Max Asnaashari

@sabaini Thanks for merging the other one, I've now rebased this one.

The hooks are currently included here, when the daemon starts: https://github.com/canonical/microceph/blob/aae9f94dd349132283b08902b23aee27795e60fe/microceph/cmd/microcephd/main.go#L64-L84

Two ways I can think to do this: One is add your own endpoint for node removal (like `DeleteClusterMember`), which updates the `PreRemove` hook with the appropriate function based on...

@UtkarshBhatthere @sabaini There is now a `force` flag in upstream `microcluster` in both the `PreRemove` and `PostRemove` hooks, which use the same value passed to `DeleteClusterMember`. Hopefully that will unblock...

I was referring to MicroCeph asking the user to specify a name for the cluster member, but this is not respected everywhere. A grep for `os.Hostname` shows that the supplied...

> @masnax how this sort of thing achieved in lxd? LXD does it exactly the way it's currently done in microcluster. But since we introduced the API extension component after...

> > Wouldn't they always implicitly happen on the leader, as only the leader POSTs to the database? > > I dont follow you here, the leader surely has direct...

> That's correct. That's why I dont understand how the deadlock you're describing is occurring and why there are notifications going on. Imagine 3 nodes have schema version 1 and...

> This 30s timeout scenario will trigger every single time we add a new schema update and a new API extension at the same time. The reason for that is...

> > The reason for that is because if we detect that we have a new schema update, we can't be sure if this update is the one that introduces...