sync group vm/107 failed - owner check failed

[larsen]

Hello,

I have tested a sync from our local server to our external server (both running PBS 1.0-8) using "user1@pbs" (the sync destination). This worked as expected.
As I now started to use this more regularly, I have "beautified" user names to "user-vm03@pbs", but the sync started to fail afterwards:

ProxmoxBackup Server 1.0-8
()
2021-03-07T23:30:00+01:00: Starting datastore sync job 'our.hostname:vm03:sync_datastore:s-2db4540b-6fb2'
2021-03-07T23:30:00+01:00: task triggered by schedule '23:30'
2021-03-07T23:30:00+01:00: Sync datastore 'sync_datastore' from 'our.hostname/vm03'
2021-03-07T23:30:00+01:00: found 1 groups to sync
2021-03-07T23:30:00+01:00: sync group vm/107 failed - owner check failed (user-vm03@pbs != user1@pbs)
2021-03-07T23:30:00+01:00: TASK ERROR: sync failed with some errors.

It works again when I create the user "user1" on the sync server. Doesn't even have to assign any permissions to that user.

I only find this old stuff on the sync server:

titan:~# grep -R user1 /etc/proxmox-backup
/etc/proxmox-backup/shadow.json:  "user1": "$5$46F2gwvrnVo$CHvuaP9pZgkL2mGVkwbRDVuTb3f9pPVJcGs/c2emoq."

How can I fix this?

 

[dcsapak]

when editing/creating a sync job, you can choose a 'local owner' user. this user has (ofc) to exists and will be used to store the local backups. if the already existing backup groups belong to a different user you get this error. you can change the owner of the existing groups via the gui on the content tab of the datastore

 

[larsen]

I had seen that the old backups were connected to the old user and have deleted them. There are no entries in the content tab anymore.

I looked around a little bit more and noticed that the name is stored in "sync_datastore/vm/107/owner". Changed it there and the sync is starting again =)

Should it be considered a bug?
Also, will there be notification mails for failed sync jobs just like there already is for backup jobs?

 

[dcsapak]

Should it be considered a bug?

it is definitely not the best behaviour we probably either should remove the owner if we remove the last snapshot, or let it be changed even without any snapshots... not sure what is better though

Also, will there be notification mails for failed sync jobs just like there already is for backup jobs?

yes you can configure that on the datastore -> options tab

 

[larsen]

Thanks!

Reported as https://bugzilla.proxmox.com/show_bug.cgi?id=3336