Hi,
I just upgraded one of our clusters in place from 4.4 to 5.1. It's a 3 host cluster with ceph as storage backend.
So far all went smooth.
The problem I stumbled upon was that some VMs (all linux kvm, all had vga set to cirrus) didn't live migrate.
This cluster has 18VMs. 13 of these are in a HA-group, 5 aren't. The live migration from a 4.4 Host to a 5.1 Host worked without problem for all 13 Hosts of the HA-group, but none of the 5 hosts that aren't members there worked.
The error message didn't give a reason why the live migration faulted.
Shutting down the machines and migrate them offline worked.
We are planning to migrate two additional 5-Host clusters with far more machines where no HA-groups are defined at all. So it would be ugly if we couldn't migrate them live.
Does anyone has an idea what could be the problem?
I noticed a similar problem in this post, but as our is more specific and that problem is claimed solved I made a new thread.
https://forum.proxmox.com/threads/p...grade-vm-migration-problem.38117/#post-188106
Best Regards
Dirk
I just upgraded one of our clusters in place from 4.4 to 5.1. It's a 3 host cluster with ceph as storage backend.
So far all went smooth.
The problem I stumbled upon was that some VMs (all linux kvm, all had vga set to cirrus) didn't live migrate.
This cluster has 18VMs. 13 of these are in a HA-group, 5 aren't. The live migration from a 4.4 Host to a 5.1 Host worked without problem for all 13 Hosts of the HA-group, but none of the 5 hosts that aren't members there worked.
The error message didn't give a reason why the live migration faulted.
Code:
Apr 12 10:33:33 starting migration of VM 111 to node 'proxdmz01' (10.10.254.50)
Apr 12 10:33:33 copying disk images
Apr 12 10:33:33 starting VM 111 on remote node 'proxdmz01'
Apr 12 10:33:37 start remote tunnel
Apr 12 10:33:38 starting online/live migration on unix:/run/qemu-server/111.migrate
Apr 12 10:33:38 migrate_set_speed: 8589934592
Apr 12 10:33:38 migrate_set_downtime: 0.1
Apr 12 10:33:38 set migration_caps
Apr 12 10:33:38 set cachesize: 214748364
Apr 12 10:33:38 start migrate command to unix:/run/qemu-server/111.migrate
Apr 12 10:33:40 migration status error: failed
Apr 12 10:33:40 ERROR: online migrate failure - aborting
Apr 12 10:33:40 aborting phase 2 - cleanup resources
Apr 12 10:33:40 migrate_cancel
Apr 12 10:33:43 ERROR: migration finished with problems (duration 00:00:11)
TASK ERROR: migration problemsShutting down the machines and migrate them offline worked.
We are planning to migrate two additional 5-Host clusters with far more machines where no HA-groups are defined at all. So it would be ugly if we couldn't migrate them live.
Does anyone has an idea what could be the problem?
I noticed a similar problem in this post, but as our is more specific and that problem is claimed solved I made a new thread.
https://forum.proxmox.com/threads/p...grade-vm-migration-problem.38117/#post-188106
Best Regards
Dirk
