race: podman-remote stop --cidfile --ignore does very weird things

[edsantiago]

I can't find a way to reproduce this actual bug but it is super reproducible in BATS:

<+014ms> # # podman-remote --url unix:///tmp/CI_TVd8/podman_tmp_zsap stop --cidfile=/tmp/CI_TVd8/podman_bats.4OGhmH/no-such-file --ignore
<+9.98s> # 0090f32ea149f1e6ebc63979910c705b82365860d95f8b034848759ee47e3c12
         # Error: no container with name or ID "47158e686535dd8c907085d24b480969c50ddaebc880152c96a9abd0417d51a3" found: no such container

That is: wtfff?

This sort of reproduces it after a long time (1-2 minutes) but still, wtf??

Window 1:

$ podman system service --timeout=0

Window 2:

$ while :;do bin/podman run --rm quay.io/libpod/testimage:20240123 date;done

Window 3:

while :;do bin/podman-remote stop --cidfile /this/does/not/exist --ignore || break;done

Window 3 spits out CID after CID after CID ??? Then, after a while, it finally dies

Error: no container with ID c0c5b067c03004b53fa800938747e6842c0966b5352a8d477ebe1a8ece3b2a0a found in database: no such container

[Luap99]

so when ignore is used the error reading the file will be ignored, then we call into the backend without any containers which should not do anything. This seems to work locally, remote however thinks that means it should stop all containers.