I have a 13 node cluster using HA. What are the best practices for setting the max_restart and max_relocate values? As it stands right now, for VMs that can run on any node, I've simply picked a restart value of 4, and a max_relocate of 10. My thinking is that the HA service will try to restart a VM 4 times on a node, then move to another node, up to 10 times.
It seems the max_relocate value in the GUI is 10. Is my thinking of setting this value to the max of either the GUI set value or the number of nodes (minus 1) valid? For instance, if I only have 7 nodes, then I should set that to 6? Will the HA algorithm pick untried nodes when relocating, or is it random which node it will try next?
Does the max_restart number apply on a per-node basis? Meaning once the VM moves to another node, it tries to restart 4 times before moving to the next node where it will also try 4 times.
This information would be good to add to the Wiki page at https://pve.proxmox.com/wiki/High_Availability
It seems the max_relocate value in the GUI is 10. Is my thinking of setting this value to the max of either the GUI set value or the number of nodes (minus 1) valid? For instance, if I only have 7 nodes, then I should set that to 6? Will the HA algorithm pick untried nodes when relocating, or is it random which node it will try next?
Does the max_restart number apply on a per-node basis? Meaning once the VM moves to another node, it tries to restart 4 times before moving to the next node where it will also try 4 times.
This information would be good to add to the Wiki page at https://pve.proxmox.com/wiki/High_Availability