Should I move VMs to new install instead of upgrade?

dtiKev

Well-Known Member
Apr 7, 2018
78
4
48
Hi!

I'm running a 2 node 5.4.15 (updated yesterday) cluster that houses a handful of legacy machines that I converted from physical to vms a few years ago. After reading some of the issues with upgrades I'm wondering if moving all Vms to one of the nodes, upgrading the other node to the latest with a fresh install, and then somehow putting the machines on that server and then wiping the other and setting it up and building a new cluster might be a safer route.

While none of my VMs are mission critical I'd rather not chance unknown errors and downtime if I can help it. Has anyone used this method instead? I'm looking for insight as to the best method of getting VMs from 5.x to a different 6.x installation.

Thanks,
dtiK
 
Well I started the upgrade last night... I moved all VMs to one of my quorate and upgraded the other. Once I rebooted I see the new version of host1 and try to cold migrate machines from the older versioned host2 to the upgraded one and get:

Code:
2020-08-06 08:26:05 starting migration of VM 100 to node 'pvn1' (10.11.12.251)
2020-08-06 08:26:05 found local disk 'local-zfs:vm-100-disk-1' (in current VM config)
2020-08-06 08:26:05 found local disk 'local:iso/virtio-win-0.1.171.iso' (in current VM config)
2020-08-06 08:26:05 can't migrate local disk 'local:iso/virtio-win-0.1.171.iso': local cdrom image
2020-08-06 08:26:05 ERROR: Failed to sync data - can't migrate VM - check log
2020-08-06 08:26:05 aborting phase 1 - cleanup resources
2020-08-06 08:26:05 ERROR: migration aborted (duration 00:00:00): Failed to sync data - can't migrate VM - check log
TASK ERROR: migration aborted
 
Last edited:
When I initiate a Migration at the GUI it get a pop-up that reads:

Method 'GET /nodes/pvn2/qemu/107/migrate' not implemented (501)

After I click OK, I am able to start the migration. Will that go away once the second host is upgraded or did I mess something up?