I’m trying to figure out how to start VMs after controller host failure.
I have a number of OpenNebula nodes in HA cluster: controller plus KVM host. In case of passive node failure VMs are starting well but if controller node fails (unexpected reboot, for example) - VMs were running on this host don’t start\reschedule. I was trying to set VM_HOOK like
VM_HOOK = [
name = “hook_vm_on_unknown”,
on = “CUSTOM”,
state = “ACTIVE”,
lcm_state = “UNKNOWN”,
command = “ft/hook_vm_on_unknown.sh”,
arguments = “$ID $PREV_STATE $PREV_LCM_STATE”,
remote=“no” ]
but don’t know which statuses should I set for the hook.
In logs I can see
Sun Jul 17 10:14:22 2016 [Z0][LCM][I]: VM running but monitor state is POWEROFF
Sun Jul 17 10:14:22 2016 [Z0][VM][I]: New LCM state is SHUTDOWN_POWEROFF
Sun Jul 17 10:14:22 2016 [Z0][VM][I]: New state is POWEROFF
Sun Jul 17 10:14:22 2016 [Z0][VM][I]: New LCM state is LCM_INIT
Is there any way to start only VMs which were running before fail? As well will VMs start if controller will start on another node?
Thanks and best regards,
Vladimir