Hi, After doing a 'radosgw-admin metadata sync init' on a secondary site in a 2-cluster multisite configuration, (See my previous post on doing a quincy upgade with sts roles manually synced between primary and secondary), and letting it sync all metadata from a master zone, I get the following as a large omap object warning on the secondary site: (from ceph health detail): "1 large objects found in pool <zone>.rgw.log" And in ceph.log: "Large omap object found. Object: 87.adcf1188:::meta.full-sync.index.0:head PG 87.abcdfdff (87.0) Key count: 1234567 Size (bytes) 0" I know about large omap objects, and because this is in the log pool, I should be able to delete it with the rados utility. (And do a deep scrub on the pg. Am I correct? Looks like old bucket metadata is still there even though the buckets have been deleted. -Chris _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |