Hi,
I've noticed a strange thing about the ceph's bucket index today. I'm using ceph 0.80 from github. I've uploaded successfully a file using multipart upload (script based on the one included here: http://lists.ceph.com/pipermail/ceph-users-ceph.com/2013-October/034940.html ). The upload works, the file is there, I haven't got any errors. The thing that troubles me is the bucket index log. When I run this command: ./radosgw-admin bilog list --bucket=bucketbig I get the operations that were involved into uploading the file and also - a delete operation! Even though I haven't deleted the object. { "op_id": "00000000006.33.5", "op_tag": "slave-1.4199.6", "op": "del", "object": "Bosphorus", "state": "complete", "index_ver": 6, "timestamp": "2014-07-30 14:57:32.000000Z", "ver": { "pool": 8, "epoch": 30}}, Happens to me every time. Here's the whole bucket log: http://pastebin.com/yr8BNPE2 I understand the fact that the parts are uploaded, merged and in the end removed. But why in the bucket index my final object is also marked for deletion? Is this the expected behaviour, or a bug? Here's also the apache's log just in case: 127.0.0.1 l - [30/Jul/2014:16:57:30 +0200] "PUT /bucketbig/ HTTP/1.1" 200 474 "{Referer}i" "Boto/2.31.1 Python/2.7.6 Linux/3.13.0-32-generic" 127.0.0.1 l - [30/Jul/2014:16:57:30 +0200] "POST /bucketbig/Bosphorus?uploads HTTP/1.1" 200 249 "{Referer}i" "Boto/2.31.1 Python/2.7.6 Linux/3.13.0-32-generic" 127.0.0.1 l - [30/Jul/2014:16:57:30 +0200] "PUT /bucketbig/Bosphorus?uploadId=2/IcMF4IP6JrMphCS6UWEQb4eDTdX-Oub&partNumber=1 HTTP/1.1" 200 - "{Referer}i" "Boto/2.31.1 Python/2.7.6 Linux/3.13.0-32-generic" 127.0.0.1 l - [30/Jul/2014:16:57:32 +0200] "GET /bucketbig/Bosphorus?uploadId=2/IcMF4IP6JrMphCS6UWEQb4eDTdX-Oub HTTP/1.1" 200 643 "{Referer}i" "Boto/2.31.1 Python/2.7.6 Linux/3.13.0-32-generic" 127.0.0.1 l - [30/Jul/2014:16:57:32 +0200] "POST /bucketbig/Bosphorus?uploadId=2/IcMF4IP6JrMphCS6UWEQb4eDTdX-Oub HTTP/1.1" 200 284 "{Referer}i" "Boto/2.31.1 Python/2.7.6 Linux/3.13.0-32-generic" Any help would be appreciated. Thanks, Patrycja Szabłowska _______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
