How to reduce timeout when migrating a VM in a Proxmox cluster when a node failure?
Trying to prepare for the move of some services a cluster of 3 node-based Proxmox and GlusterFS cluster for testing new path enabled HA on the proks, migration occurs. But the problem lies elsewhere, if it loses connection to Noda quite a long time to wait for the cluster will take the nod for sure outside and start dev on the next one. It takes 2 minutes before it begins the start of the path, plus 30 seconds to run, for a total of 2 minutes, 30 seconds of downtime, which is quite a lot. Can this timeout to reduce?
Did some digging on the forum Proxmox, at the moment apparently, these time-outs are required to operate their fence algorithm, the talk was about the emergence of certain options in future versions that will help it to diminish. Well, actually it is myself, but Oh well.
dave_Kassulke answered on April 4th 20 at 13:01
how do you reduce , there in fact need time to fencing and if you do it instantly then you can grab two of the same running path