Can you add a delay for migrate-back once a rebooted node is back online?

[MagellanTX]

When rebooting a node, HA correctly migrates the running VMs to other nodes, but when the node comes back online it immediately tries to migrate all of them back - and that is failing for me.

My migration network is an internal (Thunderbolt) mesh network and takes time for FRR to re-admit a node once it’s online. Normally, this is less than 30 seconds, but during this time, HA is trying to migrate the nodes back, and it is failing because the network is still unreachable.

Is there a way to add a delay so once a node comes back online, it will wait (say 30 seconds) before it attempts to migrate the VMs back?

Thanks!!

 

[UdoB]

Possibly this does solve your problem: https://pve.proxmox.com/wiki/High_Availability#ha_manager_node_maintenance

 

[MagellanTX]

Possibly this does solve your problem: https://pve.proxmox.com/wiki/High_Availability#ha_manager_node_maintenance

Thank You!

It's certainly better than nothing and it is a procedure I can use.

The reason I was looking for a timed delay is because I've been doing some troubleshooting with Proxmox on an Intel big.little CPU which is requiring a lot of reboots. It gets a little tedious having to enable/disable maintenance mode each time, but I guess the maintenance mode is still better than having to manually migrate the VMs individually.

Thanks again!

 

[Packman]

Have you by any chance found a better solution? I'm in the same boat.

 

[MagellanTX]

Have you by any chance found a better solution? I'm in the same boat.

No, I did not find a permanent solution; I just use the maintenance mode as they have suggested. Unfortunately, this doesn't help when a node unexpectedly crashes because it can't automatically migrate the VMs back when it comes up.

 

[Packman]

Unfortunately, this doesn't help when a node unexpectedly crashes because it can't automatically migrate the VMs back when it comes up.

That is exactly my point. Thanks for answering.