T”, it will first try to match the
topic name against the expressions in all the settings. If it finds a match, it will
use the associated type settings. Otherwise, Recording Service will try to create the topic using the
DataWriter’s registered type name. If the type name has not been registered with Recording Service ’s DomainParticipant yet, the new topic will be associated with the discovered type. If the
type name is already registered, the new topic will use the registered type.
4.10.1.1

Example

Let’s consider an example based on RTI Shapes Demo (described in the Recording Service Getting
Started Guide). You can use Shapes Demo to publish and/or subscribe to either the "ShapeType"
data type or the "ShapeExtended" data type. ShapeExtended includes the same data as ShapeType, plus two more fields: FillKind and Angle.
Suppose Shapes Demo is publishing the base ShapeType in a domain and the Record tool is subscribing to ShapeType in the same domain. The Record tool will register and record data of type
ShapeType. If a different instance of Shapes Demo—one that uses the ShapeExtended type—
starts publishing in the same domain, the Record tool will record new data from this second
instance of Shapes Demo. (Both types are compatible because one extends the other.) However,
the Record tool will only record the fields corresponding to the base ShapeType, because that is
the registered type. The Record tool will ignore the extra fields in the ShapeExtendedType (FillKind and Angle).

4-37

Recording Service Integration with Extensible Types

As described in Selecting a Type Version For a Topic “T” In a Recording Domain (Section 4.10.1),
this default behavior is non-deterministic regarding the types being recorded. That is, you have
no control over what type version will be recorded. To change this behavior, you can change the
Record tool’s configuration and provide the XML description of the desired type version via
XML. We illustrate this with the following example, whose objective is to make sure that the
Record tool records the type ShapeTypeExtended, no matter which types are being published. To
do so, the Record tool needs a file that defines the type to be recorded, such as the following file
named ShapeType.xml.

Using the following configuration, the Record tool will always record the ShapeTypeExtended
by using the XML type definition in ShapeType.xml. With the configuration below, no information from a DataWriter using ShapeTypeExtended will be lost.

shapes.dat

domain0

ShapeType.xml

ShapeTypeExtended

ShapeType

Square
Circle
Triangle

false

RTIDDS_DESERIALIZEMODE_ALWAYS

domain0

4-39

Recording Service Integration with Extensible Types

All

4.10.2

Recording Two Versions of a Type in Different Tables in Same Database
The following example shows how Recording Service can store samples of different type versions
it finds for a topic "T" for the same DDS domain ID, in different tables, and without any loss or
duplication of information (each type version is recorded completely, and no samples in one
table are present in a different one).
The type configuration settings in Recording Service allow the specification of name filters (regular POSIX fn-match expressions) which allow the specification of a subset of the Recording Service domains they will apply to. This allows the specification of different type selection settings
(as described inSelecting a Type Version For a Topic “T” In a Recording Domain (Section 4.10.1))
for different Recording Service domains.
With these tools and using RTI Shapes Demo as an example, we could define a domain where to
record the base type of Shapes (named "domain0Base") and another one to use for the extended
type ("domain0Extended") associated with the same domain ID:

RTIDDS_DESERIALIZEMODE_ALWAYS

By applying filters to the Domain Type Configuration settings, we can specify that domain
"domain0Base" should record topics "Circle", "Square" and "Triangle" with the base type in the
XML definitions, like this:

domain0Base

false

TestTypesLibrary.xml

true

4-40

Recording Service Integration with Extensible Types

ShapeType

Circle
Square
Triangle

The same way, we can associate the extended type for shapes with the extended domain:

domain0Extended

false

TestTypesLibrary.xml

false
ShapeTypeExtended

ShapeType

Circle
Square
Triangle

For the moment, we have defined the type to use with each Recording Service domain, but
because
the
default
value
for
the
TypeConsistencyEnforcementQosPolicy
is
ALLOW_TYPE_COERCION and both types are compatible, we would record each sample
twice: once for each type version. We need to set the DataReader QoS settings in the TopicGroups we create so that they will not allow type coercion:

4-41

Recording Service Integration with Extensible Types

Circle,Square,Triangle

DISALLOW_TYPE_COERCION

We also need to associate the RecordGroup we create with both Recording Service domains:

domain0Base
domain0Extended

Shapes

With all these settings together, Recording Service will create a different table for each type version of Shapes it finds when it discovers each of the Shapes Demo topics. For example, for topic
"Square", it would create table "Square$RecordGroup$domain0Base" where it would record
only
the
samples
published
with
the
base
type
version,
and
table
"Square$RecordGroup$domain0Extended", which would contain only the samples published
with the extended type version of Shapes.
Note: Because of a known issue1, for the Replay tool to work properly with the databases created
with the above settings (replaying different topics with their recorded version requires concurrent access to the database from different Replay database entitites), we need to enable the
flag in Recording Service so the index is created beforehand:

shapes.dat
true

1. RTI Issue ID RECORD-318

4-42

Chapter 5

Accessing the Record Tool from a Remote
Location

Perhaps you want to start/stop the Record tool from another machine, or even reconfigure it to
change what is being recorded. You can create a Connext DDS application that can remotely control the Record tool. This chapter explains how.
To control the Record tool from a remote location:
1. Configure the Record tool to allow remote access (see Recording Service Integration with
Extensible Types (Section 4.10).
2. Create a Connext DDS application using the provided rtirecord.idl file. You will use rtiddsgen to generate the basics and then add code to send your desired remote commands.

5.1

Overview
If the Record tool is configured to allow remote access, it creates a DataReader for a “command
request” Topic (named RTI_RECORDER_COMMAND_REQUEST_TOPIC) and a DataWriter for
“command
response
and
status”
Topic
(named
RTI_RECORDER_COMMAND_RESPONSE_TOPIC). So the Record tool will write status
updates and comman responses.
These topics’ types and names are specified in the IDL file, resource/idl/rtirecord.idl.
When the Record tool detects a remote DataReader and DataWriter of these special topics from
the same participant, the Record tool will be in a ‘connected’ state, which means it will accept
remote commands.
Your remote-access application will use the following constants:

❏ RTI_RECORDER_COMMAND_TYPE Register a type of this name, as seen in
Figure 5.1.

❏ RTI_RECORDER_COMMAND_REQUEST_TOPIC Create a DataWriter with this
Topic name, as seen in Figure 5.2.

❏ RTI_RECORDER_COMMAND_RESPONSE_TOPIC Create a DataReader with this
Topic name, as seen in Figure 5.2.
See Remote Control Messages (Section 5.3) for more information.

5-1

Establishing a Connection with the Record Tool

Figure 5.1

Registering the Message Type
RTIRemoteCtxMsgTypeSupport_register_type
(self->dds_participant, RTI_RECORDER_COMMAND_TYPE);

5.2

Establishing a Connection with the Record Tool
To establish a connection with the Record tool, your remote-access application needs:

❏ 2 Topics (one for commands, one for status and command responses)
❏ 1 DataReader
❏ 1 DataWriter
When creating the DataReader and DataWriter, use the following QoS settings:

❏ history.kind = DDS_KEEP_ALL_HISTORY_QOS
❏ reliability.kind = DDS_RELIABLE_RELIABILITY_QOS

5-2

Establishing a Connection with the Record Tool

Figure 5.2 shows how to create the Entities in your remote-control application using the C API.
(A general knowledge of Connext DDS is assumed.)
Figure 5.2

Creating the Required Entities
struct DDS_TopicQos tqos = DDS_TopicQos_INITIALIZER;
DDS_Topic * dds_topic_cmd = NULL;
DDS_Topic * dds_topic_status = NULL;
struct DDS_DataWriterQos wqos = DDS_DataWriterQos_INITIALIZER;
RTIRecorderAdminMessageDataWriter * dds_writer = NULL;
struct DDS_DataReaderQos rqos = DDS_DataReaderQos_INITIALIZER;
RTIRecorderAdminMessageDataReader * dds_reader = NULL;
...
dds_topic_cmd =
DDS_DomainParticipant_create_topic(
dds_participant, RTI_RECORDER_COMMAND_REQUEST_TOPIC,
RTI_RECORDER_COMMAND_TYPE, &tqos,
NULL, DDS_STATUS_MASK_NONE);
dds_topic_status =
DDS_DomainParticipant_create_topic(
dds_participant, RTI_RECORDER_COMMAND_RESPONSE_TOPIC,
RTI_RECORDER_COMMAND_TYPE, &tqos,
NULL, DDS_STATUS_MASK_NONE);
...
rqos.reliability.kind = DDS_RELIABLE_RELIABILITY_QOS;
rqos.history.kind = DDS_KEEP_ALL_HISTORY_QOS;
dds_reader = (RTIRemoteCtxMsgDataReader*)
DDS_Subscriber_create_datareader(
dds_subscriber, DDS_Topic_as_topicdescription(
dds_topic_status), &rqos,
NONE, DDS_STATUS_MASK_NONE);
wqos.reliability.kind = DDS_RELIABLE_RELIABILITY_QOS;
wqos.history.kind = DDS_KEEP_ALL_HISTORY_QOS;
dds_writer = (RTIRemoteCtxMsgDataWriter*)
DDS_Publisher_create_datawriter(
dds_publisher, dds_topic_cmd, &wqos,
NULL, DDS_STATUS_MASK_NONE);

5-3

Remote Control Messages

5.3

Remote Control Messages
The Record tool exchanges messages with your remote-access application by publishing and
subscribing to two special remote-access topics. Both topics use the same message format,
shown in Figure 5.3.

Figure 5.3

Top Level Structure for Remote Control Messages
struct RTIRecorderAdminMessage {
long destination_mask;
RTIRemoteAdminAddress destination;
long msg_id;
RTIRecorderAdminUnion msg;
}; //@Extensibility MUTABLE_EXTENSIBILITY

For complete details, see the IDL file, rtirecord.idl in the examples directory.
destination_mask — A field used by other RTI tools; can be ignored.
destination — The Record tool application for which the message is intended. If

accept_broadcast_commands is turned off, this structure must match that of the Record
tool. If accept_broadcast_commands is turned on, this structure can be a specific destination or all 0's.
msg_id — A user-specified integer that identifies a particular message exchange. When the

Record tool sends a response to a command, it will include the same msg_id that was
received in the command.
msg — A union of different message types. The discriminator must be set to one of the message

types listed in Table 5.1.
The code fragment in Figure 5.4 shows how to set the message type in the remote-access application.
Depending on the message type, the correct union member must also be filled in. For example,
Figure 5.5 shows how to construct a message to the Record tool to read a new configuration from
a file. In this example, the new configuration is to be read from a file on the same file-system as
the Record tool.
Figure 5.4

Assigning a Message Type (C Language)
RTIRecorderAdminMessage * msg = NULL;
...
msg = RTIRecorderAdminMessagePlugin_create_sample();
/* Handle creation errors */
if (msg == NULL) {
...
}
msg->msg._d = RTI_RECORDER_ADMIN_START;

5-4

Remote Control Messages

Figure 5.5

Sending a Command to the Record Tool to Read a New Configuration File
RTIRecorderAdminMessage * msg = NULL;
DDS_ReturnCode_t retcode;
struct DDS_SampleInfo info;
msg = RTIRemoteCtxMsgPlugin_create_sample();
/* Handle creation errors */
if (msg == NULL) {
...
}
msg->msg._d = RTI_RECORDER_ADMIN_CONFIGURE;
/* This is the last part of the configuration. If the
* configuration spans multiple samples, then only the last
* one should have this set to TRUE */
msg->msg._u.config.final_config = DDS_BOOLEAN_TRUE;
/* Tell the Record Tool that the filename to read from follows
* in the config_from_string text string */
msg->msg._u.config.config_from_file = DDS_BOOLEAN_TRUE;
/* Copy the name of file that the Record tool shall read from */
strncpy(msg->msg._u.config.config_from_string, filename,
RTI_RECORDER_CONFIG_MAX_STRING);
/* Copy the configuration name of the tag to load */
strncpy(msg->msg._u.config.config_name, cfgname,
RTI_RECORDER_CONFIG_MAX_STRING);
/* Send the configuration message to the Record tool.
* (dds_writer has been created elsewhere) */
retcode = RTIRecorderAdminMessageDataWriter_write(dds_writer,
msg, &DDS_HANDLE_NIL);
/* check for errors here ... */
while (no response) {
retcode = RTIRecorderAdminMessageDataWriter_read_next_sample(
dds_reader, msg, &info, &DDS_HANDLE_NIL);
if (retcode != DDS_RETCODE_NO_DATA) {
/* response received */
}
sleep(1);
}

5.3.1

Updating the Record Tool’s Partition QoS Policy
For each Record Group in the XML configuration, Recording Service creates a DDS Subscriber.
There is one Subscriber in each of the Record Group’s domains. Sometimes it may be useful to
change the Partition QoS settings for some of these Subscribers.

5-5

Remote Control Messages

Table 5.1

Messages Types Exchanged Between Record Tool and Remote Access Application

Direction

Message Type
(add prefix:
RTI_RECORDER_
ADMIN)

Description

RECORDER_START

Instructs the Record tool to start recording.

RECORDER_STOP

Instructs the Record tool to stop recording.
Instructs the Record tool to reconfigure according to the
contents of the message.
Stop the Record tool before sending this message:

RECORDER_
CONFIGURE

From:
RECORDER_
Your Connext DDS
SHUTDOWN
Remote-Control
Application
RECORDER_ADD
To:
The Record tool
RECORDER_DELETE

If the Record tool has already been stopped, it will read the
new configuration and restart. It will not automatically
start recording unless auto_start (see Table 4.2, “General
Properties”) is true (the default case).
If the Record tool has not already been stopped, an error is
returned.
Instructs the Record tool to shutdown and exit.
Instructs the Record tool to add entities based on the contents of the message.
Instructs the Record tool to delete entities based on the contents of the message.

RECORDER_PAUSE

Instructs the Record tool to pause entities based on the contents of the message.

RECORDER_
PARTITION

Updates the Partition QoS in all or some of the Record tool’s
Record Groups (in their associated DDS Subscribers). See
Updating the Record Tool’s Partition QoS Policy (Section
5.3.1).

RECORDER_RESUME

Instructs the Record tool to resume recording of previously
paused entities based on the contents of the message.

RECORDER_PING

Instructs the Record tool to send the recording model a to
the Remote Control application.

To:
RECORDER_INFO
Your Connext DDS
Remote-Control
Application
RECORDER_
From:
RESPONSE
The Record tool

When the Record tool publishes statistics, it periodically
sends out this message type.
Indicates that this message is a response to a command.

a. The recording model is an XML representation of two aspects of the Record tool: (1) The configuration model: the XML
configuration (similar to the XML configuration file used to configure the Record tool.) and (2) The run-time model: an
XML description of the entities that have been created based on the configuration. Note that only a a minimal model is
returned; the QoS are not returned.

For instance, suppose the Record tool needs to record data from producers that are organized
into different partitions based on their geographical distribution (different locations are represented by different Partition strings). It is possible to specify a Partition QoS policy via XML configuration when the Record tool starts up. But suppose the data producers change location—and
thus change their Partition QoS? The Record tool won't be able to keep track of the data without
changing its Partition QoS too. (For details on the Partition QoS Policy, see the RTI Connext DDS
Core Libraries User’s Manual.)
You can change the Partition QoS for any or all Record Groups by using the
RTI_RECORDER_ADMIN_PARTITION message type. This command needs the following
information:

5-6

Remote Control Messages

❏ A collection of partition strings to be applied. As described in the RTI Connext DDS Core
Libraries User’s Manual, the partitions can be POSIX regular expressions. If no partition
string is specified, the default partition (empty) is applied.

❏ A collection of Record Group names (or regular expressions matching Record Group
names) for which the partitions will be applied. The partitions will be applied to all the
DDS Subscribers created for the matching Record Groups. If no Record Group string is
specified, the partitions are applied to all the Record Groups in the Record tool’s configuration.
The partition update command is absolute, meaning that current partitions that are active for
the matching Record Groups will be discarded and the new ones will be applied instead. For
example, if a Record Group is already joined to partition "A", in order to have the Record Group
join "A" and "B," the command will need to specify both "A" and "B" in their partition string collection.
Partition updates are applied per Record Group. If a Record Group is linked with multiple
DomainParticipants, the Partition QoS policy will be updated for each Subscriber created for
each of the Record Group’s associated DomainParticipants. For example, consider the following
configuration:

...

Domain0
Domain1

...

With the above configuration, there will be two DDS DomainParticipants, "Domain0" and
"Domain1." Each of these DomainParticipants will contain a Subscriber associated with the
Record Group called "RecordGroup." Sending a partition update command specifying
"RecordGroup" will affect the Subscribers in both DomainParticipants. If you have a situation in
which different partition changes need to be specified for each Subscriber, you should specify
separate Record Groups in the Record tool’s configuration. For the above example, this would
mean splitting the Record Group in two, like this:

...

5-7

Remote Control Messages

...

Domain0

...

Domain1

...

This way, partition update commands can modify separate Partition QoS settings for each
Record Group.
The Partition QoS Policy establishes limits on the length of the strings provided in the partition
string collection. There can be up to 64 strings, with a maximum of 256 characters summed
across all strings. When the Record tool receives a partition update command, it checks this limit
by checking the length of all the partition strings in the command. If the summed length of the
strings is larger than the limit, the partitions won't be applied and an error response will be sent
back to the command issuer.
Note: For DDS Subscribers to match DDS Publishers, either of the two can use regular expressions as their partition strings, but not both at the same time. A regular expression partition has
to be matched against a specific name partition string. It is the user's responsibility to ensure that
at least one of the sides in the communication (Publisher or Subscriber) uses a pecific name partition string as the partition.
Any failure while processing the command will result in an error response from the Record tool
to the command issuer. However, after validating the partition strings passed in the command,
the processing is completed for all the Record Groups, even if an error happens while processing
any of the QoS updates for any of them.
Depending on the status of the Record tool when the partition update command is issued, the
following can happen:

❏ If the Record tool is running, the command will update the current DDS Subscribers associated with matching Record Groups. The same will happen when the Record tool is idle
(paused).

❏ If the Record tool is stopped, the command will update the stored QoS settings for the
Record Groups. Because the DDS subscriptions haven't been created yet when Recorder
is stopped, the Partition QoS settings that were updated will take effect when the Record
tool is started. The newly created DDS Subscribers will use the new Partition QoS settings as received in the partition update command.
Note: Partition updates are not instantaneous. Thus, the Record tool may lose samples while the
partition changes take effect. This behavior is not different than that of any other Connext DDS
application.

5-8

Remote Control Messages

5.3.1.1

Sending a Partition Update Command to the Record Tool

The following is the IDL definition of the type used to specify the information in a Partition
Update command:
const long PARTITION_MAX_LENGTH = 256;
const long MAX_PARTITIONS = 64;
const long MAX_RECORD_GROUP_EXPRESSIONS = 256;
const long RECORD_GROUP_EXPRESSION_MAX_LENGTH = 256;
struct RTIRecorderPartitionMessage {
sequence, MAX_PARTITIONS> partitions;
sequence,
MAX_RECORD_GROUP_EXPRESSIONS> recordGroupExpressions;
};

A maximum of 64 (MAX_PARTITIONS) partition strings can be specified, with a maximum
length of 256 characters (PARTITION_MAX_LENGTH). In a similar way, a maximum of 256
expressions to match Record Group names (MAX_RECORD_GROUP_EXPRESSIONS) can be
specified,
with
a
maximum
length
of
256
characters
(RECORD_GROUP_EXPRESSION_MAX_LENGTH).
The following code snippet in C shows how to prepare and send a partition update command to
the Record tool so that all the Record Groups in the configuration join partition "A":
RTIRemoteCtxMsgDataWriter * messageWriter = NULL;
RTIRemoteCtxMsg * message = NULL;
DDS_ReturnCode_t retcode;
/* Initialize the DDS entities used for the communication */
...
/* Create the message instance to be written to the tool */
message = RTIRemoteCtxMsgTypeSupport_create_data();
if (message == NULL) {
/* Error handling */
}
/* Set the destination information for broadcast commands
* (all zeroes)
* Set the destination mask as empty too */
message->destination.app_id = 0;
message->destination.host_id = 0;
message->destination.instance_id = 0;
message->destination_mask = 0;
/* Set the type of the message to be a partition message */
message->msg._d = RTI_REMOTECTX_MSG_RECORDER_PARTITION;
/* Set the message ID; here we set it to zero for simplicity,
* but this value should be auto-incremented with every new command
* sent to the tool in order to correlate the commands with their
* responses */
message->msg_id = 0;
/*
*
*
if

To send a partition update command for all Record Groups in a
Recording configuration, set the collection of Record Group
names (expressions) to be empty */
(!DDS_StringSeq_ensure_length(
&message->msg._u.partitions.recordGroupExpressions, 0, 0)) {
/* Error handling */

5-9

Remote Control Messages

}
/* Ensure enough space to add the "A" partition string to the
* collection of partitions*/
if (!DDS_StringSeq_ensure_length(
&message->msg._u.partitions.partitions, 1, 1)) {
/* Error handling */
}
/* Copy the string ("A") into the first string in the partition
* string collection. We've used 'strncpy' in conjunction with the
* maximum length of a partition string for correctness, even
* though we know the length of string "A" fits the maximum length
* of a partition string */
strncpy(DDS_StringSeq_get(&message->msg._u.partitions.partitions,
0), "A", PARTITION_MAX_LENGTH);
/* Write the message; recall that the 'accept_broadcast_commands'
* must be enabled in the Record tool to be able to read this
* command; if we know the destination information for the Record
* tool (app id, host id and instance id), we can use this
* information and don't have to enable 'accept_broadcast_commands'
*/
retcode = RTIRemoteCtxMsgDataWriter_write(
RTIRemoteCtxMsg_writer, message,&DDS_HANDLE_NIL);
if (retcode != DDS_RETCODE_OK) {
/* Error handling */
}

If we wanted to include a Record Group filter in our command, the filtering expression would
have to be included in the collection of Record Group expressions. For example, to apply a partition update only to Record Groups ending with the suffix "Domain1," we could use the regular
expression "*Domain1" and use the following code (in C language):
RTIRemoteCtxMsgDataWriter * messageWriter = NULL;
RTIRemoteCtxMsg * message = NULL;
DDS_ReturnCode_t retcode;
/* Initialize the DDS entities used for the communication */
...
/* Create the message instance to be written to the tool */
message = RTIRemoteCtxMsgTypeSupport_create_data();
if (message == NULL) {
/* Error handling */
}
...
/* To send a partition update command to Record Groups suffixed
* with "Domain1," use the regular expression "*Domain1"; we need
* to ensure enough space in the record group expressions
* collection to hold one entry */
if (!DDS_StringSeq_ensure_length(
&message->msg._u.partitions.recordGroupExpressions, 1, 1)) {
/* Error handling */
}
/* Copy the string ("*Domain1") into the first string in the
* record group expression string collection. We used 'strncpy' in
* conjunction with the maximum length of a record group expression
* string for correctness, even though we know the length of string
* "*Domain1" fits the maximum length of a record group expression

5-10

Using the Example Remote-Access Application—Record Shell

* string */
strncpy(DDS_StringSeq_get(
&message->msg._u.partitions.recordGroupExpressions, 0),
"*Domain1",
RECORD_GROUP_EXPRESSION_MAX_LENGTH);
/*
*
*
if

Set the length of the partition string collection to be zero;
The Record tool will set the default partition (empty) in the
matching Record Groups */
(!DDS_StringSeq_ensure_length(
&message->msg._u.partitions.partitions, 0, 0)) {
/* Error handling */

}
/* Write the message */
retcode = RTIRemoteCtxMsgDataWriter_write(
RTIRemoteCtxMsg_writer, message, &DDS_HANDLE_NIL);
if (retcode != DDS_RETCODE_OK) {
/* Error handling */
}

The example above uses an empty collection of partition strings. This means that the Record tool
will apply the default partition (empty) to the Record Groups that match the regular expression
"*Domain1" specified in the command.

