Jobs fail with error code 143

[Mehdi-Bendriss]

Description

For a job where I allocate LXD containers for an integration test. The job consistently fails as soon as I get to the 3rd allocated unit, with an error 143.

Following this similar thread, It seems it has to do with the resources of the runner.

Something worth mentioning is that I disable the swappiness of the runner due to the requirements of the software running sysctl -w vm.swappiness = 0.

It doesn't seem like there is a straightforward way to investigate it from my side. Would it be possible, in the case of this specific job, to know exactly what the culprit is?

Platforms affected

• Azure DevOps
• GitHub Actions - Standard Runners
• GitHub Actions - Larger Runners

Runner images affected

• Ubuntu 18.04
• Ubuntu 20.04
• Ubuntu 22.04
• macOS 11
• macOS 12
• Windows Server 2019
• Windows Server 2022

Image version and build link

Image: ubuntu-22.04
Version: 20230206.1

Failed Job
Indefinitely hanging job

Is it regression?

No - it's been failing consistently when provisioning a 3th LXD container. Occasionally the job will run forever (up to ~6h or so)

Expected behavior

Build / integration tests should pass

Actual behavior

Error 143 when trying to allocate a new (3rd) LXD container on the runner.
In some other case, it will indefinitely hand (for hours)

Repro steps

Run the Integration tests for HA (lxd) job in the branch of the repo, i.e via a PR -- the pipeline yaml can be found here

[mikhailkoliada]

Hello, please provide minimal repro-steps and yaml snippets to reproduce the issue :)

[Mehdi-Bendriss]

@mikhailkoliada updated, thanks :)

[erik-bershel]

Hey @Mehdi-Bendriss! May I ask you to run the same pipeline again once? I need it to get new logs for investigation.

[mikhailkoliada]

@Mehdi-Bendriss please feel free to reach us again if the problem persists with a link of a workflow run, we will gladly help you. I am going to close it for now due to inactivity.