backup speed for LXC with ZFS backend

[yarii]

Well - doing incremental backup I see that comparing backups VM(1TB) and LXC(800GB) speed there is 10x performance gap (or even more! - doing backup for LXC is significantly slower even if it's smaller)
I did some test zfs send/zfs receive to remote host not differs so much.

Local storage backend: zfs
Remote storage: Proxmox Backup Suite (newest v3).

Is it possible to make this process faster for LXC?

 

[Dunuin]

Is it possible to make this process faster for LXC?

No. Only VMs could make use of dirty-bitmapping. Means when using a VM, QEMU will keep track of what parts of the virtual disk got written to or overwritten since the last backup and will completely skip those unchanged parts so those won't need to be read and hashed again.
When using LXCs all the data has to be read and hashed each time.
So if you got a LXC and VM, each 1000 GB and both of them only changed by 1 GB since the last backup, then the VM will only have to read and backup 1 GB but the LXC has to read 1000 GB and backup 1GB.

 

[yarii]

I know that difference but in backup process incremental snapshot using zfs send / zfs receive time is also the same.
Maybe PBS/PBS Client dev should use other magic spells for backuping LXC?

 

[Dunuin]

PBS won't make use of any ZFS features like snapshots or zfs replication. PBS should work with any filesystem, no matter what storage is used for best compatibility.