Hi,
If I write too much data to a Ceph cluster such that an OSD becomes full, it seems that I am unable to delete any objects. I've read about a few options to reduce usage and get out of this situation, namely:
(A) add OSDs, or
(B) temporarily disable backfill then increase the full ratio, or
(C) temporarily disable backfill then delete some PGs, being careful to not delete all the copies of any data.
Thanks!
DavidIf I write too much data to a Ceph cluster such that an OSD becomes full, it seems that I am unable to delete any objects. I've read about a few options to reduce usage and get out of this situation, namely:
(A) add OSDs, or
(B) temporarily disable backfill then increase the full ratio, or
(C) temporarily disable backfill then delete some PGs, being careful to not delete all the copies of any data.
All of these
options require a fair amount of administrator access to the Ceph
cluster. Is there a way for a user with more restricted access (e.g. to a
single pool) to reduce usage on the OSD? For example, it'd be great if
there's a setting that allows the deletion of objects even when an OSD
is full.
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