5.4

Using the Example Remote-Access Application—Record Shell
The Record Shell is a Connext DDS application that can remotely control (start, stop, and reconfigure) the Record tool. The Record Shell is not meant as a complete solution to remotely controlling
the Record tool. Its purpose is just to give you an idea of what can be done.
To start the Record Shell, open a command prompt1change to the /bin directory,
and enter:

❏ On Linux and Mac OS X systems:
rtirecsh -domain

❏ On Windows systems:
rtirecsh.bat -domain

Table 5.2 lists the command-line options you can use when starting the Record Shell. Once it is
running, you can use the commands described in Record Shell’s Commands (Section 5.4.1).
Table 5.2

Record Shell’s Command-Line Options

Command-line Option

Description

-domain

Required.
Specifies the domain ID (an integer between 0 and 99).

-partition

Specifies an optional, comma-separated list of partition names. This option is
necessary if the Record tool is configured to enable remote access in a particular partition.

1. On Windows systems: from the Start menu, select Accessories, Command Prompt.

5-11

Using the Example Remote-Access Application—Record Shell

Table 5.2

Record Shell’s Command-Line Options

Command-line Option

5.4.1

Description

-noUdpv4

Disables the UPDv4 transport.

-updv6

Enables the UDPv6 transport.

-noShmem

Disables the shared memory transport.

-noMulticast

Disables multicast.

-verbosity

The verbosity is a bit-map that specifies what type of logging information
should be printed.
The verbosity may be:
0 — No messages
1 — Exceptions (default)
2 — Warnings
4 — Information
7 — All types

-help

Prints version information and a list of options.

Record Shell’s Commands

❏
❏
❏
❏
❏
❏
❏
❏
❏
❏
❏

add (Section 5.4.1.1)
configure (Section 5.4.1.2)
delete (Section 5.4.1.3)
exit (Section 5.4.1.4)
info (Section 5.4.1.5)
pause (Section 5.4.1.6)
resume (Section 5.4.1.7)
shutdown (Section 5.4.1.8)
start (Section 5.4.1.9)
status (Section 5.4.1.10)
stop (Section 5.4.1.11)

Several of the commands accept a argument. A model is an XML representation of two
aspects of the Record tool:

❏ The configuration model: the XML configuration (similar to the XML configuration file
used to configure the Record tool.)

❏ The run-time model: an XML description of the entities that have been created based on
the configuration.
The format of this XML is the same as the configuration format for the Record tool (see
Chapter 4: Configuring the Record Tool). The top-level tag must be followed by
.
Some examples for are:

5-12

Using the Example Remote-Access Application—Record Shell

name="RTI Shapes Demo">

Square

color

5.4.1.1

add

This command adds entities to the Record tool.
The add command has the following format:
add
5.4.1.2

configure

The Record tool can be reconfigured remotely with the configure command. There are two ways
to reconfigure the Record tool; using a local file or a remote file.
Note that the Record tool must be stopped before it can be reconfigured. When the Record tool is
reconfigured, it will shut down completely. The Record Shell will lose its connection with the
Record tool until the Record tool re-establishes remote access. If remote access is not enabled in
the new configuration, Record Shell will not reconnect to the Record tool.
The configure command has the following format:
configure [-localfile | -remotefile ]

The configuration name is used to find the matching tag to load.

❏ -localfile
Example: Assume that you want to the Record tool to use a configuration file called
myconfig.xml, which is local to the Record Shell:
RTI Record Shell> stop
RTI Record Shell> configure myrecord -localfile myconfig.xml

The Record Shell will read the contents of myconfig.xml and send it to the Record tool,
which will search for a tag . If auto_start (see Table 4.2,
“General Properties”) is true (the default case), it is not necessary to run the start command to start the Record tool. If auto_start is false in the new configuration, then issue the
start command in the Record Shell to start recording:
RTI Record Shell> start

❏ -remotefile

To configure the Record tool with the contents of a file that is local to the Record tool, use
the -remotefile option.
For example, assume that you want reconfigure the Record tool with a file called remotemyconfig.xml, which resides on the same file-system as the Record tool.
RTI Record Shell> stop
RTI Record Shell> configure myrecord -remotefile remotemyconfig.xml

The Record tool will read the contents of remotemyconfig.xml and reconfigure with the
contents of the tag . Depending on the configuration file, it
may be necessary to start it:
RTI Record Shell> start

5-13

Using the Example Remote-Access Application—Record Shell

5.4.1.3

delete

This command deletes entities from the Record tool.
The delete command has the following format:
delete
5.4.1.4

exit

This command exits the Record Shell.
RTI Record Shell> exit
5.4.1.5

info

This command shows you which Record tool session the Record Shell is connected to. The output
looks similar to this:
STATE ..: Connected to [0a0a64fe.006bbe00]
GUID ...: 0a0a64fe.006bbb00

❏
❏
5.4.1.6

STATE

Which DomainParticipant the Record tool is connected to (HOSTID.APPID].

GUID

The GUID of the Record Shell itself.

pause

This command pauses the recording of entities in the Record tool.
The pause command has the following format:
pause
5.4.1.7

resume

This command resumes the recording of already paused entities in the Record tool.
The resume command has the following format:
resume
5.4.1.8

shutdown

This command causes the Record tool to shut down and terminate.
This command can only be issued when the Record tool has been stopped.
5.4.1.9

start

The start command is used to start the Record tool. Note that this command only works after
stopping the Record tool first, since the tool is started when it is launched.
When the start command is given, the Record tool will shut down completely, delete all state and
objects and start from scratch. By default, the Record tool will create a new fileset each time it is
started.
5.4.1.10

status

When the Record tool is configured with remote access enabled, it will periodically send its current status. The Record Shell stores the most recent status. The current status is displayed with
the status command:
RTI Record Shell> status

The output is similar to the following:

5-14

Using the Example Remote-Access Application—Record Shell

Version ..........:
Timestamp ........:
State ............:
Config file ......:
Database file ....:
Received bytes ...:
Saved bytes ......:

❏
❏
❏

The Record tool’s version

Version

Timestamp
State

5.0.0
Mon Feb 18 20:02:48 2013
STOPPED
simple_config.xml
simple_config.dat_34_3
86653952
2127872 (2 %)

The timestamp of the Record tool when status message was sent.

The Record tool’s state. The following states are possible:

• IDLE
• RECORDING
• STOPPED (the Record tool has been stopped and is not recording any user data)
• RECONFIGURE
• SHUTDOWN
• DOWNLOAD (the Record tool is downloading a new configuration)

❏
❏
❏
❏
5.4.1.11

Config file The name of the file from which the Record tool read its configuration. If the
configuration was received with configure -localfile, this field is not available.
Database file
Received bytes

The file-segment currently being written to.
Total amount of data that has been written to file.

Total number of data that has is currently saved to file. Note that if the rollover property is true, then Saved bytes may be less than Received bytes.
Saved bytes

stop

This command stops the Record tool from recording user data.
RTI Record Shell> stop

5.4.2

Running Multiple Record Tools in the Same Domain
The Record Shell can only keep track of one instance of the Record tool. To control multiple copies
of the Record tool in the same domain with the Record Shell, run each Record tool instance in a
separate partition.
For the first instance of the Record tool, change the configuration file as follows:

true
domain0

RecordA

5-15

Using the Example Remote-Access Application—Record Shell

For the second instance of the Record tool, change the configuration file as follows:

true
domain0

RecordB

Then you can run the Record Shell for each partition:
rtirecsh -partition RecordA
rtirecsh -partition RecordB

5-16

Chapter 6

Using the Replay Tool

Besides replaying data with Recording Console, you can use the Replay tool directly. You may find
this method useful when you want to tie its service into your own infrastructure or software, or
if you need to use its more advanced features.
The Replay tool replays recorded data by publishing it just like the original Connext DDS application did. You can use the original domain ID, QoS settings and data rate, or make changes to test
different scenarios.
See also:

❏ Chapter 7: Configuring the Replay Tool
❏ Chapter 8: Accessing the Replay Tool from a Remote Location

6.1

Recording Data for Replay
The Replay tool can replay information that has been stored in either serialized or deserialized
form. If Replay is to be used to replay deserialized data, ensure that all of the fields of the sample
data are recorded, as Replay is unable to replay partial data.
Note: SQLite is unable to look at the individual fields in the sample data of files recorded in serialized mode.

6.2

Starting the Replay Tool
Open a command prompt1 and change to the /bin directory. Then enter:

❏ On Linux and Mac OS X systems:
rtireplay -cfgFile -cfgName

❏ On Windows systems:
rtireplay.bat -cfgFile -cfgName

Table 6.1 describes the command-line options and which ones are required.

1. On Windows systems: from the Start menu, select Accessories, Command Prompt.

6-1

Stopping the Replay Tool

The Replay tool is dynamically linked against the Connext DDS libraries. You should run the tool
from the /bin directory scripts—not from the executable files themselves. The
scripts set all the paths and variables needed for the tool to find the shared libraries and run correctly.

6.3

Stopping the Replay Tool
To stop the Replay tool, use .

Table 6.1

Replay Tool’s Command-line Options

Command-line Option
-appName

Description
Specifies an application name which is used to identify
the application for remote administration.
Default: -cfgName

-cfgFile

Required.
Used to identify the XML configuration file.
Required.

-cfgName

-domainIdBase

Identifies the configuration within the XML
configuration file. The Replay tool will load the
with the same name as this value.
Adds this value to the domain IDs in the configuration
file.
Default: 0

-forceXmlTypes

When used with XML Type Configuration, this option
instructs Replay to always use type code from the XML
file, even if an alternate is available from recorded data.

-help

Displays this information.

-identifyExecution

Appends the host name and process ID to the
appName to help using unique names.

-noAutoEnable

Use this option if you plan to enable the Replay tool
remotely.

-remoteAdministrationDomainId

Enables remote administration and sets the domain ID
for the communication.
Default: remote administration is not enabled.

-srvName

Specifies a name that will be used to identify the
service.

-verbosity

Specifies what type of logging information should be
printed.
Silent
Exceptions (both Connext DDS and the Replay tool)
Warnings (the Replay tool only)
Information (the Replay tool only)
Warnings (both Connext DDS and the Replay tool)
Tracing (the Replay tool only)
Tracing (both Connext DDS and the Replay tool)
Default: 1

-version

Prints the Replay tool’s version.

