Question about Continuous Replication/ Backups always doing Full Backups

florent

@Kraken89 is it on all VM or only one ?
we did quite a lot of changes on the replication, maybe we missed something.

are you replicating inside the same pool or toward another pool ? (are both server in the same pool ? )

Andrew

@florent I too am having issues with this... on some VMs...

Running XCP 8.3 (latest March update), XO (master 0286f).

I have a simple small CR job that replicates 3 VMs between two pools. Two VMs got stuck at Full, and one works correctly. CBT/NBD 1 enabled. Purge disabled.

It's also leaving snapshots each time. No CD in the drive this time.

I'll have to update XO and try again.

Kraken89

@florent Hello!

The VMs are replicating within the same pool. The pool only has 2 hosts and all is taking place within the same pool.

All the VMs currently live on host 1 with the local storage. The VMs are set to replicate to the local storage on host 2.

ph7

since I had my problem fixed in
https://xcp-ng.org/forum/topic/11969/timestamp-lost-in-continuous-replication/22
I tested my f5468 commit similar to Your setup with 2 Hosts in the same pool and local storage, ran CR with retention of 2 from 1 host to the other and it worked as expected.
One full and then delta.

Andrew

@florent I updated XO to master 60ba5, and still having the issue.

I don't see any errors in the backup log. It just says "isFull": true for the ones stuck on full backups.

The XO log just reports:

Mar 24 16:26:47 xo1 xo-server[21199]: 2026-03-24T20:26:47.315Z xo:backups:worker INFO starting backup
Mar 24 16:26:52 xo1 xo-server[21199]: 2026-03-24T20:26:52.359Z xo:xapi:xapi-disks INFO export through vhd
Mar 24 16:26:54 xo1 xo-server[21199]: 2026-03-24T20:26:54.410Z xo:xapi:xapi-disks INFO export through vhd
Mar 24 16:29:22 xo1 xo-server[21199]: 2026-03-24T20:29:22.250Z xo:backups:worker INFO backup has ended

Note, it reports two export through vhd and does two full backups and one delta.

Kraken89

@florent Hello,

So I am not sure what changed but it looks like the continuous replication should be working as intending now. I did reconfigure the schedule a last time before I posted last and let it run. I can see where it was still doing the full backups and where I stopped it by restarting the XOA. It seems then after the reboot and saving the schedule again it appears to working.

It is now showing delta backups and the transfers are much smaller.

I wish I had a better idea what would have caused it to work but I will continue to monitor it. I would say for now my issue is resolved.

florent

@Kraken89 I will look into the transition from the previous system to the new one

maybe that's the key

tsukraw

I thought I was having a similar issue but mine might be slightly different.

This is the first time I am trying to configure CR so it's very possible I am not setting things correctly. I have two separate pools currently.

When I run the replication job, it shows type: Delta
However, the size is the full size the VM each time which is 200GB.

Since it is a delta shouldn't the size just be showing the change data?

The two pools are physically at different sites, so the throughput is not optimal, but i never expected a delta run to take 8 hours. the full took the same 8 hours.

Andrew

@tsukraw It's running a full, not a delta. Do you have Use NBD enabled?

tsukraw

@Andrew

But at the bottom it has Type: Delta
Which I took as running a delta?

Yes, I have NBD enabled. Should I try disabling it?
Here a side-by-side comparison.

I do see the message at the top saying "backup fell back to a full"
Does this imply that the delta function didn't work, and it is truly a full again?

tsukraw

@kraken89
Sorry, didn't mean to hijack your post

Andrew

@tsukraw The type delta at the bottom is a known bug... Do you have NBD Connection enabled on a network interface (Check the pool network)?