Hello
When I try to stop a 9-based container such as rockylinux9 or ubi9 with podman, the following message appears.
Do you know the cause and solution?
●Command
$ podman container stop containername
●Message
StopSignal SIGTERM failed to stop container containername in 10 seconds, resorting to SIGKILL
●Host OS
・Rocky Linux 9.5 (Blue Onyx)
・podman version 5.2.2
Thank you in advance.
Your process inside the container doesn’t honor the signal. The question is therefore, what process do you start inside the container (as PID 1).
I have checked several versions of 8 and 9. In all cases, bash is PID 1. I don’t know why only version 9 displays the following warning.
$ podman container ls
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
d7175e10a772 docker.io/rockylinux/rockylinux:9.5 /bin/bash 27 seconds ago Up 27 seconds con-9.5
46e45c847240 docker.io/rockylinux/rockylinux:8.10 /bin/bash 19 seconds ago Up 19 seconds con-8.10
7290bff8bf90 docker.io/rockylinux/rockylinux:9.1 /bin/bash 11 seconds ago Up 11 seconds con-9.1
2776197e4e5a docker.io/rockylinux/rockylinux:8.4 /bin/bash 5 seconds ago Up 5 seconds con-8.4
$ podman container top con-9.5
USER PID PPID %CPU ELAPSED TTY TIME COMMAND
root 1 0 0.000 35.432794893s pts/0 0s /bin/bash
$ podman container top con-8.10
USER PID PPID %CPU ELAPSED TTY TIME COMMAND
root 1 0 0.000 30.984995722s pts/0 0s /bin/bash
$ podman container top con-9.1
USER PID PPID %CPU ELAPSED TTY TIME COMMAND
root 1 0 0.000 28.40745646s pts/0 0s /bin/bash
$ podman container top con-8.4
USER PID PPID %CPU ELAPSED TTY TIME COMMAND
root 1 0 0.000 25.120534778s pts/0 0s /bin/bash
$
$ podman container stop con-9.5
WARN[0010] StopSignal SIGTERM failed to stop container con-9.5 in 10 seconds, resorting to SIGKILL
con-9.5
$ podman container stop con-8.10
con-8.10
$ podman container stop con-9.1
WARN[0010] StopSignal SIGTERM failed to stop container con-9.1 in 10 seconds, resorting to SIGKILL
con-9.1
$ podman container stop con-8.4
con-8.4
$ podman container ls
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
$
I was able to confirm that if I add “–stop-signal SIGHUP” when running the container (podman container run), it will terminate without warning when I run podman stop.
I thought this was because PID1 was bash and the TERM signal was ignored, but I don’t understand why the 8 series container terminated normally.
I checked “/proc/1/status” in 8 and 9.
8: SigIgn: 0000000000380004
9: SigIgn: 0000000000384004
Based on this alone, it seems that 9 ignores SIGTERM, but 8 does not. I thought this might be related.
You can use the following cmds to inspect your containers setup, the container does NOT have to be running :
show the command that runs by default in the container :
podman inspect --format "{{ .Config.Cmd }}" <container>
show the default stop signal used in the container :
podman inspect --format "{{ .Config.StopSignal }}" <container>
You can also pass cli parameters to alter the behavoir at run time, or edit the container or the compose/container file, edit the image with other cmds as well.