@olivierlambert I would think HP servers would be a top tier for IPMI support.... please...

@gduperrey Up and working on several Intel pools.

@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.

@acebmxer Yes, 22 or 24 should work correctly. Some early versions of 24 had issues, but not currently. I'm running 24 and it is working.

@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.

@acebmxer Use current Node LTS version... Currently 24.14.0
Or you can use 22.22.1.... use the latest version.

22.3 is the oldest version allowed to be supported (don't use one that old).

@florent I rebuilt my XCP hosting environment (everything is faster and bigger stuffed into one rack).... and this issue is now worse.

The main changes to this new setup are 2x40Gb networking, faster NFS NVMe NAS, faster pool servers, more memory, much faster CR destination machine with ZFS.

Running XCP 8.3 (March 2026 updates) and XO (master a2e33).

Replication is leaving many attached to control domain every day with NBD connection set to 2. Changing it to 1 seems to resolve the issue (no more stuck to control domain).

@yann Item Opened on Gitlab.

@olivierlambert Error causes tools to fail to load... >256 IP addresses on an interface (IPv4 and IPv6):

xen-guest-agent[514]: Error: Custom { kind: OutOfMemory, error: "no free slot for a new IP address" }

Best is to account for all the IP addresses.... second best is to ignore the extra IP addresses and not fail and exit.

@florent XO also can't migrate a VM to a new pool with a CD in the drive... It does generate an error, but the error is unclear.

@gduperrey The new OpenSSL/SSH blocks existing/working RSA keys from older SSH clients. While you can still use a password for SSH, it will block old keys from working which will break things (not good for existing LTS installs). To maintain compatibility add PubkeyAcceptedAlgorithms +ssh-rsa to /etc/ssh/sshd_config

@Pa3docypris I also had this problem with one single Debian VM. I would delete the snapshot and they would just keep adding every backup. I have lots of the same VMs and lots of other VMs on the pool. Three VMs were basically exactly the same but only one had an issue.

I tried deleting all snapshots on the VM. Changing the CBT state. Doing a full backup up of the VM.... same problem. Snapshots just built up.

I found a CD in the VM drive and it would not eject (with a XCP error). So there was something strange with the VM state. I shutdown the VM, cleared the CD, and restarted the VM. After that, backups worked normally for that VM. So I won't blame XCP or XO, it seems the VM was just in a random strange state that caused problems.

@dcskinner @acebmxer I totally agree.... XOA is the Vates provided software appliance XO system.

Compiling XO yourself is XO from source or XO community edition or anything other than XOA. Calling it XOA adds confusion because it's not the same thing that Vates provides.

@MajorP93 @olivierlambert I think 8.3 has reached LTS so I don't think the drivers are getting major upgrades (it's up to Vates). It seams reasonable to add the 8125 alt driver and new 8126/8127 drivers as an installable option.

Realtek keeps releasing new chips and new versions of the existing chips (like the 8125 in @paha machine) that don't work with the existing older drivers. It's an ongoing issue even with mainline Linux kernels.

@paha Try installing r8125-module-alt from my web page.

@whyberg Nope... that feature is part of the CPU. But is that the actual problem? Maybe... As that machine is too old to do any real work, it's a fine toy. There are better faster cheap test machines available.

@florent Delta backup is also leaving old snapshots on some VMs. It should only have one (current) snapshot for the nightly backup. This is an issue on 1 of 100 VMs.

XCP (Jan 2026 update) and XO (91c5d) are current.

@whyberg Normally not 10x...

@whyberg The E5430 does not support EPT/SLAT.... without EPT the system uses shadow paging so virtual machines will be slower. Also, you can't run XCP 8.3.

@florent Yes.