6-2

Performance and Indexing

6.4

Performance and Indexing
The Replay tool replays stored samples in the same order in which they were received, using
SQLite indexes to retrieve the samples in sorted order. SQLite automatically builds indexes
when opening an SQLite table for sorted access, and for large tables the process of building the
index may take some time.
To improve initialization performance, the Replay tool attempts to create and store indexes (rather
than depend upon automatic indexing) for the tables that it will be replaying; this saves initialization time on subsequent replays.
The Replay tool's ability to store indexes is controlled by the parameter under
(see Database (Input File) Properties (Section 7.4)). The default value for
is false; this allows the Replay tool to write the table indices to the database. If you
change to true, the Replay tool will display a message during initialization for each
table opened, stating that it was unable to store the table index.
In summary, the replay performance of the Replay tool is not affected by the parameter. The Replay tool will use the fastest means of retrieving samples in either case. But setting the
option to false (the default) may help improve initialization performance.

6-3

Chapter 7

Configuring the Replay Tool

When you start the Replay tool, you must specify a configuration file in XML format. In that file,
you can set properties that control the data source, which topics to replay, and attributes such as
the replay speed. This chapter describes how to write a configuration file.

7.1

How to Load Replay’s XML Configuration File
The Replay tool loads its XML configuration from multiple locations. This section presents the
various approaches, listed in load order.
The first three locations only contain QoS Profiles and are inherited from Connext DDS (see
Chapter 15 in the RTI Connext DDS Core Libraries User's Manual).

❏ /resource/xml/NDDS_QOS_PROFILES.xml
This file contains the Connext DDS default QoS values; it is loaded automatically if it
exists. (First to be loaded.)

❏ File in NDDS_QOS_PROFILES
The files (or XML strings) separated by semicolons referenced in this environment variable are loaded automatically.

❏ /resource/xml/RTI_REPLAY_SERVICE.xml
This file contains the default configuration for the Replay tool; it is loaded if it exists.
RTI_REPLAY_SERVICE.xml defines a configuration that replays all topics on domain 0.

❏ /USER_REPLAY_SERVICE.xml
This file is loaded automatically if it exists.

❏ The file specified with the command-line option, -cfgFile (see Table 6.1 on page 6-2).
You may use a combination of the above approaches.

7-1

General Format

7.2

General Format
The Replay configuration file uses XML format. The main sections use the following top-level
tags:
Top-level Tag

Reference Section

General Properties for Replay (Section 7.3)

Database (Input File) Properties (Section 7.4)

Session Properties (Section 7.5)

Replay Topic Properties (Section 7.6)

Time Control Properties (Section 7.7)

Remote Administration Properties (Section 7.8)

Type Configuration (Section 7.9)

The XML configuration file used by Replay has a simple hierarchical format. The replay_service
is configured to replay data contained in one or more replay_database. Each replay_database is
associated with a DomainParticipant, and must contain one or more session. Each session is associated with a Publisher, and corresponds to a unique execution thread. Each session contains one
or more replay_topic, each of which is associated with a DataWriter, and contains a filter expression that specifies what information contained in the data base should be replayed. Each of the
four major levels— replay_service, replay_database, session, and replay_topic—may contain a
time_control element that allows control over such features as the rate of replay, how much of
the available data to be replayed, and for coordination.
Much of Replay’s configuration has been designed to be compatible with the Record tool, so
familiarity with the Record tool’s concepts and configuration will be helpful. See Chapter 4: Configuring the Record Tool.
Let’s look at a very basic configuration, just to get an idea of its contents. You will learn the
meaning of each line as you read the rest of this chapter.

yes

replay_database.dat_0_0

7-2

General Properties for Replay

*
*
*
*

7.3

General Properties for Replay
Table 7.1 describes optional properties that control the Replay tool’s main module.
All properties are optional except replay_database.
These properties must be specified inside and
tags, where String is the name to be assigned to the service entity when it is
created. This name will be used during remote administration unless it is overridden by the
element.

Table 7.1

Replay Service Properties

Property

Syntax

administration

Remote Admin.
Properties

auto_exit

DDS_Boolean

Description
Configures the DomainParticipant that can be used to
remotely control Replay via the rtireplaysh utility.
See Remote Administration Properties (Section 7.8).
The Remote Administration Properties must specify a
domain_id. You may also specify a name, participant_qos,
publisher_qos, subscriber_qos, datareader_qos, and
datawriter_qos.
Controls whether or not the Replay tool should terminate
when all the available data specified in the initial
configuration has been replayed.
Default: True

7-3

Database (Input File) Properties

Table 7.1

Replay Service Properties

Property

Syntax

replay_database

time_control

Time Control Properties

7.4

Description

Required. Specifies configuration properties that describe

how to replay the information from a database. This
Replay Database Properties element can be repeated.

See Database (Input File) Properties (Section 7.4)
Specifies time configuration properties to be applied to
the Replay tool as a whole.
See Time Control Properties (Section 7.7).

Database (Input File) Properties
Table 7.2 describes the source of the data that Replay will replay.
All properties are optional except session.
These properties must be specified inside and
tags, where String is the name to be assigned to the database entity when it
is created. This name will be used during remote administration.

Table 7.2

Replay Database Properties

Properties
under

filename

Syntax

String

Description
Specifies the name of the fileset that
contains the data to be replayed.
Do not use both and
at the same time.
Default: Undefined

fileset

participant

Specifies a fileset to be used by the
Replay tool. This fileset must contain the
file prefix to be searched for, the set
number, and the start and end segments.

String
DDS_Long
DDS_Long
DDS_Long

Do not use both and
at the same time.

Participant Properties

See Table 7.3, “Participant Properties”

Note: The fileset replay feature is not
compatible with the reposition
command for timestamps outside of the
current segment being replayed.

7-4

Database (Input File) Properties

Table 7.2

Replay Database Properties

Properties
under

readonly

Syntax

DDS_Boolean

Description
Specifies if Replay should open the data
file in read-only mode (true), or readwrite mode (false). Setting this option to
false is useful to enable indexing of
older database files.
Default: False
See Performance and Indexing (Section
6.4).
Required

Session Properties

session

The configuration properties that
describe how to replay the information
in a session. This element can be
repeated.
See Table 7.4, “Session Properties”

time_control

type_config

Time Control Properties

XML Properties

The time configuration properties to be
applied to the replay database.
See Table 7.8, “Time Control Properties”
Optional XML type configuration for
this replay_database. This option is useful
when type codes have not been
recorded in the database, or when
specifying types that are too large to be
recorded in the database.
See Table 7.13, “XML Type
Configuration Properties”

Table 7.3

Participant Properties

Properties
under

domain_id

participant_qos

Syntax

DDS_Long

DDS_QosPolicy

Description

Sets the domain ID. Default: 0
Configures the DomainParticipant’s QoS policies. See the
RTI Connext DDS Core Libraries User’s Manual’s chapter on
Configuring QoS with XML.
Defaults: See the RTI Connext DDS API Reference HTML
documentation on DomainParticipants.
See the Note: below for more information.

Note:

❏ If you do not set the property in the settings, the
Replay tool will automatically build a participant name and set it using the prefix "RTI
Replay: ". This is for compatibility with RTI Administration Console. If this property is

7-5

Session Properties

changed, the Replay tool won't override the property, but compatibility between the
Replay tool and RTI Administration Console will be broken if the participant name is not
prefixed with "RTI Replay: " (notice the space after the colon).

7.4.1

Enabling Monitoring Library with Replay
This section only applies if you want to use RTI Monitoring Library (included in Connext DDS),
which enables Connext DDS applications to provide monitoring data. The monitoring data can
be visualized with RTI Monitor, a separate GUI application that can run on the same host as
Monitoring Library or on a different host. Recording Service is statically linked to Monitoring
Library (you do not have to install it separately).
To enable monitoring in the Replay tool, use the same approach described in Enabling Monitoring Library in the Record Tool (Section 4.7.1). In the section, include the
rti.monitor.library property with the value rtimonitoring. For example:

rti.monitor.library
rtimonitoring
false

7.5

Session Properties
Table 7.4 describes the Session’s properties.
All properties are optional except replay_topic.
These properties must be specified inside and tags, where
String is the name to be assigned to the session entity when it is created. This name will be used
during remote administration.

Table 7.4

Session Properties

Properties
under

Syntax

publisher_qos

DDS_QosPolicy

replay_topic

Replay Topic Properties

Description
Configures the Publisher’s QoS policies. See the RTI
Connext DDS Core Libraries User’s Manual’s chapter on
Configuring QoS with XML.
Defaults: See the RTI Connext DDS API Reference HTML
documentation on Publishers.
Required. The configuration properties that describes the
topics to be replayed, and the associated DataWriter
configuration. This element can be repeated. See Table 7.5,
“Replay Topic Properties”

7-6

Replay Topic Properties

Table 7.4

Session Properties

Properties
under

7.6

Syntax

Description

thread

Thread Properties

Configures the properties for the execution thread.

time_control

Time Control Properties

Configures the time control properties to be applied to the
Session. See Table 7.8, “Time Control Properties”

Replay Topic Properties
Table 7.5 describes the Topics’ properties.
All properties are optional except input.
These properties must be specified within and
tags, where String is the name to be assigned to the replay topic entity when it
is created. This name will be used during remote administration.
All input properties (Table 7.6) are required, except for type_name, which is optional.
All output properties (Table 7.7) are optional.

Table 7.5

Replay Topic Properties

Properties
under

Syntax

input

Input Properties

Required. Configures the topics that are to be replayed
from the database. See Table 7.6, “Input Properties”.

output

Output Properties

Configures the attributes to be used in writing the replayed
topics. See Table 7.7, “Output Properties”.

time_control

Time Control Properties

Table 7.6

Description

Specifies time configuration properties to be applied to the
Session. See Table 7.8, “Time Control Properties”.

Input Properties

Properties
under

Syntax

Description

domain_name

String

Required. Specifies the name of the domain_name that was
specified in the Record tool’s configuration file or a regular
or wildcard expression.

record_group_
name

String

Required. Specifies the name of the record_group that was
specified in the Record tool’s configuration file or a regular
or wildcard expression.

7-7

Replay Topic Properties

Table 7.6

Input Properties

Properties
under

Table 7.7

Syntax

topic_name

String

Required. Specifies the name of the topic_name that was
specified in the Record tool’s configuration file or a regular
or wildcard expression.

type_name

String

Specifies the name of the type_code to be used in writing
matching topics. This parameter will default to “*” if not
specified. Replay will search for a matching type name only
within matching topic records.

Output Properties

Properties
under

7.6.1

Description

Syntax

Description

datawriter_qos

DDS_DataWriterQos

Specifies the QoS settings for all DataWriters created for this
Replay_Topic. A DataWriter is created for each Topic that
matches the topic_expr. All the DataWriters for the
Replay_Topic will use the same set of QoS policies. You can
specify all of the QoS policies with this datawriter_qos
property.
See the RTI Connext DDS Core Libraries User’s Manual’s
chapter on Configuring QoS with XML.

topic_name

String

Specifies the name to be assigned to the topic when
creating a DataWriter to write the data to be replayed.

topic_qos

DDS_TopicQos

Specifies the QoS settings to be applied to the topic when
creating a DataWriter to write the data to be replayed.

type_name

String

Specifies the name to be assigned to the type when creating
a DataWriter to write the data to be replayed.

Type Selection
This sections explains how the Replay tool obtains the type version (TypeCode) and the registered type name that will be used to replay a topic “T”.

