.mgr will always have a single PG. If for some reason that PG wasn't placed on any OSDs, you'd have a bigger problem.
> On Jul 11, 2026, at 11:14 AM, listy via ceph-users <[email protected]> wrote: > > Yes, changing: > > -> $ ceph osd pool set .mgr crush_rule replicated_rule_MONERO > > fixes: > > -> $ ceph osd pool autoscale-status > > What was (not)happening there? _.mgr_ had no OSDs assigned, if so how did it > work, showing no errors (which I could spot)? You had a situation that was more complex than the PG autoscaler is equipped to deal with. Like I said, I have an RFE in to prevent this kind of thing. > > thanks, L. > _______________________________________________ > ceph-users mailing list -- [email protected] > To unsubscribe send an email to [email protected] _______________________________________________ ceph-users mailing list -- [email protected] To unsubscribe send an email to [email protected]
