Hello,
I have a two nodes DRBD with reactor (and one arbiter) based on Ubuntu 24, i followed guide NFS High Availability on RHEL 9 with DRBD Reactor: A How-To Technical Guide - LINBIT
My clients are Proxmox Host with default settings.
When i used “evict” or “reboot” failover worked perfectly before going in production, VM just “pause” some IO but continue.
Yesterday i updated the primary node DRBD, I wasn’t paying attention, there was a netplan update, which apparently cut the network for a short while… So Reactor change “primary” but all the VM stuck with “read only” filesystem…
I know this problem from having had it in the past with other platforms or operating systems but I was surprised that it happened to me like that.
My DRBD HA NFS is now in production and I can’t run any more tests, but I’m planning to set up a second one. Are there any best practices? Specially with nfs hard / soft mount..
Thanks
Yathus