HA managed VMs does not start, when node is cold powered down

[fomichevk]

Hello.

I have Proxmox 2 RC1 installed on 5 nodes in cluster.

pve-manager: 2.0-35 (pve-manager/2.0/d07f49c3)
running kernel: 2.6.32-7-pve
proxmox-ve-2.6.32: 2.0-60
pve-kernel-2.6.32-7-pve: 2.6.32-60
lvm2: 2.02.88-2pve1
clvm: 2.02.88-2pve1
corosync-pve: 1.4.1-1
openais-pve: 1.1.4-2
libqb: 0.10.1-2
redhat-cluster-pve: 3.1.8-3
resource-agents-pve: 3.9.2-3
fence-agents-pve: 3.1.7-1
pve-cluster: 1.0-23
qemu-server: 2.0-20
pve-firmware: 1.0-15
libpve-common-perl: 1.0-15
libpve-access-control: 1.0-16
libpve-storage-perl: 2.0-12
vncterm: 1.0-2
vzctl: 3.0.30-2pve1
vzprocps: 2.0.11-2
vzquota: 3.0.12-3
pve-qemu-kvm: 1.0-4
ksm-control-daemon: 1.1-1

Shared disk storage exports LVM VGs via SRP (Infiniband).
Every VG is available on every node.
Fencing is configured by fence_ipmilan, and working well.

For testing I've created some VMs and put them to HA cluster.
Rgmanager is running on all nodes. When I stop rgmanager, all VMs, running on node, starts on other nodes in cluster.
BUT when I power down node (e.g. pull of power cords) VMs, worked on this node, does not starting on other nodes, and I could not start this VMs even manually!

Is this normal?

 

[dietmar]

Is this normal?

Does fencing work if power is down (should work)?

 

[fomichevk]

Does fencing work if power is down (should work)?

Yes, fencing is working when node powered off (e.g. power button pressed), node is starting over IPMI. But, when node _physically_ loss the power (PSU fails, motherboard burns, etc.) fencing just can't work - IPMI is unavailable.

 

[tom]

in this case you need to use power fencing like APC power switches - using IPMI is cheap but does not work in all cases like this one.