XCP-NG Pool Warm Migration not working

Computer Hardware & Server Infrastructure Builds
1

Hy @all,
since some weeks, I’m migrated from Proxmox to xcp-ng and am really impressed on how easy and versatile things are. I’m still struggling on some basic things, though.

I have 2 Minisforum MS01 setup almost identical with the 2 Ethernet-Port bonded for outward traffic. One SFP+ Port ist dedicated for Connection to my Synology NAS, the other one is dedicated for replication between the two hosts.

While I set up the second host the last few days, I stepped up to xcp-ng version 8.3 rc1. I added both hosts to a single pool and since then, I gives me the error, that I can’t access the second host

image
image1740×410 34.1 KB

with the error-message being empty.

I also can do a warm migration from one host to the other. Sometimes it’s working from Host2 to host1, but it never works the other way around. The error I got is:
Vm target of warm migration not found for 8ced24c5-8623-d2e5-2b7b-4208b3a21f9b on SR a5417970-bfca-2e46-a10a-384e1f9708fd

in the logs I also see severyl errors of this kind:
OPERATION_BLOCKED(OpaqueRef:faf19443-5923-b28e-611a-472fa309eb35, true)

2

I can´see anything in the Logs-section, and I’m fairly new to XOA/XCP-NG. So please tell me, what else you need to know to track down my errors.

One thing important to note is, that I did a fresh install of xcp-ng 8.3-rc1 on both machines.

3

I have not done any testing with warm migration with version 8.3 (I don’t really use it in 8.2 either) mostly because all of our clients have similar systems. All the logs are in /var/log/ and you can grep for the UUID of the VM and see what other error messages there are related to the failed migration.

4

Thanks, Tom for your help,

I didn’t think of a software problem. My initial thought was, that it is a layer 8 problem - since I’m fairly new to xcp-ng and xen.

I took the advice from your YouTube-video and migrated everything over to a NFS-Share on my Synology. I have a 10Gig connection on the Synology and both of my hosts, dedicated for Storage. Somehow, it’s only using the normal Ethernet 1Gig connection. I’ve configured this 10gig connection in the pool settings under Misc for “Backup Network”. Since the Servers have a second 10Gig Port, I used this one for “Default migration network” in the same section.
I also added the NFS-Pool under “Default SR”, “Suspend SR” and “Crash Dump SR”.

Somehow, it’s not using the 10Gig connection for Migration and communication to the NFS share. When I look at the stats of the hosts, I can see no traffic coming in/out of this network-port.

What am I missing here?

5

I would need to better understand your network to answer, but storage should be on the same subnet.