[Podman] Re: Newbie network test?

On 7/12/2022 12:41 PM, Chris Evich wrote: --- I'm not an expert, but this seems like a significant clue. Every single one of your httpd containers exits like this. I think something's wrong/broken with the image itself. Try something simpler, like running a simple busybox or alpine image, with port redirects. --- Searching led to this: https://github.com/containers/podman/issues/1118 ----- [vagrant@localhost ~]$ sudo -s [root@localhost vagrant]# podman run busybox echo hello world [root@localhost vagrant]# echo $? 139 podman version 0.7.1 Can you add a --log-level=debug between podman and run to get debug-level logs and try and see what's going wrong? ----- I get the 139 fail even without the port redirect: pi@raspberrypi:~/tripod $ podman --version podman version 3.0.1 pi@raspberrypi:~/tripod $ podman run --rm -it busybox echo hello world pi@raspberrypi:~/tripod $ echo $? 139 But lots of debug runs on the web have that crun/runc/kata error 139 and it doesn't seem to be the main problem... Regular echo works: pi@raspberrypi:~/tripod $ echo hello world hello world pi@raspberrypi:~/tripod $ podman --log-level=debug run busybox echo hello world [Full output in attached file...] ... INFO[0000] Error initializing configured OCI runtime runc: no valid executable found for OCI runtime runc: invalid argument INFO[0000] Error initializing configured OCI runtime kata: no valid executable found for OCI runtime kata: invalid argument ... [conmon:d]: failed to write to /proc/self/oom_score_adj: Permission denied ... EBU[0002] Received: 7703 INFO[0002] Got Conmon PID as 7700 DEBU[0002] Created container 96483125cc3a225c33a354bf903970705b0d89daa66f3cbc3fe6d65a36269e5b in OCI runtime DEBU[0002] Attaching to container 96483125cc3a225c33a354bf903970705b0d89daa66f3cbc3fe6d65a36269e5b DEBU[0002] Starting container 96483125cc3a225c33a354bf903970705b0d89daa66f3cbc3fe6d65a36269e5b with command [echo hello world] DEBU[0002] Started container 96483125cc3a225c33a354bf903970705b0d89daa66f3cbc3fe6d65a36269e5b DEBU[0002] Enabling signal proxying DEBU[0002] Called run.PersistentPostRunE(podman --log-level=debug run busybox echo hello world) pi@raspberrypi:~/tripod $ https://devopscube.com/podman-tutorial-beginners/ ----- If you don’t have runc installed, you might get the following error. Error: default OCI runtime "runc" not found: invalid argument ----- pi@raspberrypi:~/tripod $ sudo apt-get -y install runc [Apparently successful] --> runc error gone, but still: INFO[0000] Error initializing configured OCI runtime kata: no valid executable found for OCI runtime kata: invalid argument DEBU[0000] using runtime "/usr/bin/crun" DEBU[0000] using runtime "/usr/bin/runc" Seems to be a hot mess of confusion about crun vs. runc - sites saying one or the other is correct, that the problems are fixed, and that they really aren't... Looks like I need both! https://github.com/kata-containers/documentation/blob/master/install/snap... ----- Kata Containers can be installed in any Linux distribution that supports snapd. NOTE: From Kata Containers 2.x, only the Containerd Runtime V2 (Shim API) is supported, note that some container engines (docker, podman, etc) may not be able to run Kata Containers 2.x. ----- Clearly a 32-bit Pi can't run Kata Containers 2.x. Apparently the 1.x version can supposedly be installed: pi@raspberrypi:~ $ sudo apt-get -y install snapd ... Job for snapd.service failed because the control process exited with error code. See "systemctl status snapd.service" and "journalctl -xe" for details. --- Jul 12 16:51:04 raspberrypi systemd[1]: snapd.service: Failed with result 'exit-code'. Jul 12 16:51:04 raspberrypi systemd[1]: Failed to start Snap Daemon. Jul 12 16:51:04 raspberrypi systemd[1]: snapd.service: Triggering OnFailure= dependencies. Jul 12 16:51:04 raspberrypi systemd[1]: snapd.service: Failed to enqueue OnFailure= job, ignoring: Unit snapd.failure.service not found. --- --- Jul 12 14:08:50 raspberrypi /usr/bin/podman[7705]: time="2022-07-12T14:08:50-07:00" level=debug msg="Initializing event backend journald" Jul 12 14:08:50 raspberrypi /usr/bin/podman[7705]: time="2022-07-12T14:08:50-07:00" level=info msg="Error initializing configured OCI runtime kata: no valid executable found for OCI runtime kata: invalid argument" Jul 12 14:08:50 raspberrypi /usr/bin/podman[7705]: time="2022-07-12T14:08:50-07:00" level=debug msg="using runtime \"/usr/bin/crun\"" Jul 12 14:08:50 raspberrypi /usr/bin/podman[7705]: time="2022-07-12T14:08:50-07:00" level=info msg="Error initializing configured OCI runtime runc: no valid executable found for OCI runtime runc: invalid argument" ... Jul 12 16:07:45 raspberrypi conmon[8181]: conmon b2b1159879569be673d9 <nwarn>: stdio_input read failed Input/output error ... ░░ The unit UNIT has successfully entered the 'dead' state. Jul 12 16:50:59 raspberrypi snapd[8454]: cannot run daemon: state startup errors: [cannot obtain snap-seccomp version information: Jul 12 16:50:59 raspberrypi snapd[8454]: ----- Jul 12 16:50:59 raspberrypi snapd[8454]: 2022/07/12 16:50:59 unknown kernel arch "armv6l" Jul 12 16:50:59 raspberrypi snapd[8454]: panic: unknown kernel arch "armv6l" --- Looks to me like running this on a 32-bit Pi is just not going to work, despite claims it can! Loren On 7/12/2022 12:41 PM, Chris Evich wrote: > On 7/8/22 20:32, Loren Amelang wrote: >> > I'm not an expert, but this seems like a significant clue.  Every single one of your httpd containers exits like this.  I think something's wrong/broken with the image itself. > > Try something simpler, like running a simple busybox or alpine image, with port redirects.  You can easily do client/server debugging interactively using nc (or netcat or ncat, whichever is available). > > Point is, see if you can get communication flowing across a redirected port in a very simple setup first.  If this works, then it suggests the problem is inside the httpd image (or your usage of it). >