Hi,
With reference to my post in the thread If the shutdown and startup are timed correctly, a zombie listen port will be created:
My container has been running fairly stable for about two weeks but has started mis-behaving.
- First noted since the load average is high; currently about 16 when normally about 2. No obvious culprits in
top
but this is the same MO as I have seen on last two occasions of this issue. - The culprit (or at least the most obviously impacted) container is running Plex Media Server
- The Plex web frontend seems to be working for browsing the libraries etc but will not play media and has stopped pulling DVR guide data, etc.
- Attempted a shutdown of the service within the container which failed.
systemctl status plexmediaserver
shows the service is now in a failed state. - Attempt to stop the container and
lxc stop plex
never returns. After five minutes, +c out. -
lxc info
below. Note the network interfaces have gone. Log messages are quite benign I think.
Name: plex
Location: none
Remote: unix://
Architecture: x86_64
Created: 2021/06/06 12:50 UTC
Status: Running
Type: container
Profiles: default
Pid: 22671
Resources:
Processes: 6
Disk usage:
root: 16.76GB
CPU usage:
CPU usage (in seconds): 159936
Memory usage:
Memory (current): 230.21MB
Memory (peak): 12.46GB
Snapshots:
snap0 (taken at 2021/06/15 11:41 UTC) (stateless)
Log:
lxc plex 20210731102443.919 WARN conf - conf.c:lxc_map_ids:3389 - newuidmap binary is missing
lxc plex 20210731102443.920 WARN conf - conf.c:lxc_map_ids:3395 - newgidmap binary is missing
lxc plex 20210731103736.799 WARN conf - conf.c:lxc_map_ids:3389 - newuidmap binary is missing
lxc plex 20210731103736.799 WARN conf - conf.c:lxc_map_ids:3395 - newgidmap binary is missing
- Attempt to stop the container with ‘-f’ flag, still seems to run forever. No change to output of
lxc info
. - The host is running Ubuntu 20.04.2.
- Running LXD snap.
tracking: 4.0/stable/ubuntu-20.04
andinstalled: 4.0.7
Happy to leave this container in a dodgy state for next day or so but will need to recover it for next week. Only way I have to recover is to restart the host.
Regards,
Whytey