Can't create VM on cluster hosts other than primary node.

sjalloq

New Member
Jan 2, 2025
4
0
1
Hi,

I have a 3-node cluster running Ceph and I followed the Simple Routing mesh configuration. The cluster is up and running and I've been creating VMs on the primary host. I've just tried creating a new VM on the other 2 hosts in the cluster and the 'Create' step seems to hang.

I've just restarted node 2 and tried again and this time it seems to be fine. I still have tasks hanging on node 3 so can put some debug info here if needed. Any thoughts on what would cause this? Restarting seems heavy handed.
 
Here's the journalctl from today that shows some task timeouts. Also, here is the Ceph status:

Code:
root@pve-prod-03:/var/log# pveceph status
  cluster:
    id:     09bf080b-be37-4e58-b14f-e1a8ecce63f2
    health: HEALTH_OK

  services:
    mon: 3 daemons, quorum pve-prod-01,pve-prod-02,pve-prod-03 (age 23m)
    mgr: pve-prod-01(active, since 12d)
    mds: 1/1 daemons up
    osd: 6 osds: 6 up (since 23m), 6 in (since 12d)

  data:
    volumes: 1/1 healthy
    pools:   4 pools, 97 pgs
    objects: 4.30k objects, 16 GiB
    usage:   47 GiB used, 35 TiB / 35 TiB avail
    pgs:     97 active+clean
 
Without context what you did on the other nodes while this config was written, it's hard to guess. Hopefully you rebooted a node, then the data would be consistent.

However, the qemu hangs are strange.
 
This was all a fresh setup/installation so perhaps the nodes have never been rebooted. Dunno, I did it just before the holiday break and can't remember now. I'll reboot and hope this doesn't continue.