Finetune HA

[komupisi]

Hello,

I've configured test cluster with 3 servers, in cluster, with Ceph. Only 1 KVM virtual machine, in 'alwayson' HA group (request state started).

In my example I shutdown cluster1 server, where this test VM resides and measure time, until proxmox detects cluster1 server is dead and autostarts this VM on another server (cluster2 or cluster3). Right now this process is very long, up to 4 minutes.

Can someone please point me, where I can check and finetune timeouts for such HA operation?

Thank you

gay sex tube for all tastes and ages https://tubegay.name/en/

 

[badji]

By default the base time is 30 seconds.
4 minutes is weird, I think this slowness is due to a problem with your Ceph cluster network and the cephFS share directory.