❏ First, if there are XML type-configuration settings () for the replay database (see Section 7.9) and the name of the topic to be replayed matches any of the topic
name expressions provided using the tag within , the Replay tool will use
the XML definition of the type. The registered type name will be determined as follows:
1. The
type
name
explicitly
.
2. If

defined
false,

by
the

the
name

user
provided

in
in

3. The canonical name (fully qualified name in the XML file) of the type.

❏ Second, if there is no XML type definition, the Replay tool will first try to get a candidate
registered type name as follows:

7-8

Time Control Properties

1. The
type
name
explicitly
.

defined

the

user

2. If there is no explicit type name, the Replay tool will try to get the name using the column type_name in the DCPSPublication table. If the name is not unique, the Replay
tool will report an error.

❏ Third, if a valid registered type name was obtained, the Replay tool will try to get a type
definition (TypeCode) associated with the given registered type name as follows:
1. Using the type definitions under
2. If there is no type definition available in XML, the tool will try to get the type definition using the typecode column in the DCPSPublication table. If there is no available
or unique type definition, the Replay tool will report an error.

7.7

Time Control Properties
The element can be applied to any of Replay’s major entities: ,
, , and .

❏ The index time of the is the earliest index time of all of its component
replay_database entities.

❏ The index time of a is the earliest timestamp of the database, taken
from its creation log.

❏ The index time of a is the earliest timestamp of the topic, taken from the
first recorded sample of the topic.
All time control properties are optional.
The start_time and stop_time values of a child entity are constrained by the start_time and
stop_time settings of its parent entities. If a start_time or stop_time value is explicitly specified
and constrained by one of its parent entities, Replay will issue a warning that the value has been
truncated.
The start_mode of a child entity overrides the start_mode setting of its parent entities.
The time_mode value of a child entity cannot be applied to a parent entity (e.g.,
TOPIC_RELATIVE cannot be applied to the time_control time_mode element of a ,
, or ).

7.8

Remote Administration Properties
The Replay tool can be controlled remotely, by either the rtireplaysh utility, or by a Connext DDS
application that reads and writes the remote administration topic.
For security reasons, Remote Administration is turned off in the Replay tool by default.
The Remote Administration section of the configuration file is used to enable Remote Administration and configure its behavior. This section is not required in the configuration file.
The rtireplaysh utility or a remote application can send commands to the Replay tool to:

❏ Start/stop/pause/resume/step the Replay tool or any of its individual entities.

7-9

Remote Administration Properties

Table 7.8

Time Control Properties

Properties
under

Syntax

Description
Specifies the replay rate, expressed as a multiple of the original
rate at which the data was recorded. (Therefore 1 means the
same as the original rate, 2 means twice as fast, etc.)

Real Number

rate

Although this rate may be specified as a real decimal number,
the internal resolution of the rate value is stored as a
percentage with two decimal places. The rate may also be
configured with the special value “AS_FAST_AS_POSSIBLE”,
which directs Replay to replay the data without any
intervening time between samples.
The minimum value is 0.01 (1% of the original rate.)
Default: 1

start_mode

Start Mode

Sets the starting mode of the entity, as described in Table 7.9,
“Start Mode Values”. Default: AUTOMATIC
The time to offset the selected entity’s starting time from its
parent entity. This value is used for synchronizing data that is
replayed from different sources.
When applied to the , Replay will delay for
the number of seconds specified between the creation of the
entities and the start of replay. (To allow for discovery, for
example).

start_offset

DDS_Duration

When applied to a , this time is the
amount of offset between the index time of the replay_service
and the index time of the database.
When applied to a , this time allows an
additional DDS_Boolean element, , which if set
TRUE directs Replay to automatically calculate the offset
between the index time of the and the
index time of the . This value should not be
applied to or .
Default: 0

start_time

stop_time

DDS_Duration

The time of the recorded data at which replay is to begin. The
time is interpreted based on the setting of time_mode.
Although expressed as and , the internal
resolution of Replay is limited to milliseconds. Default: 0
The time of the recorded data at which replay is to stop. The
time is interpreted based on the setting of time_mode.
Although expressed as and , the internal
resolution of Replay is limited to milliseconds.
Default: Infinity

time_mode

Time Mode

Describes how the start_time and stop_time parameters
should be interpreted. See Table 7.10, “Time Mode Values”.
Default: DATABASE_RELATIVE, except when applied to a
replay_service entity, whose default is SERVICE_RELATIVE.

❏ Change the speed of replay of the Replay tool or any of its individual entities.
❏ Query the status of the Replay tool or any of its individual entities.

7-10

Remote Administration Properties

Table 7.9

Start Mode Values

Enumeration
Value
AUTOMATIC

Description
Replay of the entity begins automatically. For subordinate entities, replay begins when
parent replay starts.
Replay of the selected section begins automatically, and is restarted immediately after
the last data sample of the entity has been replayed. For example, replay_topics with
start_mode LOOP will each restart as soon as each topic has completing its replay, while
a session with start_mode LOOP will restart only when all of its topics have completed
replay.

LOOP

Note: Currently this mode is operational only for session and replay_topic entities.
MANUAL

MATCHED

Replay of the entity begins when explicitly directed by remote administration. When an
entity is manually started, all of its child entities with AUTOMATIC start_mode will
also be started at the same time, and so forth continuing to the lowest child.
CURRENTLY NOT SUPPORTED.
Replay begins after each child DataWriter has detected at least one matched reader.

Table 7.10 Time Mode Values

Enumeration
Value

Description

ABSOLUTE

The start_time and stop_time values are in absolute timestamps and will be used
without modification.

DATABASE_
RELATIVE

The start_time and stop_time values are relative to the replay_database’s index time (i.e.,
the index time of the replay_database will be added to the start_time and stop_time
values, if specified).

SERVICE_
RELATIVE

The start_time and stop_time values are relative to the replay_service’s index time (i.e.,
the index time of the replay_service will be added to the start_time and stop_time values,
if specified).
CURRENTLY NOT SUPPORTED.

TOPIC_
RELATIVE

The start_time and stop_time values are relative to the earliest timestamp of the topic
(i.e., the index time of the topic will be added to the start_time and stop_time values, if
specified).

❏ Reposition the Replay tool or any of its entities to any point in the replay range specified
for playback.
Chapter 8: Accessing the Replay Tool from a Remote Location describes the command format
and individual commands available in rtireplaysh, Replay’s remote administration utility.
Table 7.11 describes the Remote Administration properties. All remote administration properties
are optional.
All Remote Administration properties must be specified inside and tags.

7.8.1

Enabling RTI Distributed Logger in the Replay Tool
The Replay tool provides integrated support for RTI Distributed Logger (included in Connext
DDS).
When you enable Distributed Logger, the Replay tool will publish its log messages to Connext
DDS. Then you can use RTI Monitor1 to visualize the log message data. Since the data is pro1. RTI Monitor is a separate GUI application that can run on the same host as your application or on a different host.

7-11

Remote Administration Properties

Table 7.11 Remote Administration Properties

Properties under

Syntax

datareader_qos

DDS_DataReaderQos

datawriter_qos

DDS_DataWriterQos

distributed_logger

Distributed Logger
Properties

domain_id

DDS_Long

name

String

participant_qos

DDS_ParticipantQos

publisher_qos

DDS_PublisherQos

Description
Configures the QoS for the DataReader created by
the Replay tool’s Remote Access module.
Defaults: See the RTI Connext DDS API Reference
HTML documentation on DataReaders.
Configures the QoS for the DataWriter created by
the Replay tool’s Remote Access module.
Defaults: See the RTI Connext DDS API Reference
HTML documentation on DataWriters.
Configures RTI Distributed Logger.
See Enabling RTI Distributed Logger in the
Replay Tool (Section 7.8.1).
Sets the domain ID.
Default: 0
Assigns a name to the replay service. You can use
this name when sending commands via Replay
Shell (see Section 6.4).
Configures the QoS for the Participant created by
the Replay tool’s Remote Access module.
Defaults: See the RTI Connext DDS API Reference
HTML documentation on DomainParticipants.
Configures the QoS for the Publisher created by
the Replay tool’s Remote Access module.
Defaults: See the RTI Connext DDS API Reference
HTML documentation on Publishers.
Specifies, in seconds and nanoseconds, the period
between each status message sent by the Replay
Service to the Replay Shell.

status_period

DDS_Duration

When this value is set to zero (the default), no
status message is sent.
Applications that want to periodically poll the
status of the Replay service they administer
should provide a value for this property.
Default: 0 (no status is sent)

subscriber_qos

DDS_QosPolicy

Configures the QoS for the Subscriber created by
the Replay tool’s Remote Access module.
Defaults: See the RTI Connext DDS API Reference
HTML documentation on Subscribers.

vided in a Connext DDS topic, you can also use rtiddsspy or even write your own visualization
tool.
To enable Distributed Logger, modify the Replay tool’s XML configuration file. In the section, add the tag as shown in the example below.

Replay Service using Distributed Logger
99

7-12

Type Configuration

true

There are more configuration tags that you can use to control Distributed Logger’s behavior. For
example, you can specify a filter so that only certain types of log messages are published. For
details, see the RTI Distributed Logger section of the RTI Connext DDS Core Libraries User’s Manual.

7.9

Type Configuration
The element allows you to pass type configuration information to the Replay tool
in the form of XML type-configuration files.
Table 7.12 describes the Type Configuration properties. All Type Config properties are optional.

Table 7.12 Type Configuration Properties

Properties
under

xml

Syntax

XML Type Configuration
Properties

Description
Allows you to specify XML type-configuration
files, the path in which to find them, and other
properties related to the registration of the types
with the DomainParticipants.
The XML type configuration for this domain
group. See Table 7.13, “XML Type Configuration
Properties”.

Table 7.13 XML Type Configuration Properties

Properties
under

file_group

Syntax

File Group Properties

Description
Allows you to specify XML file groups; each of
these files contain type definitions to be used by
the Replay tool. The element tag can be repeated.
See Table 7.14, “File Group Properties”.

max_sequence

Integer

The default sequence size, in case there are
unbounded sequences in the type definitions
specified in the any of the files specified in any of
the file groups.

max_string

Integer

The default string size, in case there are
unbounded strings in the type definitions
specified in any of the files specified in any of the
file groups.

7-13

Recording Service Integration with Extensible Types

Table 7.13 XML Type Configuration Properties

Properties
under

path

Syntax

Path

register_top_level Boolean

Description
A list of the paths (relative or absolute) to be used
when searching for the XML type definition files.
The tag can be repeated.
Whether or not to register the top-level types
found in the type definitions with their canonical
names. Do this with any of the files defined in
any of the file groups. Default: TRUE.

Table 7.14 File Group Properties

Properties under

Syntax

Description

file_name

File Name string

A list of file name strings, specifying files containing
the type definitions. The element tag can be repeated.

max_sequence

Integer

The default sequence size in
case there are
unbounded sequences in the type definitions specified in the any of the files specified in this specific file
group.

max_string

Integer

The default string size in case there are unbounded
strings in the type definitions specified in any of the
files specified in this specific file group .

Boolean

Whether or not to register the top-level types found
in the type definitions with their canonical names. Do
this with any of the files found in this specific file
group. The value of this setting overrides the value of
the upper-level setting (see Table 7.13, “XML Type
Configuration Properties”). Default: TRUE.

register_top_level

type

Type
Registration
Properties

Specific type properties. A list of type registration
properties used to define how a type found in the
files has to be registered by Recorder in the DomainParticipants. See Table 7.15, “Type Registration Properties”

7.10

Recording Service Integration with Extensible Types
The Replay tool includes partial support for the "Extensible and Dynamic Topic Types for DDS"
specification from the Object Management Group (OMG)1. This section assumes that you are

1. http://www.omg.org/spec/DDS-XTypes/

7-14

Recording Service Integration with Extensible Types

Table 7.15 Type Registration Properties

Properties
under

register_top_
level

Syntax

Boolean

Description
Whether or not to register this type's canonical name
(as defined in tag type_name) with the DomainParticipant. This name is registered alongside any of the
registration
type
names
defined
in
tag
registered_type_name above. Default: TRUE.

registered_
type_name

Registration
name string

When registering the type with the DomainParticipant, this setting defines a list of names to register
the type with. The element tag can be repeated.

topics

Topic name string

type_name

Type Name string

This is a list of regular POSIX fn-name expressions
(every entry is equivalent to an expression). A topic whose name matches any of the
expressions will be recorded using the type definition designated by . If a topic matches
more than one expression in different
entries, the first one that was matched will be used.
A string representing the canonical type name as
defined in the XML type definition for the type.

familiar with Extensible Types and you have read the RTI Connext DDS Core Libraries Getting
Started Guide Addendum for Extensible Types.

