podman stop: rootless netns ref counter out of sync, counter is at -1, resetting it back to 0 #21569
Description
<+012ms> # $ podman pod rm test_pod
<+328ms> # time="2024-02-08T08:28:16-06:00" level=error msg="rootless netns ref counter out of sync, counter is at -1, resetting it back to 0"
# 2ef9c5c99f00484285a5a7394e9d276d97c436311f197e7d2a0fa05ea78e69b5
Seen mostly in system tests, BUT ALSO IN E2E TESTS WHERE IT IS IMPOSSIBLE TO DETECT. Because it happens in test cleanup, where rm does not trigger a test failure. So I caught all these by accident. THIS IS HAPPENING A LOT MORE THAN WE ARE AWARE OF.
- fedora-38 : sys podman fedora-38 rootless host boltdb
- PR Upgrade tests: reenable, but revamped #21535
- 02-08 09:30 in [sys] [700] podman play --service-container
- PR Upgrade tests: reenable, but revamped #21535
- fedora-39 : int podman fedora-39 rootless host sqlite
- 02-06 08:21 in Podman kube play test with reserved init annotation in yaml
- 01-30 09:16 in Podman kube play override with tcp should keep udp from YAML file
- 01-18 08:40 in Podman kube play secret as volume support - multiple volumes
- fedora-39β : sys podman fedora-39β rootless host sqlite
- PR CI: test overlay and vfs #20161
- 10-17-2023 15:39 in [sys] podman play --service-container
- PR CI: test overlay and vfs #20161
- rawhide : int podman rawhide rootless host sqlite
- 02-06 11:21 in Podman kube play secret as volume support - multiple volumes
- 02-04 12:39 in Podman kube play test restartPolicy
- rawhide : sys podman rawhide rootless host sqlite
- PR AppleHV: update LastUp time #21246
- 01-12 13:11 in [sys] [700] podman play
- PR AppleHV: update LastUp time #21246
| x | x | x | x | x | x |
|---|---|---|---|---|---|
| int(5) | podman(8) | rawhide(3) | rootless(8) | host(8) | sqlite(7) |
| sys(3) | fedora-39(3) | boltdb(1) | |||
| fedora-39β(1) | |||||
| fedora-38(1) |