Incremental from Ceph cluster to PBS not working

[solv]

Hi,
Just testing PBS - and VM's on my ZFS backed pool backing up to my ZFS backed PBS datastore seem to work fine when it comes to incrementals/dirty chunk only backups, but the ones I've tested from my Ceph/RBD backed pool in my cluster going to the same PBS datastore do a full backup each time they are triggered.
Is this just a limitation of RBD or just a feature that needs to be worked out as PBS develops?

PBS is fantastic btw, thank you so much for developing it!

 

[tom]

No, this should/must work. Make sure that the VM is always running between the backups.

Post your backup log.

 

[solv]

Thanks, I'll do a bit more testing. I did shut it down between backups - only because after the first successful backup, the second time it spat out the error to do with qemu not supporting this type of backup (which is resolved by a reboot on machines that were running when proxmox was upgraded to 6.2-10) which I did think was weird.

I've actually just nuked the cluster and put more RAM and an SSD into each of the lab machines, so I'll get back to you once I can reproduce it.
Thanks for the quick response

 

[solv]

Hi,
AFter rebuilding the cluster with more RAM and SSD's for proxmox install I tested this again and it is now performing as expected with incremental backups. Thanks!