@jvanabra "merge synchronously" option is there for log purpose. When you click on your backup job you can get all logs from this job here: [image: 1777971428328-78941893-5585-486a-b9bc-7d4c59873aa6-image.jpeg]

Xen Project covered this as XSA-489.

@olivierlambert In my understanding, Anti affinity tags only looks for the tags that are attached to the VM. For example: Datacenter 1 Host 1 VM1-1 Host 2 VM1-2 Datacenter 2 Host 3 VM2-1 Host 4 VM2-2 VM1-1 and VM1-2 should be able to migrate to Host 1 and Host 2. VM2-1 and VM2-2 should be able to migrate to Host 3 and Host 4. But VM1-1 and VM1-2 should never be able to migrate to Host 3 and Host 4. VM2-1 and VM2-2 should never be able to migrate to Host 1 and Host 2. With the current (Anti) Affinity tags, it will only look to the tags attached to the VM, so in the example it would be possible for VM1-1 to go to Host 4, if VM2-2 would be at Host 3 together with VM2-1. But if we could tag the hosts, it would not be possible for VM1-1 and VM1-2 to be on Host 3 and Host 4

@Aleksander There are plenty of examples on this forum. We have used Nvidia T4, it works very well.

Hi @olivierlambert and @pilow Thank you for your answers, it helps a lot, Regards, Olivier

@john.c This is a good idea as well. If the previous suggestion of trying to do a DNS round robin approach doesn't work, this may be my best option. Thanks for taking the time to provide your suggestion!

@gregoire said: @olivierlambert I added this feature request in the backlog regarding RPU improvements. Hello, Thanks all ! Totally agree with @poddingue , be able to exclude VM which has: PCI attached devices Local storage ( maybe ? ) Could be a great option !

@TechGrips Sorry, there is no quick test to be sure the VM is not corrupted. The usual way would be to make a healthcheck. We cannot be sure everything is ok as it concerns multiple tar linked to each other. If it keeps warning you on the same backups, it may be due to a faulty parent. You would need for this to create a new chain of snapshot

I solved the problem by installing and setting up a new Ubuntu Server. I exported the XO configuration from the old server and imported it onto the new one. What can I say—the backup started working again right away.

@julienXOvates Got it. I didn't get a chance to work on this over the weekend, so sometime this week, I'll just bounce the host and see if it comes back. Right now, while I can SSH into it, the admin console appears to be frozen and unresponsive (over iLO).

@poddingue I see backups in the tasks view fine, just not mirror ones. I'm talking about v5, haven't converted to v6 yet because of missing features.

Note that we have 0 XO or XCP-ng dev based in the US neither

yes we'll enhance this behavior in a coming release ! thanks

Ah very good, so it was even easier than this. You had the Xen blk driver but instead of using an UUID, the appliance was having a hardcoded sda. Keep us posted

@delacosta456 said: the pool host maximum limit is 3200% right right

@yannsionneau Thank you very much. Please let me know if there’s anything I can help with.

Hello @MathieuRA , @florent Moving the discussion to this thread as it seems to be tackling the same issue: https://xcp-ng.org/forum/topic/11892/xoa-memory-usage/28?_=1776334896974 I reported my findings so far over there Best regards

Thank you so much for your feedback, @Vagrantin !

@yannsionneau Thanks for the update! We are eagerly awaiting your findings! Thanks for looking into it.

@Bastien-Nollet I'm running c1e5f btw