Thanks for the help.
So the problem was an issue with the crushmap. Ceph03 should not have been
listed in ‘default’ too – especially with the higher weight. It was basically
listed twice. When the system was originally installed, ceph03 blew up, so they
replaced it with a system using the same
Hi Goncalo! Thanks for the explanation of the pg identifier. That helped a lot.
I thought XX in pg XX.YYY was an OSD, not a pool.
So pool 19 was the pool I created yesterday in an attempt to replace the broken
'volumes' pool (6}. I just deleted the new pool and all the 19.XXX unclean
messages