It's getting even worse now

LINBIT SDS Integrations Proxmox VE
1

Hi there,

as you can see in the topic “Migrating Proxmox-VM(s) fails with 'Wrong medium type” I’m kind of struggling with a small Proxmox/DRBD cluster.

Yesterday I’ve been running updates on all nodes just to make sure, that everything is up to date, and maybe the mentioned problem may solve itself. :wink:

Before the update I wanted to hibernate all the VMs on node1 where they have been running on … but that didn’t work!

Then I shut all the VMs down, ran updates and rebooted.
Afterwards I had to “drbdadm primary vm-1xx-disk-1” on all the disks that are in use (DRBD-distributed to two other nodes).

And now … the disks reside in ‘datapool’ and have a 00000 added to their name, plus there’s a …state_suspende$DATE_00000 of all the disks there.
The disks are defined to be in ‘drbdstorage’ where their are also visible, but node1 uses them from ‘datapool’ which it treats as a local storage.

So migrating to the other nodes now works … but it syncs the disks to the other node instead of relying on the underlying DRBD device (in ‘drbdstorage’).

What is that?
And how can I get rid of it?

Seems like I’m a bit lost with all those pools, storages, groups and so on.
Setting it up worked like a charm, but now the vm-disks seem to have ‘moved’ somehow and started behaving like thei’re on local storage instead on DRBD storage anymore.

Can please anybody show me a way out of this … ?

Best regards
Matthias

ps

datapool:

linstor storage-pool create zfs nodeX LinstorData datapool
linstor resource make-available --diskful nodeX vm-10x-disk-1
etc.

/etc/pve/storage:

drbd: drbdstorage
content images, rootdir
controller 10.10.10.1,10.10.10.2,10.10.10.3
resourcegroup pve-rg
preferlocal yes

2

What does the health of LINSTOR and DRBD look like at this point?

linstor node list
linstor storage-pool list
linstor resource list

If things are “very bad” after forcing things to primary and migrating VMs to non-LINSTOR storage, and you want to get them back onto LINSTOR replicated storage, you might be best to cleanup (delete) LINSTOR volumes that were somehow migrated to local storage, and after the LINSTOR cluster looks healthy again, migrate them to LINSTOR replicated storage.

Hard to say how you got there in the first place without detailed steps and logs.

1 Like