❏ The Replay tool can publish topics associated with optional, final, mutable, and extensible
types.

❏ Users can provide an XML definition of the type that must be used by the Replay tool to
publish a topic. The XML description supports structure inheritance and type mutability
(described in the RTI Connext DDS Core Libraries Getting Started Guide Addendum for
Extensible Types).

7.10.1

Selecting the Type Version to use when Replaying a Topic
With Extensible Types, it is possible to have multiple versions of a type associated with a topic
“T”. Recording Service allows you to select the type version that will be used to record the data
(see Section 4.10). However, Recording Service persists the type description (TypeCode) corresponding to each one of the discovered versions in the DCPSPublication table.
The type used by the Replay tool to replay a certain topic is obtained as described in Type Selection (Section 7.6.1). When a unique type name and type-code exist for a topic in the DCPSPublication table, the Replay tool will be able to use that information to replay it. In any other case,
XML Type Configuration is needed. As an example, imagine that we have the database recorded
in the example shown in Selecting a Type Version For a Topic “T” In a Recording Domain (Section 4.10.1). It is recorded using the extended type version of the Shapes Demo types. In order to
replay the tables in the database as samples of the base type version, we could configure the
Replay tool in the following manner:

7-15

Recording Service Integration with Extensible Types

true

shapes.dat_0_0

false

ShapeType.xml

true
ShapeType

Circle
Square
Triangle

*
*
*
*

The example uses the same ShapeType.xml file defined in Selecting a Type Version For a Topic
“T” In a Recording Domain (Section 4.10.1) and Recording Two Versions of a Type in Different
Tables in Same Database (Section 4.10.2). By using the tag (see Table 7.15, “Type Registration Properties”) in the settings in the XML Type Configuration, we make a direct
mapping of the topics to the type to be used for replay. The above configuration will result in all
Shapes Demo topics being replayed with the base ShapeType type definition.

7.10.2

Replaying Topics with Different Type Versions Stored in Different Tables
If the Record tool was set up to record each type version for a topic in its own table (see Recording Two Versions of a Type in Different Tables in Same Database (Section 4.10.2)), it seems interesting to be able to do the same with the Replay tool and have every type topic correctly replayed
with the original type version. This can be done by using different Replay database entities con-

7-16

Recording Service Integration with Extensible Types

figured to replay each of the recorded versions (distinguished by the name of the Recording Service domain that recorded them).
Consider the following Replay configuration:

true

shapes.dat_0_0

false

ShapeType.xml

true
ShapeType

Circle
Square
Triangle

*
domain0Base
*
*

shapes.dat_0_0

false

ShapeType.xml

7-17

Recording Service Integration with Extensible Types

false
ShapeTypeExtended

ShapeType

Circle
Square
Triangle

*
domain0Extended
*
*

In the above configuration, there are two Replay database entities, each replaying the base and
the extended version of the topics, respectively. This is done by specifying the adequate domain
name in the tags (one database replays domain
"domain0Base" and the other one replays "domain0Extended"). For each of these database entities, different XML Type Configuration settings are provided. Topics "Circle", "Square" and "Triangle" are mapped to the type ShapeType in the base Database entity, and they are mapped to
the type "ShapeTypeExtended" in the extended Database entity.
Note: There is a known issue when accessing the same database file from different Replay database entities when the indexes haven't been created yet1. There are three ways to avoid this
issue:
1. Record with the flag set to true so that the indexes
are created by Recording Service while recording.
2. Set the to true. This will disable write-mode access to the
database and thus indexes won't be created. However, this can affect the performance
while replaying the data (see Performance and Indexing (Section 6.4)).

1. RTI Issue ID RECORD-318

7-18

Recording Service Integration with Extensible Types

3. Index the database tables offline by using any SQLite access tool (for an example, see
Chapter 9: Viewing Recorded Data) the indexes on the table can be created with the following SQL query (all on one line):
CREATE INDEX IF NOT EXISTS
"index$[table name]" ON "[table name]"(reception_timestamp);

For additional information on how to provide XML types to the Replay tool, see Type Configuration (Section 7.9).

7-19

Chapter 8

Accessing the Replay Tool from a Remote
Location

The Replay Shell is a Connext DDS application that can remotely control the Replay tool.
To start the Replay Shell:

Open a command prompt1, change to the /bin directory, and enter:

❏ On Linux and Mac OS X systems:
rtireplaysh [options]

❏ On Windows systems:
rtireplaysh.bat [options]

Table 8.1 lists the command-line options you can use when starting the Replay Shell. Once it is
started, you can use the commands in Table 8.2.
Table 8.1

Replay Shell’s Command-Line Options

Command-line Option
-cmdFile
-domainId

Description
A file that contains commands to be run.
Specifies the domain ID, an integer between 0 and 232.
Default: 0

-help

Prints version information and a list of options.

-timeout

Maximum number of seconds to wait for a remote response.
Default: 15 seconds

-verbosity

Specifies what type of logging information should be printed.
Silent
Exceptions (both Connext DDS and the Replay tool)
Warnings (the Replay tool only)
Information (the Replay tool only)
Warnings (both Connext DDS and the Replay tool)
Tracing (the Replay tool only)
Tracing (both Connext DDS and the Replay tool)
Default: 1

1. On Windows systems: from the Start menu, select Accessories, Command Prompt.

8-1

The Replay Shell commands use this format:
[entity] [value]

where:

❏ is one of the supported commands (see Table 8.2).
❏ is the name given to the Replay service by one of the following, in
descending order of precedence:
• The value specified with the -appName command-line option used when starting the
Replay tool (highest precedence)
• The value for the element (see page 7-12)
• The value for the name attribute (lowest precedence) (see Section 7.3)

❏ [entity] is any one of the service entities expressed in this hierarchical form: [::[::]].
Note: In this release, not all commands are supported for all entity levels. Please see the Recording Service Release Notes for details on which modes are currently supported.
The database-name must match a name from a tag in the configuration file that you specified when starting the Replay tool , such as:

Similarly, if you specify a session-name, it must match a name from a tag
within the specified database, such as:

If you specify a topic-name, it must match a name from a tag within the
specified session, such as:

If you do not specify an entity, the command is applied to the replay service itself.

❏ value depends on the command, see Table 8.2. Not all commands require a value.
Table 8.2

Replay Shell’s Commands

Command
exit

Description
Exits the shell.
Repositions an entity to a specific point in the playback range (relative to the entity's start
and end times).

goto

This command takes a timestamp argument, which is a string of digits of the form
"SSSSSSSSSSUUUUUU". The first ten digits specify seconds and the last six digits specify
microseconds.

pause

Pauses replay of an entity.

query

Returns the status of an entity, including:
• If the entity is enabled, started, pending, paused, and completed
• The number of child topics owned by the entity
• The number of active child topics owned by the entity
The format of this status is "[ccccc dd dd]", where each c is either 'T' or 'F', and 'dd' is a decimal number. The 'T' and 'F' entries represent enabled, started, pending, paused, and completed. The d’s are decimal numbers for how many child topics are owned by the entity
and how many active child topics are owned by the entity.

8-2

Table 8.2

Replay Shell’s Commands

Command

Description
Changes the replay rate of an entity.

rate

The rate is a multiplier from 0.1 to 4 billion. It replays at the speed of the multiplier (2 = 2x,
0.5 = 1/2x, etc.)
Default: 1

resume

Resume replay of an entity.

start

Starts replay of an entity.

step

Replays a single sample from the entity.

stop

Stops replay of an entity.

8-3

Chapter 9

Viewing Recorded Data

The data can be viewed with any command-line or GUI-based tool capable of reading SQLite
databases.

9-1

9-2

Chapter 10 Converting and Exporting Recorded Data
Recording Service includes a Converter tool that enables serialized or deserialized data recorded
with the Record tool to be exported to CSV, HTML or XML formats. It also allows you to convert
serialized data tables into deserialized data tables so that the recorded data can be easily
explored via SQL/SQLite in a human-readable way.
To work with the database files created by the Record tool, Converter needs the
replay_compatibility flag to be turned on (see Table 4.2), otherwise conversion is not guaranteed to work properly.
Tables that were recorded with field filters (field_expr) cannot be converted.
General Usage

The script to launch Converter is in /bin/. To start the tool, open a command
prompt, change to /bin, and enter:

❏ On Linux and Max OS X systems:
rtirecconv [options] fileset|filename

❏ On Windows systems:
rtirecconv.bat [options] fileset|filename

The parameters fileset or filename represent the name of a recorded file set such as
recorded_db.dat_0 or a concrete file segment such as recorded_db.dat_0_0.

10.1

Exporting Data
Converter can export data recorded by the Record tool or Recording Console in three different formats: CSV (comma-separated values), XML (default export format if not specified) and HTML.
Even if the exported table is serialized, the output format of the data is deserialized. By default,
only valid samples (data for which the valid_data Sample Info value was true) are exported;
Sample Info columns are not exported.
Here is an example of how a samples from RTI Shapes Demo is exported by Converter into XML
format (external XML formatting was applied after the conversion):

10-1

Exporting Data

1381746283759122998
MAGENTA
134
140
30
1
0.000000

...

There are several options available to select how and what should be exported:
Flag
-format

Values
[ xml | csv | html ]

Description
Specifies what format to use to export the data. If this flag is not
specified, XML format is used.

-includeInfo

If present in the call to Converter, the Sample Info fields will be
exported with each sample.

-includeNonData

If present in the call to Converter, samples will be exported even if
the associated Sample Info valid_data property is false.

-compact

[ auto | yes | no ]

The Record tool may store arrays and sequences of single-byte
primitive types (characters and octets) in compact mode. This
means that the array or sequence is stored in one only blob
column in the recorded database. Converter can detect how data
is stored in this regard and follow the same pattern when
exporting the data.
The auto mode (default) will detect the compactness of singlebyte collections and use it for exporting the data. There are ways
to force the converted data to be compact or non-compact. By
specifying yes, the single-byte collections will be exported in
compact form. By specifying no, the data will be exported in noncompact form.
This flag is used to specify how to export values of character type
in arrays or sequences. The default format is 'text'. Explanation of
the two formats:

-decodeChar

[ hex | text ]

Text: If the character being exported is printable, it's printed
directly. Otherwise, it's printed as two hex digits preceded by the
escape character '\' (e.g. '\0B'). If exporting to XML or HTML,
special/reserved characters like '>', '<', etc. are exported as their
equivalent, '>', '<' and so forth.
Hex: Every character is exported as two hex digits. Every
exported character is separated by a dash. For example, 0B-FA18-D3... and so on.
This flag is used to specify how to export byte values in arrays or
sequences.

