Closed edsantiago closed 7 months ago
Theory: We're doing something with libimage after system reset
deleted all directories associated with Podman, causing a panic as we can't create lock files in deleted directories.
could be a race condition where the go routine runs after the r.libimageRuntime.Shutdown()
and we should wait for the goroutine to exit before .libimageRuntime.Shutdown()
exits?
Events being a goroutine is unnecessary, usually, we can probably just convert it to a direct call to add an event
Could this be related to #17957?
A friendly reminder that this issue had no activity for 30 days.
No further instances seen since this January 17 one. Searching my flake logs I see two others, both in March 2023, but those logs have long since evaporated. Copied from my logs:
[+1399s] Running: podman [...] rm -fa -t 0
[+1399s] 62942d61068a7046a3ff6df59c6e16ce33107761e2643a24f5063ed8e37b995e
[+1399s] output: 62942d61068a7046a3ff6df59c6e16ce33107761e2643a24f5063ed8e37b995e
[+1399s] panic: creating lock file directory: mkdir /tmp/podman_test2754512739/root/vfs-containers: no such file or directory
and
[+0940s] Verify podman containers.conf usage
[+0940s] using journald for container with container log_tag
[+0940s] /var/tmp/go/src/github.com/containers/podman/test/e2e/containers_conf_test.go:230
[no "Running" lines in log]
[+0940s] panic: creating lock file directory: mkdir /tmp/podman_test2316220714/runroot/vfs-layers: no such file or directory
Not seen again since Jan 17.
Seen just now in f38 containerized: