Failure to clone submodules if job gets aborted early

Generated by Generative AI

No

Operating System:

<PUT OS INFO HERE>

ROS version or commit hash:

rolling

RMW implementation (if applicable):

No response

RMW Configuration (if applicable):

No response

Client library (if applicable):

No response

'ros2 doctor --report' output

<COPY OUTPUT HERE>

Steps to reproduce issue

1. Start a job on an agent on ci.ros2
2. Abort it while it's cloning the repositories.
3. Run a second job on the same agent on ci.ros2
4. Jobs fails with failure to clone submodules.

Expected behavior

Job abort should not have cascading effects on subsequent builds.

Actual behavior

Workspace stays in a broken state making subsequent git clone request fail.

Additional information

Reported by @skyegalaxy happening both on ci_linux, ci_linux-rhel.

Failing job : https://ci.ros2.org/job/ci_linux-rhel/9128/

 > git submodule update --init --recursive linux_docker_resources/rticonnextdds-src # timeout=10
20:21:03 hudson.plugins.git.GitException: Command "git submodule update --init --recursive linux_docker_resources/rticonnextdds-src" returned status code 1:
20:21:03 stdout: 
20:21:03 stderr: error: The following untracked working tree files would be overwritten by checkout:
20:21:03 	README.md
20:21:03 	openssl-1.1.1k-6.0.1.25-host-x64Linux.rtipkg
20:21:03 	openssl-3.0.12-7.3.0-host-x64Linux.rtipkg
20:21:03 	openssl-3.0.12-7.3.0-target-x64Linux4gcc7.3.0.rtipkg
20:21:03 	openssl-3.5.5-7.7.0-host-x64Linux.rtipkg
20:21:03 	openssl-3.5.5-7.7.0-target-x64Linux4gcc8.5.0.rtipkg
20:21:03 	rti_connext_dds-6.0.1-pro-host-x64Linux.run.000
20:21:03 	rti_connext_dds-6.0.1-pro-host-x64Linux.run.001
20:21:03 	rti_connext_dds-6.0.1-pro-host-x64Linux.run.002
20:21:03 	rti_connext_dds-6.0.1-pro-host-x64Linux.run.003
20:21:03 	rti_connext_dds-6.0.1.25-pro-host-x64Linux.rtipkg.000
20:21:03 	rti_connext_dds-6.0.1.25-pro-host-x64Linux.rtipkg.001
20:21:03 	rti_connext_dds-6.0.1.25-pro-target-x64Linux4gcc7.3.0.rtipkg.000
20:21:03 	rti_connext_dds-6.0.1.25-pro-target-x64Linux4gcc7.3.0.rtipkg.001
20:21:03 	rti_connext_dds-7.3.0-pro-host-x64Linux.run.001
20:21:03 	rti_connext_dds-7.3.0-pro-host-x64Linux.run.002
20:21:03 	rti_connext_dds-7.3.0-pro-host-x64Linux.run.003
20:21:03 	rti_connext_dds-7.3.0-pro-host-x64Linux.run.004
20:21:03 	rti_connext_dds-7.3.0-pro-host-x64Linux.run.005
20:21:03 	rti_connext_dds-7.3.0-pro-host-x64Linux.run.006
20:21:03 	rti_connext_dds-7.3.0-pro-host-x64Linux.run.007
20:21:03 	rti_connext_dds-7.3.0-pro-target-x64Linux4gcc7.3.0.rtipkg.001
20:21:03 	rti_connext_dds-7.3.0-pro-target-x64Linux4gcc7.3.0.rtipkg.002
20:21:03 	rti_connext_dds-7.3.0-pro-target-x64Linux4gcc7.3.0.rtipkg.003
20:21:03 	rti_connext_dds-7.3.0-pro-target-x64Linux4gcc7.3.0.rtipkg.004
20:21:03 	rti_connext_dds-7.7.0-pro-host-x64Linux.run.000
20:21:03 Please move or remove them before you switch branches.
20:21:03 Aborting

Aborted job: https://ci.ros2.org/job/ci_linux-rhel/9127/

18:40:03  > git submodule update --init --recursive linux_docker_resources/rticonnextdds-src # timeout=10
18:44:35 $ ssh-agent -k
18:44:35 unset SSH_AUTH_SOCK;
18:44:35 unset SSH_AGENT_PID;
18:44:35 echo Agent pid 237946 killed;
18:44:35 [ssh-agent] Stopped.
18:44:35 Build was aborted

Start with the aborted ci_linux-rhel/9127 and failing ci_linux-rhel/9128 logs, focusing on the git submodule update for linux_docker_resources/rticonnextdds-src. Trace how the CI job handles an aborted clone and reuses the agent workspace. Done means a job aborted during cloning leaves the workspace usable and the next job clones its submodules successfully.