QA (public) » openQA Project (public)

Motivation¶

In https://openqa.opensuse.org/tests/2335254#comments we observed that there are two comments:

system wrote 2022-05-09 11:39:58 +0000
Restarting because RETRY is set to 2 (and only restarted 0 times so far)

and then 8s later

geekotest wrote 2022-05-09 11:40:06 +0000
Automatic investigation jobs

so openQA automatically triggered a retry because we know of the presence of sporadic issues and signal the intent to retry to mitigate hence the investigation and also the notification should be avoided.

Acceptance criteria¶

• AC1: No investigation jobs and notifications are triggered on jobs that are automatically retried by openQA
• AC2: investigation jobs and notifications are still triggered on the last retried job if it fails with an unknown issue

Suggestions¶

• Understand when and where we trigger the job done hooks and retrying
  • retry is handled in https://github.com/os-autoinst/openQA/blob/master/lib/OpenQA/Schema/Result/Jobs.pm#L2032
  • job_done_hooks are triggered as part of https://github.com/os-autoinst/openQA/blob/master/lib/OpenQA/Schema/Result/Jobs.pm#L2022
• It needs to be checked if we can simply move the retry call further above so that it happens before we call any job_done_hooks or if this breaks something else. If it works, good, if not then we need to find a different way to ensure that openqa-investigate only handles jobs that would not be restarted anyway.