[SOLVED] Proxmox HA with starting LXC's and VMs with node wait_for_agent_lock during backup task

[mmeissner]

Hi,

I have Proxmox configured with HA (2 Nodes + 1 qdevice). Some Container and VM's are setup for HA. I had to shutdown primary node for some maintanance to upgrade RAM and run some test. HA migrated the LXC's and VM's that are configured for HA to the secondary node, all worked fine.
At the time I rebooted the maintained primary node the Cluster started to migrate the LXC's and VM's back to the primary node as expected.
However after migration and shutdown of the LXC's and VM's on the secondary node, they got stuck at starting on the primary.
When I checked the primary node I saw that the primary started a backup of a VM that takes quite a while (one that is not configured for HA) .
The node is currently backing up the VM and is stuck on "wait_for_agent_lock" and the LXC's and VM's are in "starting" state.
I assume as soon the Backup is finished the VM'x and LXC's are getting started. I can also see that the replication for the VM's and LXC's finished, the same as the Bulk start VM's and Containers job finished. However they are not running and HA shows them in starting state with "wait_for_agent_lock".
Is it suposed to be this way or can I do something to prevent this ?

Would be happy about some help here

 

[mmeissner]

Ok I found the problem. I had swapped month and day during setup in the BIOS. That means the system booted up with a time some month into the "future". Short time after the system was fully loaded and tasks were already fired off, the server got an update from the timeserver and reset the time to the actual value. This short period must have caused the hiccup and some other issue I've meanwhile fixed.