A lot of my VM's and CT's are for fun home hosting stuff, but a few of them are important. My cluster has 3 nodes and while it has enough resources if all the nodes are up, memory get a little tight if one of them goes down. There's enough, but it depends on how the VM's wind up being re-balanced in a failure and auto-restart scenario.
Example:
I don 't care if there isn't enough ram/cpu for plex, home assistant, torrent server, etc.
But I do care if some of these systems run out of resources: vm router, vm pbx, ct nextcloud.
How do I prioritize my important VMs & CTs for CPU and Memory? Or even say in failure scenario, just shut these non-essential vm's down, even if they are not on the node that failed.
EDIT: If i'm reading the documentation right i could set memory shares and cpu units? Default is 1,000 for memory shares, so i could just set it to any value higher than 1,000? likewise for the default on cpu units? But memory shares doesn't seem to be an option for CTs?
Example:
I don 't care if there isn't enough ram/cpu for plex, home assistant, torrent server, etc.
But I do care if some of these systems run out of resources: vm router, vm pbx, ct nextcloud.
How do I prioritize my important VMs & CTs for CPU and Memory? Or even say in failure scenario, just shut these non-essential vm's down, even if they are not on the node that failed.
EDIT: If i'm reading the documentation right i could set memory shares and cpu units? Default is 1,000 for memory shares, so i could just set it to any value higher than 1,000? likewise for the default on cpu units? But memory shares doesn't seem to be an option for CTs?
Last edited: