[PVE-User] Cluster Host Node Rebooted and VMs Did not restart? UPDATE
jmr.richardson at gmail.com
Fri Jul 5 19:39:08 CEST 2019
> Ran into an interesting problem today. Topology is a 4-Node Cluster on v.5.2, been up and running without issues for 10 months or so. Several VMs running on node 1, 2 & 3, maybe 30% loaded. I keep node 4 empty for HA and new build VM testing. All VMs are set for HA and auto start on node reboot.
> Node 1 rebooted, no indication why, nothing in the logs identifying cause of the node rebooting. When node 1 came back up it was 1 hour ahead of time and none of the VMs would start and HA did not start up the VMs on empty node 4.
I tested this scenario in the lab running PVE 5.4-7 and cannot
reproduce the original symptoms above. I manually set 1 node hardware
clock +1 hour, rebooted and when it came back up, the system time and
hardware clock synchronized to correct time within a minute or so and
VMs started up OK.
So my question, is there a startup routine in PVE 5.4-7 that
synchronizes system and hardware clock shortly after boot up that was
not present in PVE 5.2?
Engineering for the Masses
Chasing the Azeotrope
More information about the pve-user