-decodeOctet

[ hex | text ]

Text: If the byte being exported is printable, it's printed directly
as a character would be printed. Otherwise, it's printed as two
hex digits preceded by the escape character '\' (e.g. \0B). If
exporting to XML or HTML, special/reserved characters like '>',
'<', etc. are exported as their equivalent, '>', '<' and so forth.
Hex (default): every character is exported as two hex digits.
Every exported character is separated by a dash. For example,
0B-FA-18-D3 represents four bytes.

10-2

Deserializing Serialized Tables

Flag

-time

Values

[ epoch | gmt ]

Description
Every sample that is exported includes a "Timestamp" field
which with the reception timestamp of the sample. This
timestamp can be exported in two ways:
epoch: The timestamp in nanoseconds since January 1 1970.
gmt: The timestamp in GMT time format. For example: Fri Oct 11
17:25:03 2013.

10.2

Deserializing Serialized Tables
Samples stored in serialized format are not human-readable. Exporting the data automatically
deserializes serialized data but it may be the case that we need to keep the data as a SQLite database for inspection or data handling, or because we still want to be able to use the Replay tool.
Converter provides an option to process a SQLite database and transform all serialized tables
into deserialized tables: -deserialize. This option takes precedence over any other format option,
like HTML or XML. When -deserialize is included the format specifier is ignored.
By default, the new database file is created in the same location as the original one and the new
file is prefixed with "deserialized." followed by the original file's name. The two files, original
and deserialized, are equivalent. If the Replay tool was able to work with the original file, it will
be able to replay the deserialized file.

10.3

Handling Data Types
Converter infers the type-code that should be used to deserialize and/or export the data from the
DCPSPublication table and the topic name. However, the type-code may be missing, either
because it wasn't published (not propagated by DomainParticipants during discovery) or
because it was filtered by using field selection properties. It may also be desirable to force types
to be loaded using XML configuration, e.g. in cases where different versions of the types were
published and we want fine-grained control of the type being exported.
If the type-code is not directly accessible from the DCPSPublication table, Converter can provide
XML representation of the types via the Recorder XML configuration file and name. The following command-line parameters can be used to provide Converter with XML type information::
Flag

-typeConfig

-forceXMLTypes

Parameters

Description

When this option is present and a correct Recorder XML
configuration file and configuration name are provided,
the XML type configuration settings in the configuration
(see Recording Service Integration with Extensible Types

(Section 4.10) will be used to match the topics and types in
the database and provide type-codes for them.
This option forces Converter to obtain the type-codes from
the XML configuration provided via the -typeConfig
option, whenever possible.

10-3

Selecting Output Files

10.4

Selecting Output Files
Converter extracts the file set from the input file (or set) provided by the user and detects all
available segments for conversion. Conversion is performed on every segment, and there is only
one output file, which aggregates all the conversion results. So, for example, if we have file segments rti_record_default.dat_0_0 and rti_record_default.dat_0_1, running Converter on any of
the files or the file set (rti_record_default.dat_0) will convert both segments into the same file.
Note: It is not possible to convert a single segment with Converter. If there is more than one
recorded file in a fileset Converter will automatically detect it and convert all of them.
The output file is created by appending the appropriate file extension depending on the conversion format. For example, if we are converting the fileset mentioned above into XML format, the
output file would be called rti_record_default.dat_0.xml, located in the same directory as the
original fileset.
You can, however, change the output file. The command-line options to alter the output file are
as follows:
Flag

Parameters

Description

-outputFile

If this option is present and has a valid file name parameter, the conversion
results will be placed in the file passed as a parameter instead of the default
filename derived from the converted fileset's name.
This option has a string parameter used as a prefix for each of the output
files created.
This option cannot be used at the same time as -outputFile.

-filePrefix

When this option is present, an output file is created for each table in the
fileset, prefixed with the prefix provided by the user.
For example, if a recording contains tables Table1 and Table2 and the filePrefix option is used with prefix "toXML", two output files will be
created: toXML_Table1.xml and toXML_Table2.xml.

10.5

Exporting Discovery Tables
Converter's default behavior is to export just the user topic tables, except when deserializing,
where the discovery tables are copied automatically to the deserialized SQLite database. It is
possible to tell Converter to export the discovery tables when converting. If the -includeDiscovery option is present, Converter will export the discovery tables present in the recorded file
alongside the user data tables.

10.6

Filtering User Topic Tables
Converter provides the ability to filter the tables to be exported. If the -tableExpr
command-line option is used, Converter will export or convert tables that follow the patterns
described by this option. The expression parameter for this option is a POSIX fnmatch expression. The tableExpr option can be repeated. If the name of a table matches any of the expressions
provided with tableExpr, it is converted/exported; otherwise it isn't.

10-4

Filtering User Topic Tables

Discovery tables are not affected by the tableExpr option. If -includeDiscovery is included, the
discovery tables will be exported normally. The -deserialize option will automatically include
the discovery tables in the output database file, no matter what table expression is used. However, user topic tables will be filtered according to the table expression.

10-5

Chapter 11 Example Configuration Files
This chapter shows how to configure the Record tool for a variety of situations:

❏
❏
❏
❏
❏
❏

11.1

How to Record All Topics in a Single Domain (Section 11.1)
How To Record a Subset of Data from Multiple Domains (Section 11.2)
How To Record Data to Multiple Files (Section 11.3)
How To Record Serialized Data (Section 11.4)
How To Record Using Best-Effort Reliability (Section 11.5)
How To Enable Remote Access (Section 11.6)

How to Record All Topics in a Single Domain
Scenario

You have a system with several nodes using domain ID 54. You want all the data in this system
to be recorded to a single file called mydomaindata. When the file is full, recording should stop.
The typecodes are available from the system.
Configuration File

mydomaindata
1 GB

mydomain

11-1

How To Record a Subset of Data from Multiple Domains

All

Expected Outcome

The expected outcome is a single file about 4 GB with all the data in a file called
mydomaindata_0_0. By default, the Record tool will store deserialized data, so the file will have
one column for each field in the topic.

11.2

How To Record a Subset of Data from Multiple Domains
Scenario

You have a system with multiple domains, including domain IDs 54 and 98, and hundreds of
topics whose names contain “Sensor” (such as TemperatureSensor, HeatSensor, SensorTypes,
etc.), in addition to hundreds of other topics. You only want to record the topics that start with
“Sensor”, and from each of these topics, you only want to record the fields whose name includes
“value” (such as value_max, value_min, current_value).
Configuration File

mydomaindata
1 GB

Sensor*

*value*

mydomain54

All

mydomain98

All

11-2

How To Record Data to Multiple Files

Expected Outcome

The expected outcome is a single file about 4 GB, with all the data in a file called
mydomaindata_0_0. By default, the Record tool will store deserialized data; so the file will have
one column per field in the topic.

11.3

How To Record Data to Multiple Files
Scenario

The Record tool is recording data on a system that supports files up to 4 GB in size. However,
you want to record more than 4 GB of data.
Configuration File

mydomaindata
2000 kB
1000
yes

...

Expected Outcome

Up to 1,000 files will be created if necessary, named mydomaindata_0_0, mydomaindata_0_1,
etc., up to mydomaindata_0_999.

11.4

How To Record Serialized Data
Scenario

Due to space limitations and speed, you want to store serialized data.
Configuration File

...

RTIDDS_DESERIALIZEMODE_NEVER

...

11-3

How To Record Using Best-Effort Reliability

Expected Outcome

All samples will be stored in a single column, along with SampleInfo and other meta-data.

11.5

How To Record Using Best-Effort Reliability
Scenario

You have a system with multiple DataWriters of the same topic. Some of these use best-effort
reliability, while others use strict reliability. You want to minimize the impact that the Record tool
has on the system.
Configuration File

...

Sensor*

*value*

BEST_EFFORT_RELIABILITY_QOS

...

Expected Outcome

The Record tool will use DataReaders with best-effort Reliability to record all data.

11.6

How To Enable Remote Access
Scenario

The Record tool is part of a larger system that must reach a steady state before it starts recording.
The Record tool should use domain ID 54 and partition “rti” for communication with the controller.
Configuration File

...

yes
10
domain54

11-4

How To Enable Remote Access

rti

...

Expected Outcome

The Record tool will communicate with a remote controller on domain ID 54 using partition
“rti.” Status information will be published every 10 seconds.

11-5

Appendix A Fields Available for Recording
This appendix lists the fields that are available for the Record tool to store in both user topic
tables and discovery tables. See Section 4.5.1 to learn how to include or exclude fields from the
recorded tables.

A.1

User Topic Tables
The fields in Table A.1 are available for storing with the user data samples being recorded in
user topic tables. Fields coming from the SampleInfo objects are prefixed with "SampleInfo_".
More information on the meaning of the fields can be found in the RTI Connext DDS API Reference HTML documentation. The units of the timestamp fields are nanoseconds.
The Record tool stores two metadata fields with the user topic tables: the domain ID and the
table prefix (see the shared_table property in Table 4.13 for more information on table prefixes).
User data fields are stored after column SampleInfo_valid_data. When recording in serialized
format, the column storing the sample is called rti_serialized_sample. The sample's endianness
and length are stored in columns rti_serialized_endian and rti_serialized_length, respectively.

Table A.1

Fields Available to Store in User Topic Tables

Field Name

SQL Type

SampleInfo_reception_timestamp

INTEGER

SampleInfo_source_timestamp

INTEGER

SampleInfo_valid_data

INTEGER

SampleInfo_publication_seq_nr

INTEGER

SampleInfo_subscription_seq_nr

INTEGER

SampleInfo_sample_state

INTEGER

SampleInfo_instance_state

INTEGER

SampleInfo_instance_handle

BLOB

SampleInfo_publication_handle

BLOB

SampleInfo_disposed_generation_count

INTEGER

SampleInfo_no_writers_generation_count

INTEGER

SampleInfo_sample_rank

INTEGER

SampleInfo_generation_rank

INTEGER

SampleInfo_absolute_generation_rank

INTEGER

A-1

DCPSParticipant Table (Discovery)

Table A.1

Fields Available to Store in User Topic Tables

Field Name

A.2

SQL Type

SampleInfo_original_publication_virtual_guid

BLOB

SampleInfo_original_publication_virtual_seq_nr

INTEGER

SampleInfo_related_original_publication_virtual_guid

BLOB

SampleInfo_flag

INTEGER

SampleInfo_source_guid

BLOB

SampleInfo_related_source_guid

BLOB

SampleInfo_related_subscription_guid

BLOB

SampleInfo_related_original_publication_virtual_seq_nr

INTEGER

Metadata_domain_id

INTEGER

Metadata_table_prefix

TEXT

DCPSParticipant Table (Discovery)
The DCPSParticipant table stores the samples of the built-in discovery topic DCPSParticipant
samples received by the Record tool. Samples of this topic are prefixed with "ParticipantData_".
Alongside every sample of the discovery topic, the Sample Info object received is also stored
(and prefixed with "SampleInfo_"). For more information on this topic, see the RTI Connext DDS
Core Libraries API Reference HTML documentation. The Record tool stores two metadata fields
with this table: the domain ID and the name of the Recorder Domain object as defined in the
recorder configuration that generated the discovery traffic.

