PVE 7 - Replication is not stable and logs are incompletes

[ikus060]

Hello,

With PVE 7, I have a LXC server with replication target with another PVE 7 node and the replication is not working well.

When schedule manually, the process is starting for ~3-4 min and stop working in the middle without anymore information regarding and error or failure.

Looking at the running process I see the replication not running (no more zfs send or received process on both node) and confirm the ZFS dataset is not completed imported either.

In the logs, I see incomplets logs. e.g.:

2021-12-05 20:38:57 120-2: 20:38:57   4.30G   rpool/data/subvol-120-disk-2@zfs-auto-snap_monthly-2021-01-01-1152
2021-12-05 20:38:58 120-2: 20:38:58   4.39G   rpool/data/subvol-120-disk-2@zfs-auto-snap_monthly-2021-01-01-1152
2021-12-05 20:38:59 120-2: 20:38:59   4.50G   rpool/data/subvol-120-disk-2@zfs-auto-snap_monthly-2021-01-01-1152
2021-12-05 20:39:00 120-2: 20:39:00   4.59G   rpool/data/subvol-120-disk-2@zfs-auto-snap_monthly-2021-01-01-1152
2021-12-05 20:39:01 120-2: 20:39:01   4.68G   rpool/data/subvol-120-disk-2@zfs-auto-snap_monthly-2021-01-01-1152
2021-12-05 20:39:02 120-2: 20:39:02   4.77G   rpool/data/

Like the process was terminated right in the middle.

I've try running the command line to send and receive the ZFS dataset manually and it's working without any error.

pvesm export data:subvol-120-disk-2 zfs - -with-snapshots 1 -snapshot __replicate_120-2_1638754681__ | /usr/bin/ssh -e none -o 'BatchMode=yes' -o 'HostKeyAlias=boxter' root@192.168.14.30 -- pvesm import data:subvol-120-disk-2 zfs - -with-snapshots 1 -snapshot __replicate_120-2_1638754681__ -allow-rename 0

It's probably a bug or something. Any body experiences something similar ?

 

[fiona]

Hi,
please see this thread for more information. To fix the issue it should be enough to

1. Disable all replication jobs and wait for currently running ones to finish (there might be some running in the background even if the UI doesn't display it, check with ps aux | grep pvesm).
2. Upgrade to pve-manager >= 7.1-6.
3. Re-enable the replication jobs.

 

[BeWu]

Hello,

Just want to confirm that had the same issue, and upgrade 7.1-5 -> 7.1-7 solved the problem .

All the best Everyone!
Bartek

 

[ikus060]

I've also restarted pvescheduler service after upgrading.