podman-remote: incorrect exit status of 'run --rm'

[edsantiago]

Like #3795 but with podman-remote instead. Unclear if this is a regression because we've never tested it.

$ podman-remote run --rm alpine /bin/false
$ echo $?
0    <----- expected 1

Also:

window1# podman-remote run --name foo alpine sleep 30

window2# podman-remote rm -f foo   [will take 10 seconds]

window1# echo $?          # after waiting those 10 seconds for prompt to return
0      <------ expected 137, because the container was terminated

[rhatdan]

This works if you do a podman exec, but not run. We must have the checks for this turned off for remote containers.

[github-actions]

This issue had no activity for 30 days. In the absence of activity or the "do-not-close" label, the issue will be automatically closed within 7 days.

[rhatdan]

We want this fixed, I don't know if it is still a bug. I think it is depending on how you setup podman-remote.

[edsantiago]

I think this is fixed, or at least partly so:

$ ./bin/podman-remote run --rm alpine /bin/false
$ echo $?
1       ! yay

$ ./bin/podman-remote run --rm alpine sh -c 'exit 4'
$ echo $?
4       ! yay

However, there's a new weirdness going on, and I think this may be related to some CI failures we were discussing last month: podman-remote rm -f is returning immediately, even before the container exits:

[10:07:52]$ ./bin/podman-remote run --name foo alpine sleep 30 &
[1] 11804
[10:08:09]$ ./bin/podman-remote rm -f foo
257088a67990163317762fdaee2817a9472770e9c22048b59d561901832b86ee
[10:08:10]$
[1]  + exit 137   ./bin/podman-remote run --name foo alpine sleep 30
[10:08:15]$

That is: I background-run a sleep container and immediately try to rm it. The rm command exits immediately, and then, five seconds later, the container exits. The container exit status is correct, I'm just surprised that podman-remote rm doesn't wait for the exit. Weren't we looking into a related problem recently in CI?

Leaving issue open for now because (1) I want to make sure this behavior is intentional, (2) I want to remind myself to reenable the podman-remote exit-status CI checks so this stays fixed, and (3) I'd like opinions on whether I should file this comment as a separate issue.

[edsantiago]

Update: The exit-status bug was fixed in #3934, and CI tests were enabled then for podman-remote. Thanks, Dan.

Closing, but someone please lmk if the above immediate-exit concern deserves an issue.