Skip to content
This repository has been archived by the owner on Feb 4, 2021. It is now read-only.

Connext: Automatic participant index failed to initialize #93

Closed
dhood opened this issue Mar 10, 2018 · 5 comments
Closed

Connext: Automatic participant index failed to initialize #93

dhood opened this issue Mar 10, 2018 · 5 comments

Comments

@dhood
Copy link
Member

dhood commented Mar 10, 2018

https://ci.ros2.org/view/nightly/job/nightly_linux_repeated/957/testReport/junit/(root)/projectroot/test_tlsf__rmw_connext_cpp/

[D0108|ENABLE]DDS_DomainParticipant_enableI:Automatic participant index failed to initialize. PLEASE VERIFY CONSISTENT TRANSPORT / DISCOVERY CONFIGURATION. 
DDSDomainParticipant_impl::createI:ERROR: Failed to auto-enable entity

This post says it's from a network interface configuration issue.

Failed on repeat 12 and there have been CI jobs not showing this error since then so it seems like bad luck.

@dhood dhood closed this as completed Mar 10, 2018
@dhood
Copy link
Member Author

dhood commented Mar 10, 2018

@dhood
Copy link
Member Author

dhood commented Mar 16, 2018

This is still an issue, though it has low probability. It can interfere with testing of flaky tests where we're repeating many times (e.g. it failed on repeat 417 in this flakiness-testing build: https://ci.ros2.org/job/ci_linux/4085).

The response on the forum was:

The error that you are facing is due there is not interface up and running when the participant is created.
So, there are two workarounds:

  1. There is always an interface up and running on your system
  2. It is possible to set the <participant_id> QoS manually. You can find more information in the “8.5.9 WIRE_PROTOCOL QosPolicy (DDS Extension)” section of the User’s Manual.
    Example:
<participant_qos>
    <wire_protocol>
        <participant_id>1</participant_id>
    </wire_protocol>
</participant_qos>

Setting the participant ID manually isn't really viable even if it's do-able for tests. This should be resolved for users that want automatic participant ID selection.

@ros2/team Is anyone aware of issues that would impact the interface availability in the docker containers? Is there something we can improve there, or -- given that the interface is fine for all other runs of the tests -- is this more likely an issue in how connext is working with the interface?

@dhood
Copy link
Member Author

dhood commented Mar 20, 2018

@dhood
Copy link
Member Author

dhood commented Apr 3, 2018

I asked for specifics on what changed between connext versions but haven't heard back: https://community.rti.com/forum-topic/please-verify-constistent-transport-discovery-configuration

@dirk-thomas
Copy link
Member

Since we haven't seen this for quite a while I will go ahead and close the ticket for now.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants