LXC on PVE7 crashes with UNSUPPORTED(1)

[amace]

A few months ago I installed new hardware and migrated our old systems from pve6 to pve7. One LXC container, a lightly loaded email server, crashes after 5-10 days putting the whole server filesystem in "read-only" mode. It effects all other LXC/VM on the machine and the only recovery is reboot. I haven't seen any errors in the usual logs so I ran the container in debug mode and logged to a file. After 6 days, right after the backup finishes successfully, I see this:

lxc-start 2313 20220821004802.840 NOTICE start - ../src/lxc/start.c:signal_handler:449 - Received 17 from pid 2258 instead of container init 2262
MARK - BOOTED SUCESSFULLY -----------------------------------------------------
lxc-start 2313 20220827051523.570 DEBUG start - ../src/lxc/start.c:signal_handler:467 - Container init process 2262 exited
lxc-start 2313 20220827051523.571 DEBUG start - ../src/lxc/start.c:__lxc_start:2107 - UNSUPPORTED(1) - Container "2313" is rebooting
lxc-start 2313 20220827051523.571 INFO error - ../src/lxc/error.c:lxc_error_set_and_log:34 - Child <2262> ended on signal UNSUPPORTED(1)
lxc-start 2313 20220827051523.673 INFO network - ../src/lxc/network.c:lxc_delete_network_priv:3680 - Removed interface "veth2313i0" from ""
lxc-start 2313 20220827051523.674 DEBUG network - ../src/lxc/network.c:lxc_delete_network:4173 - Deleted network devices
lxc-start 2313 20220827051524.533 WARN cgfsng - ../src/lxc/cgroups/cgfsng.c:cgroup_tree_remove:483 - Read-only file system - Failed to destroy 10(lxc/2313)
lxc-start 2313 20220827051524.534 ERROR cgfsng - ../src/lxc/cgroups/cgfsng.c:__cgroup_tree_create:711 - Read-only file system - Failed to create 10(lxc.pivot)
lxc-start 2313 20220827051524.535 WARN cgfsng - ../src/lxc/cgroups/cgfsng.c:cgfsng_monitor_destroy:901 - Read-only file system - Failed to create pivot cgroup 10(lxc.pivot)
lxc-start 2313 20220827051524.559 INFO conf - ../src/lxc/conf.c:run_script_argv:337 - Executing script "/usr/share/lxc/hooks/lxc-pve-poststop-hook" for container "2313", config section "lxc"
lxc-start 2313 20220827051528.669 ERROR conf - ../src/lxc/conf.c:run_buffer:321 - Script exited with status 1
lxc-start 2313 20220827051528.669 ERROR start - ../src/lxc/start.c:lxc_end:988 - Failed to run lxc.hook.post-stop for container "2313"
lxc-start 2313 20220827051528.669 WARN start - ../src/lxc/start.c:lxc_end:990 - Container will be stopped instead of rebooted

I'm not finding much about the error, "ended on signal UNSUPPORTED(1)". Has anyone seen this before?

Andy

 

[LnxBil]

"Read-only file system" implies something wrong with the filesystem. Could you describe your system in detail and look for filesystem errors in dmesg?

 

[amace]

I've checked the hardware out and it is fine. The problem follows the container. it does the same thing after being moved to a different node that was running flawlessly.

After I posted, I found the container was set to privileged so I've backed up and restored as unprivileged. I'm hopeful that will fix it. There's nothing else that's different about the installation - other machines are fine and I can't find anyone with the same problem on google.

 

[amace]

Just to follow up, setting unprivileged fixed the problem. Not sure how they got set that way to begin with but that's what it was.