On Mon, Nov 21, 2016 at 3:33 PM, Graham Allan <gta@xxxxxxx> wrote: > > > On 11/21/2016 05:23 PM, Yehuda Sadeh-Weinraub wrote: >> >> >> Seems like bucket was sharded, but for some reason the bucket instance >> info does not specify that. I don't know why that would happen, but >> maybe running mixed versions could be the culprit. You can try >> modifying the bucket instance info for this bucket, change the num >> shards param to 32. >> >> Yehuda > > > Yes! That seems to do it... > > And of course in retrospect I can see num_shards = 32 for my working bucket > but 0 for this one. > > so I did: >> >> rados get --pool .rgw.buckets.ec42 >> default.712449.19__shadow_tcga_test/rnaseq/Aligned.out.sam.2~10nUehK5BnyXdhhiOqTL2JdpLfDCd0k.11_76 >> lemming > > > edit "num_shards": 32 > >> # radosgw-admin metadata put bucket.instance:tcga:default.712449.19 < >> tcga.json > > > and this bucket is now visible again! Thanks so much! > > I wonder how this happened. It looks like this affects ~25/680 buckets. > > Could be that this is a secondary issue, maybe trying to fix the buckets because of the original issue caused this to happen (radosgw-admin bucket check maybe broke it?). Yehuda > Graham > -- > Graham Allan > Minnesota Supercomputing Institute - gta@xxxxxxx _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |