Implement helper for processing specific failed CI jobs

[cevich]

This is an incredibly basic implementation, and a simple "does it run" test. It is needed in the short-term for use in [other scripts], to help process and debug incidents of "CI Agent Stopped Responding" tasks. Adding it into this repo. as a reference for future similar/related scripts.

Ref: containers/podman#8068 Intended use by [other scipts]:

• Allow debugging of leftover VMs for which a task failed due to "CI Agent Stopped Responding" (see Cirrus CI: agent stopped responding podman#8068 (comment))
• Allow cleanup of all leftover VMs for which a task failed due to any reason.

[github-actions]

Successfully triggered github-actions/success task to indicate successful run of cirrus-ci_retrospective integration and unit testing from this PR's 24da5bec8f9a95f8d658e60320920ca11e6fa702.

[github-actions]

Successfully triggered github-actions/success task to indicate successful run of cirrus-ci_retrospective integration and unit testing from this PR's 09ca62a11a03b1684f74fd5d518f4c791212aceb.

[cevich]

hrmmm...tests not working right...

[github-actions]

Successfully triggered github-actions/success task to indicate successful run of cirrus-ci_retrospective integration and unit testing from this PR's 42fc88c5c18f324bd7f9a090416936cdadf915dc.

[github-actions]

Successfully triggered github-actions/success task to indicate successful run of cirrus-ci_retrospective integration and unit testing from this PR's 5d872906807fe80132992d2f5afd47ccbedd2fae.

[github-actions]

Successfully triggered github-actions/success task to indicate successful run of cirrus-ci_retrospective integration and unit testing from this PR's b1ab5020a6866330bb60e79739b5e1073c5e6036.

[github-actions]

Successfully triggered github-actions/success task to indicate successful run of cirrus-ci_retrospective integration and unit testing from this PR's 7624b3d308db2292bfde0fd442f759035f7c2f75.

[cevich]

@edsantiago PTAL when you have a moment. Note: This is very basic and minimal code, but functional as-is: https://cirrus-ci.com/task/5343953614536704?command=main#L669-L679 I expect it will either be improved/better tested in the future or completely removed if useless.

[cevich]

Update: Fixed up README + altered output format to include task names: Example output: https://cirrus-ci.com/task/5065071724331008?command=main#L679-L669

[github-actions]

Successfully triggered github-actions/success task to indicate successful run of cirrus-ci_retrospective integration and unit testing from this PR's 585c5da3c5271e177969da39ed62e936286a7f5c.

[cevich]

nvm Ed, it's probably not worth your time.

[edsantiago]

OK, thanks. The pip requirement blocked me from testing on my laptop; I was planning to review on a virt later this morning. The only suggestion I was going to make (which is as far as I got) was that it would seem friendlier to use a PR as argument, not a long Cirrus ID.

[cevich]

@edsantiago Near as I can tell, if you run pip3 with --user as indicated, you can just clean out the referenced directory after. This works properly as both a user and as root. The other common practice is to use virtualenv (very simple tool) in a temp-dir, activate it, then pip3 will be sensitized to use it instead.

Anyway...I seriously considered using a PR instead of commit, but was in a rush to debug that agent-stopped-responding problem, and not confident in my ability to quickly code translation from PR no. -> latest SHA. Certainly a great idea for future improvement though. Thanks.

[cevich]

oh...also, the only reason pip3 is even required is for the gql module, I looked and looked but could not find an RPM for it. Though I didn't look far beyond the default/standard repos. (note: python3-py-gql.noarch is confusingly, not it).