argo-workflows: v3.0: wait containers do not exit
Users are reporting that wait containers do not exit after the main container.
Please make sure that you do not report problems with <= v3.0.0-rc6 - these have know issues with killing injected sidecars.
Questions:
- What executor are you using?
- Do you have sidecars?
- Do you have injected sidecars (e.g. Istio)?
- Are you using outputs artifacts, outputs.result?
- What is the last 10 log lines of the wait container?
- Are you archiving logs? i.e
archiveLogs: true
About this issue
- Original URL
- State: closed
- Created 3 years ago
- Comments: 15 (15 by maintainers)
Commits related to this issue
- fix(executor): Always poll for Docker injected sidecars. Resolves #5448 (#5479) — committed to argoproj/argo-workflows by alexec 3 years ago
- fix(executor): Always poll for Docker injected sidecars. Resolves #5448 (#5479) — committed to argoproj/argo-workflows by alexec 3 years ago
We have now run about 20 different workflows through with your dev version of argoexec and haven’t found the problem once. I think from our point of view its safe to say that this is resolved.
After debugging with @tim-sendible on Zoom, it appears that the Vault sidecar sometimes do not start for > 5 seconds after the wait container starts. I’ve pushed a new build of
argoproj/argoexec:dev-dockerthat allows 30s for sidecars to start. Can you please test?