Use retry logic from containers/common

[rhatdan]

What type of PR is this?

/kind api-change
/kind bug
/kind cleanup
/kind deprecation
/kind design
/kind documentation
/kind failing-test
/kind feature
/kind flake
/kind other

What this PR does / why we need it:

How to verify it

Which issue(s) this PR fixes:

Special notes for your reviewer:

Does this PR introduce a user-facing change?

Default values for --retry and --retry-delay can be set in containers.conf

[rhatdan]

@edsantiago Did the new crun make it into this VM release?

[edsantiago]

Post-2024-01-02 VMs are in podman only, due to the rawhide hang. I am building new VMs right this moment, hoping to get an rc6 kernel. It will take ~2h to find out.

You can, if you wish, bump up to c20240222t143004z-f39f38d13 here for testing purposes. I don't recommend merging with those VMs because CI is likely to need many, many reruns.

[rhatdan]

Already bumped to c20240222t143004z-f39f38d13 and still seeing crun errors?

[edsantiago]

Oh, sorry - I didn't actually look at your diffs. My bad.

The error I see is

[+0050s]   Error: unable to parse value provided "" as --retry-delay: time: invalid duration ""

No wait I see crun error in conformance. It looks like conformance is using debian??????? Debian and crun??? That seems weird. And yes, crun is 1.14.3 on all fedoras but 1.14.1 in debian. Do you know why conformance is using debian + crun (instead of runc)?

[rhatdan]

No idea, that seems wrong. I will look at the retry-delay one.

[edsantiago]

I don't know what to do here. Best I can tell, crun-1.14.3 is not even built yet for debian. And I can't figure out where the "use crun in debian" logic is coming from.

[edsantiago]

This is weird, in setup:

[+0003s] dpkg: considering removing runc in favour of containerd.io ...
[+0003s] dpkg: yes, will remove runc in favour of containerd.io

Next step is to see if this happens in older CI runs with older VMs. I will do so.

UPDATE: yes, this is happening right now. Next question: when did this (removing runc) start happening? That will be much harder to answer.

UPDATE 2: it happened December 22 2023, which is before January 2 2024, therefore it is not something new in c20240201t143038z-f39f38d13. We do not have CI logs beyond 60 days so I'm calling this a dead end.

Situation:

• buildah conformance tests have been running on debian + crun for at least two months
  • they have been running just fine
• As of this writing (2024-02-26), debian has a broken crun. There is no fixed crun anywhere in the pipeline AFAICT.

Recommendation: find a Debian maintainer who will build crun-1.14.3. My only clue here is something called Salsa, which suggests that Faidon Liambotis at Debian might be a good person to contact.

UPDATE 3: I've sent email to Liambotis.

[rhatdan]

@edsantiago I turned off the update for runtime-spec until crun gets updated for Debian. This will allow us to merge this and get buildah updated in podman.

[rhatdan]

@containers/podman-maintainers PTAL
@nalind @flouthoc @umohnani8 @giuseppe @vrothberg @ashley-cui @TomSweeneyRedHat @mheon @baude @Luap99 PTAL

[rhatdan]

@lsm5 PTAL, once this gets in your PR can probably get in as well.

[flouthoc]

LGTM

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: flouthoc, rhatdan

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [flouthoc,rhatdan]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[vrothberg]

Code and tests LGTM but the replace caught my eyes.

[vrothberg]

Why the replace?

[rhatdan]

crun on debian is not updated to a version that will work, so in order to move forward we are not advancing the runtime-spec. As soon as debian moves forward this can be removed.

This will not effect the effect the eventual podman build PR

[TomSweeneyRedHat]

Same question as @vrothberg about the replace, that was unexpected.
LGTM otherwise. The branch is out of date, I'll update it now.