Table A.2

Fields Available to Store in DCPSParticipant Tables

Field Name

SQL Type

SampleInfo_reception_timestamp

INTEGER

SampleInfo_source_timestamp

INTEGER

SampleInfo_valid_data

INTEGER

ParticipantData_key_0

INTEGER

ParticipantData_key_1

INTEGER

ParticipantData_key_2

INTEGER

ParticipantData_key_3

INTEGER

ParticipantData_userdata

BLOB

ParticipantData_participant_name

TEXT

ParticipantData_property

TEXT

ParticipantData_rtps_protocol_version

TEXT

ParticipantData_rtps_vendor_id

BLOB

ParticipantData_dds_builtin_endpoints

INTEGER

ParticipantData_default_unicast_locators

TEXT

ParticipantData_product_version

TEXT

SampleInfo_publication_seq_nr

INTEGER

SampleInfo_subscription_seq_nr

INTEGER

A-2

DCPSPublication Table (Discovery)

Table A.2

Fields Available to Store in DCPSParticipant Tables

Field Name

A.3

SQL Type

SampleInfo_sample_state

INTEGER

SampleInfo_instance_state

INTEGER

SampleInfo_instance_handle

BLOB

SampleInfo_publication_handle

BLOB

SampleInfo_disposed_generation_count

INTEGER

SampleInfo_no_writers_generation_count

INTEGER

SampleInfo_sample_rank

INTEGER

SampleInfo_generation_rank

INTEGER

SampleInfo_absolute_generation_rank

INTEGER

SampleInfo_original_publication_virtual_guid

BLOB

SampleInfo_original_publication_virtual_seq_nr

INTEGER

SampleInfo_related_original_publication_virtual_guid

BLOB

SampleInfo_related_original_publication_virtual_seq_nr

INTEGER

SampleInfo_flag

INTEGER

SampleInfo_source_guid

BLOB

SampleInfo_related_source_guid

BLOB

SampleInfo_related_subscription_guid

BLOB

Metadata_domain_id

INTEGER

Metadata_domain_name

TEXT

DCPSPublication Table (Discovery)
This table stores the samples of the built-in discovery topic DCPSPublication samples that are
received by Recorder. Samples of this topic are prefixed with "PublicationData_". Alongside
every sample of the discovery topic, the Sample Info object received is also stored (and prefixed
with "SampleInfo_"). For more information on this topic, see the RTI Connext DDS Core Libraries
API Reference HTML documentation. The Record tool stores two metadata fields with this table:
the domain ID and the name of the Recorder Domain object as defined in the recorder configuration that generated the discovery traffic.

Table A.3

Fields Available to Store in DCPSPublication Tables

Field Name

SQL Type

SampleInfo_reception_timestamp

INTEGER

SampleInfo_source_timestamp

INTEGER

SampleInfo_valid_data

INTEGER

PublicationData_typecode_length

INTEGER

PublicationData_topic_name

TEXT

PublicationData_type_name

TEXT

PublicationData_typecode

BLOB

PublicationData_key_0

INTEGER

A-3

DCPSPublication Table (Discovery)

Table A.3

Fields Available to Store in DCPSPublication Tables

Field Name

SQL Type

PublicationData_key_1

INTEGER

PublicationData_key_2

INTEGER

PublicationData_key_3

INTEGER

PublicationData_participant_key_0

INTEGER

PublicationData_participant_key_1

INTEGER

PublicationData_participant_key_2

INTEGER

PublicationData_participant_key_3

INTEGER

PublicationData_topic_data

BLOB

PublicationData_destination_order_kind

INTEGER

PublicationData_destination_order_scope

INTEGER

PublicationData_destination_order_source_timestamp_tolerance

INTEGER

PublicationData_presentation_access_scope

INTEGER

PublicationData_presentation_coherent_access

INTEGER

PublicationData_presentation_ordered_access

INTEGER

PublicationData_partition

TEXT

PublicationData_group_data

BLOB

PublicationData_publisher_key_0

INTEGER

PublicationData_publisher_key_1

INTEGER

PublicationData_publisher_key_2

INTEGER

PublicationData_publisher_key_3

INTEGER

PublicationData_durability_kind

INTEGER

PublicationData_durability_direct_communication

INTEGER

PublicationData_deadline_period

INTEGER

PublicationData_latency_budget_duration

INTEGER

PublicationData_liveliness_kind

INTEGER

PublicationData_liveliness_lease_duration

INTEGER

PublicationData_reliability_kind

INTEGER

PublicationData_reliability_acknowledgment_kind

INTEGER

PublicationData_reliability_max_blocking_time

INTEGER

PublicationData_ownership_kind

INTEGER

PublicationData_ownership_strength

INTEGER

PublicationData_user_data

BLOB

PublicationData_property

BLOB

PublicationData_virtual_guid

BLOB

PublicationData_rtps_protocol_version

TEXT

PublicationData_rtps_vendor_id

BLOB

PublicationData_product_version

TEXT

PublicationData_disable_positive_acks

INTEGER

PublicationData_locator_filter

TEXT

PublicationData_lifespan_duration

INTEGER

PublicationData_durability_service_service_cleanup_delay

INTEGER

A-4

DCPSSubscription Table (Discovery)

Table A.3

Fields Available to Store in DCPSPublication Tables

Field Name

A.4

SQL Type

PublicationData_durability_service_history_kind

INTEGER

PublicationData_durability_service_history_depth

INTEGER

PublicationData_durability_service_max_samples

INTEGER

PublicationData_durability_service_max_instances

INTEGER

PublicationData_durability_service_max_samples_per_instance

INTEGER

PublicationData_unicast_locators

TEXT

PublicationData_publication_name

TEXT

PublicationData_publication_role_name

TEXT

SampleInfo_publication_seq_nr

INTEGER

SampleInfo_subscription_seq_nr

INTEGER

SampleInfo_sample_state

INTEGER

SampleInfo_instance_state

INTEGER

SampleInfo_instance_handle

BLOB

SampleInfo_publication_handle

BLOB

SampleInfo_disposed_generation_count

INTEGER

SampleInfo_no_writers_generation_count

INTEGER

SampleInfo_sample_rank

INTEGER

SampleInfo_generation_rank

INTEGER

SampleInfo_absolute_generation_rank

INTEGER

SampleInfo_original_publication_virtual_guid

BLOB

SampleInfo_original_publication_virtual_seq_nr

INTEGER

SampleInfo_related_original_publication_virtual_guid

BLOB

SampleInfo_related_original_publication_virtual_seq_nr

INTEGER

SampleInfo_flag

INTEGER

SampleInfo_source_guid

BLOB

SampleInfo_related_source_guid

BLOB

SampleInfo_related_subscription_guid

BLOB

Metadata_domain_id

INTEGER

Metadata_domain_name

TEXT

DCPSSubscription Table (Discovery)
This table stores the samples of the built-in discovery topic DCPSSubscription samples that are
received by the Record tool. Samples of this topic are prefixed with "SubscriptionData_". Alongside every sample of the discovery topic, the Sample Info object received is also stored (and prefixed with "SampleInfo_"). For more information on this topic, see the RTI Connext DDS Core
Libraries API Reference HTML documentation. The Record tool stores two metadata fields with
this table: the domain ID and the name of the Recorder Domain object as defined in the recorder
configuration that generated the discovery traffic.

A-5

DCPSSubscription Table (Discovery)

Table A.4

Fields Available to Store in DCPSSubscription Tables

Field Name

SQL Type

SampleInfo_reception_timestamp

INTEGER

SampleInfo_source_timestamp

INTEGER

SampleInfo_valid_data

INTEGER

SubscriptionData_key_0

INTEGER

SubscriptionData_key_1

INTEGER

SubscriptionData_key_2

INTEGER

SubscriptionData_key_3

INTEGER

SubscriptionData_participant_key_0

INTEGER

SubscriptionData_participant_key_1

INTEGER

SubscriptionData_participant_key_2

INTEGER

SubscriptionData_participant_key_3

INTEGER

SubscriptionData_topic_name

TEXT

SubscriptionData_type_name

TEXT

SubscriptionData_destination_order_kind

INTEGER

SubscriptionData_destination_order_scope

INTEGER

SubscriptionData_destination_order_source_timestamp_tolerance

INTEGER

SubscriptionData_presentation_access_scope

INTEGER

SubscriptionData_presentation_coherent_access

INTEGER

SubscriptionData_presentation_ordered_access

INTEGER

SubscriptionData_partition

TEXT

SubscriptionData_group_data

BLOB

SubscriptionData_topic_data

BLOB

SubscriptionData_subscriber_key_0

INTEGER

SubscriptionData_subscriber_key_1

INTEGER

SubscriptionData_subscriber_key_2

INTEGER

SubscriptionData_subscriber_key_3

INTEGER

SubscriptionData_durability_kind

INTEGER

SubscriptionData_durability_direct_communication

INTEGER

SubscriptionData_deadline_period

INTEGER

SubscriptionData_latency_budget_duration

INTEGER

SubscriptionData_liveliness_kind

INTEGER

SubscriptionData_liveliness_lease_duration

INTEGER

SubscriptionData_reliability_kind

INTEGER

SubscriptionData_reliability_acknowledgment_kind

INTEGER

SubscriptionData_reliability_max_blocking_time

INTEGER

SubscriptionData_ownership_kind

INTEGER

SubscriptionData_user_data

BLOB

SubscriptionData_property

BLOB

SubscriptionData_unicast_locators

TEXT

SubscriptionData_multicast_locators

TEXT

SubscriptionData_virtual_guid

BLOB

A-6

DCPSSubscription Table (Discovery)

Table A.4

Fields Available to Store in DCPSSubscription Tables

Field Name

SQL Type

SubscriptionData_rtps_protocol_version

TEXT

SubscriptionData_rtps_vendor_id

BLOB

SubscriptionData_product_version

TEXT

SubscriptionData_disable_positive_acks

INTEGER

SubscriptionData_time_based_filter_minimum_separation

INTEGER

SubscriptionData_content_filter_property

TEXT

SubscriptionData_subscription_name

TEXT

SubscriptionData_subscription_role_name

TEXT

SampleInfo_publication_seq_nr

INTEGER

SampleInfo_subscription_seq_nr

INTEGER

SampleInfo_sample_state

INTEGER

SampleInfo_instance_state

INTEGER

SampleInfo_instance_handle

BLOB

SampleInfo_publication_handle

BLOB

SampleInfo_disposed_generation_count

INTEGER

SampleInfo_no_writers_generation_count

INTEGER

SampleInfo_sample_rank

INTEGER

SampleInfo_generation_rank

INTEGER

SampleInfo_absolute_generation_rank

INTEGER

SampleInfo_original_publication_virtual_guid

BLOB

SampleInfo_original_publication_virtual_seq_nr

INTEGER

SampleInfo_related_original_publication_virtual_guid

BLOB

SampleInfo_related_original_publication_virtual_seq_nr

INTEGER

SampleInfo_flag

INTEGER

SampleInfo_source_guid

BLOB

SampleInfo_related_source_guid

BLOB

SampleInfo_related_subscription_guid

BLOB

Metadata_domain_id

INTEGER

Metadata_domain_name

TEXT

A-7

Source Exif Data:

File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.5
Linearized                      : Yes
Author                          : adelle
Create Date                     : 2016:03:31 09:15:41Z
Modify Date                     : 2016:03:31 09:15:41Z
Tagged PDF                      : Yes
XMP Toolkit                     : Adobe XMP Core 5.2-c001 63.139439, 2010/09/27-13:37:26
Creator Tool                    : FrameMaker 10.0.2
Producer                        : Acrobat Elements 10.0.0 (Windows)
Format                          : application/pdf
Title                           : RTI_Recording_Service_UsersManual.book
Creator                         : adelle
Document ID                     : uuid:a256d783-6460-4958-b6ed-940bb35ce072
Instance ID                     : uuid:ecc057a1-a5d0-48cc-81ad-084f0bd3837e
Page Mode                       : UseOutlines
Page Count                      : 126

EXIF Metadata provided by EXIF.tools

RTI_Recording_Service_UsersManual RTI Recording Service Users Manual

Navigation menu

Versions of this User Manual:

Views

Navigation