Hi all I have a Ceph cluster consisting of two zonegroups with metadata syncing enabled. I need to change the owner of a bucket that is located in the secondary zonegroup. I followed the steps below: Unlinked the bucket from the old user on the secondary zonegroup: bash Copy code $ radosgw-admin bucket unlink --uid OLD_UID -b test-change-owner Linked the bucket to the new user on the secondary zonegroup: bash Copy code $ radosgw-admin bucket link --uid NEW_UID -b test-change-owner Changed the owner of the bucket on the primary (master) zonegroup: bash Copy code $ radosgw-admin bucket chown --uid NEW_UID -b test-change-owner After executing the last command on the primary zonegroup, the bucket owner was successfully changed. However, the ownership of the objects within the bucket still remains with the old user. When I executed the same radosgw-admin bucket chown command on the secondary zonegroup, I received a warning about inconsistent metadata between zones, but the bucket owner was changed successfully on the secondary zonegroup. My questions are: What is the best way to change the owner of a bucket in a multi-zonegroup cluster? What are the potential impacts of running the chown command on the secondary zonegroup? Is it possible to have inconsistent metadata between zones in this case? _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |