Travis Nielsen

@mikesmitty Since the mgr pod is getting cycled with each reconcile, could you capture the pod spec and diff to see what is changing? The pod specs are expected to...

Ok, finally tracked down this is related to the exporter which is a new component in Reef v18. #13597 should fix this in Rook v1.13.3. The key log entry indicating...

> Logs are the same I sent before. Just to confirm, you still see this in your operator log? If so, there must be some other location I missed for...

@Nathanael-Mtd I can't see how that code path can be hit in v1.13.3. Just to humor me, does the first line of the rook operator log show that version? If...

> But the issue is not about the rook-ceph-exporter deploy objects which are deleted, but rook-ceph-exporter Service object. Aha, that's what I was missing! How did you recognize that it...

@briend Thanks for the details in that issue, it does look like it explains why the service keeps getting deleted. With a single cluster testing I couldn't repro this, but...

You can change the threshold for the alert to allow for a higher threshold, something like: `ceph config set mon mon_data_avail_warn 0.2` (less than 20% free) as described in [this...

> I would suggest the best solution to this - is to give a larger than 5 min (as given in alert rule in rook repo) window - for ceph...

@parth-gr Looks like you make a change in #13307 that would be resolved by `make crds`, but not clear if `make codegen` would be affected? Or what change did you...

The separate `repository` and `tag` settings are already available in the operator chart [values](https://github.com/rook/rook/blob/release-1.13/deploy/charts/rook-ceph/values.yaml#L5-L10). Are you referring to the [cephVersion.image](https://github.com/rook/rook/blob/release-1.13/deploy/charts/rook-ceph-cluster/values.yaml#L98) in the cluster chart?