On Thu, Aug 2, 2012 at 6:49 AM, Sylvain Munaut <s.munaut@xxxxxxxxxxxxxxxxxxxx> wrote: > > Hi, > > > Is there a document explaining how radosgw maps things to rados ? I was actually working on such a document. There are some general architectural blog posts in ceph.com, however, some may have been obsolete by now. > > My situation is that I'm currently running various failure scenarios > to see how to recover. One of them includes the complete loss of a > placement group and in my case this lead into a weirdly inconsistent > bucket. I could still add and get object from it, but I couldn't do > list or delete ... ( tracing the error, the cls_head call was failing > in rgw_rados ). Will need logs in order to say what happened exactly, if that's of any interest. You removed some crucial data for the full operation of that bucket. It may also be that some of the data was still cached so that you were able to complete some operations but not the others. > > I also tried starting from scratch by stopping radosgw and deleting > all .rgw .rgw.control .rgw.buckets pools ... and when I restarted, a > 's3 -u list' _still_ showed the previous buckets names ?!? .rgw.buckets pool holds all the objects data by default. .rgw holds all the buckets meta information, aka 'head' (e.g., bucket acls, bucket instance id) .rgw.control holds a notification object that is being used for cache control the users bucket index itself is in .users.uid, where it is kept under an object named <uid>.buckets. Hope that helps, Yehuda -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
