-
Notifications
You must be signed in to change notification settings - Fork 10
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Feature
: host groups - honour HA groups
#65
Comments
Hey @antitbone, thanks for your input. When I initially integrated the affinity / anti-affinity rules I also thought about something like that. However, it's not really possible to do it in a user friendly way. While we have tags for VMs, we only have notes fields for nodes. This means, this could be abused in a dirty way, or something like an additional matrix in YAML or JSON could be provided which might look like:
or
So, from a technical perspective this isn't difficult but I'm not sure which way could be the best one for a user to provide the information / matrix. Happy to hear feedback! Cheers, |
It does not seem possible to have tags at the host level in proxmox.
But the assignment of vm to a group could be done by a tag. Antoine |
Hey Antonie, right, but this is already included in Proxmox: https://github.com/gyptazy/ProxLB?tab=readme-ov-file#affinity-stay-together However, this only keeps the VMs together but does not have any correlation to the nodes on this level. So they might still be moved to completely different "rooms" or "places. Cheers, |
Do the currently possible tags respect ha groups? Antoine |
No, they don't do it yet. |
This would solve the problem of placing VMs on a stretched cluster. |
+1 |
Hello
Is it planned to implement cluster management on multiple computer rooms?
It would be interesting to be able to define host groups and limit the migration of VMs to host groups.
This will allow the distribution of vms over several computer rooms so that the hosted services can withstand the loss of a room.
And have a degraded mode in which host groups are not respected if one of the host groups is not available.
Example with a ceph stretch cluster on Two computer rooms + a witness. Only the compute is to be migrated
If the cluster is correctly sized it is possible to survive the loss of a computer room.
Antoine
The text was updated successfully, but these errors were encountered: