Your bucket index got corrupted. I believe there is no easy way to restore the index other than downloading existing objects and re-upload them, correct me if anybody else know a better way.

You can check out all your objects in that bucket with:

rados -p .rgw.buckets ls | grep default.32785769.2

Plus what's your region map? where is the your bucket index stored? from the naming of data pool, you seems using same pool for bucket index?

Regards

Stanley


On 31/08/17 9:01 PM, Rudenko Aleksandr wrote:
Hi,

Maybe someone have thoughts?

---
Best regards,

Alexander Rudenko



On 30 Aug 2017, at 12:28, Rudenko Aleksandr <arude...@croc.ru <mailto:arude...@croc.ru>> wrote:

Hi,

I use ceph 0.94.10(hammer) with radosgw as S3-compatible object store.

I have few objects in some bucket with strange problem.

I use awscli as s3 client.

GET/HEAD objects work fine but list object doesn’t.
In objects list I don’t see these objects.

Object metadata:

radosgw-admin bi list --bucket={my-bucket} --object={my-object}

Return [].

But:

rados -p .rgw.buckets stat default.32785769.2_{my-object}

.rgw.buckets/default.32785769.2_{my-object} mtime 2017-08-15 18:07:29.000000, size 97430


Bucket versioning not enabled.
Bucket has more 13M objects.

Where can I find the problem?

---
Best regards,

Alexander Rudenko






_______________________________________________
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
_____________________________________________________________________________

This email has been filtered by SMX. For more info visit http://smxemail.com
_____________________________________________________________________________

_______________________________________________
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

Reply via email to