ZFS Snapshot hangs

[greg]

Hello Community

I have a cron job like this:

for ct in /etc/pve/local/lxc/*.conf; do

        ct=$(basename "$ct")

        ctID="${ct%.*}"

        echo "Snapshot $snapname for $ctID"

        /usr/sbin/pct snapshot $ctID $snapname

done

which has been running for years.

A few days ago I upgraded my cluster from 5 to 6, and on one node, one the snapshot process hangs, making the CT unresponsive:

root     13190  0.0  0.0   6644  2660 ?        Ss   03:00   0:00      \_ /bin/bash /root/LOCAL/backup/daily_snapshots.sh
root     14424  0.0  0.3 323100 104728 ?       S    03:00   0:05          \_ /usr/bin/perl -T /usr/sbin/pct snapshot 114 daily_20200812                               
root     14426  0.0  0.2 332028 96948 ?        Ss   03:00   0:00              \_ task UPID:sysv4:0000385A:076C7E9F:5F333F19:vzsnapshot:114:root@pam:                  
root     14432  0.0  0.2 332028 91976 ?        S    03:00   0:00                  \_ task UPID:sysv4:0000385A:076C7E9F:5F333F19:vzsnapshot:114:root@pam:

I have to `kill -9` the last process.

A replicate job which starts at another time seems to be working fine:

(from the target node: )

ct_B/ct/subvol-114-disk-1@__replicate_114-0_1597269600__  creation              jeu. août 13  9:04 2020  -

The syslog only contains numerous lines like this, but I'm not sure it's related:

Aug 13 00:00:09 sysv4 lxcfs[6963]: utils.c: 331: read_file_fuse: Write to cache was truncated

All nodes are up-to-date.

Any idea?

Thanks in advance

Regards

 

[greg]

The GUI show: Error: failed to sync container namespace

 

[fiona]

Hi,

The GUI show: Error: failed to sync container namespace

is this the only line in the task log (you can double click the task to get the full log)? Were either of the host or the container rebooted since you installed the upgrade?