Hi,
I've been benchmarking my Luminous test cluster, the s3 user has deleted
all objects and buckets, and yet the RGW data pool is using 7TiB of data:
default.rgw.buckets.data 11 7.16TiB 3.27 212TiB 1975644
There are no buckets left (radosgw-admin bucket list returns []), and
the only user is using no quota.
radosgw-admin gc list doesn't show anything pending gc; if I do
--include-all there are some:
[
{
"tag": "01a3b9f4-d6e8-4ac6-a44f-3ebb53dcee1b.3099907.1022170\u0000",
"time": "2018-12-06 13:36:59.0.88218s",
"objs": [
{
"pool": "default.rgw.buckets.data",
"oid":
"01a3b9f4-d6e8-4ac6-a44f-3ebb53dcee1b.3665142.15__multipart_b713be7d5b86b2fa51830f7c13092223.2~7_mvHIZc-L8mOFy51hkGnZbn4ihgOXR.1",
"key": "",
"instance": ""
},
[continues for 16k lines]
What am I meant to do about this? OK, it's a test system so I could blow
the pool away and start again, but I'd like to know what the underlying
issue is and how I'd manage this on a production cluster.
We've previously had data-loss issues with using orphans find (i.e. it
found things that were not orphans)... :(
Regards,
Matthew
--
The Wellcome Sanger Institute is operated by Genome Research
Limited, a charity registered in England with number 1021457 and a
company registered in England with number 2742969, whose registered
office is 215 Euston Road, London, NW1 2BE.
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
