All of lore.kernel.org
 help / color / mirror / Atom feed
* rgw: object null version delete
@ 2016-06-02  9:21 Yang Joseph
  2016-06-02 13:17 ` Orit Wasserman
  2016-06-14  3:13 ` rgw: object listing hangs on versioned bucket KIMURA Osamu
  0 siblings, 2 replies; 7+ messages in thread
From: Yang Joseph @ 2016-06-02  9:21 UTC (permalink / raw)
  To: ceph-devel; +Cc: Javen Wu

Hello,

Radosgw version Hammer (0.94.5) can not delete a null version key which 
created before
turn on bucket versioning[1]. And its value can still be accessed. In 
order to
solve this problem, I applied the changes of [2]. But sometimes the test 
case
can pass, sometimes not.

For all my osd daemons, I set breakpoints at 
rgw_bucket_unlink_instance/rgw_bucket_read_olh_log.
I noticed that a failed case's unlink/read_olh_log requests never 
reached osd side.

How to fix this problem? Any suggestions?

Thx,

joseph

ref:

[1] How to reproduce the bug:

     - create bucket
     - put key AAA
     - turn on bucket versioning
     - read AAA // expect 404
     - delete AAA
     - delete bucket

[2] hammer: rgw: convert plain object to versioned (with null version) 
when removing #8755 
https://github.com/ceph/ceph/pull/8755/commits/12cf255eb2ae666afb29df40d61de754257f7f28?diff=split

[3] bilog

# radosgw-admin bilog list --bucket=4e00f488-28a1-11e6-a9e6-002590ae43ca 
--cluster rgwltt
[
     {
         "op_id": "00000000001.77.2",
         "op_tag": "rgwltt-rgwltt.6547.113",
         "op": "write",
         "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
         "instance": "",
         "state": "pending",
         "index_ver": 1,
         "timestamp": "0.000000",
         "ver": {
             "pool": -1,
             "epoch": 0
         },
         "versioned": false
     },
     {
         "op_id": "00000000002.78.3",
         "op_tag": "rgwltt-rgwltt.6547.113",
         "op": "write",
         "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
         "instance": "",
         "state": "complete",
         "index_ver": 2,
         "timestamp": "2016-06-02 09:06:34.000000Z",
         "ver": {
             "pool": 15,
             "epoch": 17
         },
         "versioned": false
     },
     {
         "op_id": "00000000003.79.5",
         "op_tag": "00000000574ff71azhrp4zx1y11qcsv2",
         "op": "unlink_instance",
         "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
         "instance": "",
         "state": "complete",
         "index_ver": 3,
         "timestamp": "2016-06-02 09:06:34.787122Z",
         "ver": {
             "pool": -1,
             "epoch": 2
         },
         "versioned": true
     },
     {
         "op_id": "00000000004.80.5",
         "op_tag": "00000000574ff71cgeeopd12arjscnds",
         "op": "unlink_instance",
         "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
         "instance": "",
         "state": "complete",
         "index_ver": 4,
         "timestamp": "2016-06-02 09:06:36.760707Z",
         "ver": {
             "pool": -1,
             "epoch": 3
         },
         "versioned": true
     }

]



^ permalink raw reply	[flat|nested] 7+ messages in thread
* Re: rgw: object null version delete
  2016-06-02  9:21 rgw: object null version delete Yang Joseph
@ 2016-06-02 13:17 ` Orit Wasserman
  2016-06-07  2:30   ` Yang Joseph
  2016-06-14  3:13 ` rgw: object listing hangs on versioned bucket KIMURA Osamu
  1 sibling, 1 reply; 7+ messages in thread
From: Orit Wasserman @ 2016-06-02 13:17 UTC (permalink / raw)
  To: Yang Joseph; +Cc: ceph-devel, Javen Wu

The fix is being backported to hammer:
http://tracker.ceph.com/issues/15254

On Thu, Jun 2, 2016 at 11:21 AM, Yang Joseph <joseph.yang@xtaotech.com> wrote:
> Hello,
>
> Radosgw version Hammer (0.94.5) can not delete a null version key which
> created before
> turn on bucket versioning[1]. And its value can still be accessed. In order
> to
> solve this problem, I applied the changes of [2]. But sometimes the test
> case
> can pass, sometimes not.
>
> For all my osd daemons, I set breakpoints at
> rgw_bucket_unlink_instance/rgw_bucket_read_olh_log.
> I noticed that a failed case's unlink/read_olh_log requests never reached
> osd side.
>
> How to fix this problem? Any suggestions?
>
> Thx,
>
> joseph
>
> ref:
>
> [1] How to reproduce the bug:
>
>     - create bucket
>     - put key AAA
>     - turn on bucket versioning
>     - read AAA // expect 404
>     - delete AAA
>     - delete bucket
>
> [2] hammer: rgw: convert plain object to versioned (with null version) when
> removing #8755
> https://github.com/ceph/ceph/pull/8755/commits/12cf255eb2ae666afb29df40d61de754257f7f28?diff=split
>
> [3] bilog
>
> # radosgw-admin bilog list --bucket=4e00f488-28a1-11e6-a9e6-002590ae43ca
> --cluster rgwltt
> [
>     {
>         "op_id": "00000000001.77.2",
>         "op_tag": "rgwltt-rgwltt.6547.113",
>         "op": "write",
>         "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
>         "instance": "",
>         "state": "pending",
>         "index_ver": 1,
>         "timestamp": "0.000000",
>         "ver": {
>             "pool": -1,
>             "epoch": 0
>         },
>         "versioned": false
>     },
>     {
>         "op_id": "00000000002.78.3",
>         "op_tag": "rgwltt-rgwltt.6547.113",
>         "op": "write",
>         "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
>         "instance": "",
>         "state": "complete",
>         "index_ver": 2,
>         "timestamp": "2016-06-02 09:06:34.000000Z",
>         "ver": {
>             "pool": 15,
>             "epoch": 17
>         },
>         "versioned": false
>     },
>     {
>         "op_id": "00000000003.79.5",
>         "op_tag": "00000000574ff71azhrp4zx1y11qcsv2",
>         "op": "unlink_instance",
>         "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
>         "instance": "",
>         "state": "complete",
>         "index_ver": 3,
>         "timestamp": "2016-06-02 09:06:34.787122Z",
>         "ver": {
>             "pool": -1,
>             "epoch": 2
>         },
>         "versioned": true
>     },
>     {
>         "op_id": "00000000004.80.5",
>         "op_tag": "00000000574ff71cgeeopd12arjscnds",
>         "op": "unlink_instance",
>         "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
>         "instance": "",
>         "state": "complete",
>         "index_ver": 4,
>         "timestamp": "2016-06-02 09:06:36.760707Z",
>         "ver": {
>             "pool": -1,
>             "epoch": 3
>         },
>         "versioned": true
>     }
>
> ]
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

^ permalink raw reply	[flat|nested] 7+ messages in thread
* Re: rgw: object null version delete
  2016-06-02 13:17 ` Orit Wasserman
@ 2016-06-07  2:30   ` Yang Joseph
  2016-06-07  8:32     ` Orit Wasserman
  0 siblings, 1 reply; 7+ messages in thread
From: Yang Joseph @ 2016-06-07  2:30 UTC (permalink / raw)
  To: Orit Wasserman; +Cc: ceph-devel, Javen Wu

When will this work be done? May I know the plan?

On 06/02/2016 09:17 PM, Orit Wasserman wrote:
> The fix is being backported to hammer:
> http://tracker.ceph.com/issues/15254
>
> On Thu, Jun 2, 2016 at 11:21 AM, Yang Joseph <joseph.yang@xtaotech.com> wrote:
>> Hello,
>>
>> Radosgw version Hammer (0.94.5) can not delete a null version key which
>> created before
>> turn on bucket versioning[1]. And its value can still be accessed. In order
>> to
>> solve this problem, I applied the changes of [2]. But sometimes the test
>> case
>> can pass, sometimes not.
>>
>> For all my osd daemons, I set breakpoints at
>> rgw_bucket_unlink_instance/rgw_bucket_read_olh_log.
>> I noticed that a failed case's unlink/read_olh_log requests never reached
>> osd side.
>>
>> How to fix this problem? Any suggestions?
>>
>> Thx,
>>
>> joseph
>>
>> ref:
>>
>> [1] How to reproduce the bug:
>>
>>      - create bucket
>>      - put key AAA
>>      - turn on bucket versioning
>>      - read AAA // expect 404
>>      - delete AAA
>>      - delete bucket
>>
>> [2] hammer: rgw: convert plain object to versioned (with null version) when
>> removing #8755
>> https://github.com/ceph/ceph/pull/8755/commits/12cf255eb2ae666afb29df40d61de754257f7f28?diff=split
>>
>> [3] bilog
>>
>> # radosgw-admin bilog list --bucket=4e00f488-28a1-11e6-a9e6-002590ae43ca
>> --cluster rgwltt
>> [
>>      {
>>          "op_id": "00000000001.77.2",
>>          "op_tag": "rgwltt-rgwltt.6547.113",
>>          "op": "write",
>>          "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
>>          "instance": "",
>>          "state": "pending",
>>          "index_ver": 1,
>>          "timestamp": "0.000000",
>>          "ver": {
>>              "pool": -1,
>>              "epoch": 0
>>          },
>>          "versioned": false
>>      },
>>      {
>>          "op_id": "00000000002.78.3",
>>          "op_tag": "rgwltt-rgwltt.6547.113",
>>          "op": "write",
>>          "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
>>          "instance": "",
>>          "state": "complete",
>>          "index_ver": 2,
>>          "timestamp": "2016-06-02 09:06:34.000000Z",
>>          "ver": {
>>              "pool": 15,
>>              "epoch": 17
>>          },
>>          "versioned": false
>>      },
>>      {
>>          "op_id": "00000000003.79.5",
>>          "op_tag": "00000000574ff71azhrp4zx1y11qcsv2",
>>          "op": "unlink_instance",
>>          "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
>>          "instance": "",
>>          "state": "complete",
>>          "index_ver": 3,
>>          "timestamp": "2016-06-02 09:06:34.787122Z",
>>          "ver": {
>>              "pool": -1,
>>              "epoch": 2
>>          },
>>          "versioned": true
>>      },
>>      {
>>          "op_id": "00000000004.80.5",
>>          "op_tag": "00000000574ff71cgeeopd12arjscnds",
>>          "op": "unlink_instance",
>>          "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
>>          "instance": "",
>>          "state": "complete",
>>          "index_ver": 4,
>>          "timestamp": "2016-06-02 09:06:36.760707Z",
>>          "ver": {
>>              "pool": -1,
>>              "epoch": 3
>>          },
>>          "versioned": true
>>      }
>>
>> ]
>>
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html


^ permalink raw reply	[flat|nested] 7+ messages in thread
* Re: rgw: object null version delete
  2016-06-07  2:30   ` Yang Joseph
@ 2016-06-07  8:32     ` Orit Wasserman
  0 siblings, 0 replies; 7+ messages in thread
From: Orit Wasserman @ 2016-06-07  8:32 UTC (permalink / raw)
  To: Yang Joseph; +Cc: ceph-devel, Javen Wu

It should be included in the next hammer release

On Tue, Jun 7, 2016 at 4:30 AM, Yang Joseph <joseph.yang@xtaotech.com> wrote:
> When will this work be done? May I know the plan?
>
>
> On 06/02/2016 09:17 PM, Orit Wasserman wrote:
>>
>> The fix is being backported to hammer:
>> http://tracker.ceph.com/issues/15254
>>
>> On Thu, Jun 2, 2016 at 11:21 AM, Yang Joseph <joseph.yang@xtaotech.com>
>> wrote:
>>>
>>> Hello,
>>>
>>> Radosgw version Hammer (0.94.5) can not delete a null version key which
>>> created before
>>> turn on bucket versioning[1]. And its value can still be accessed. In
>>> order
>>> to
>>> solve this problem, I applied the changes of [2]. But sometimes the test
>>> case
>>> can pass, sometimes not.
>>>
>>> For all my osd daemons, I set breakpoints at
>>> rgw_bucket_unlink_instance/rgw_bucket_read_olh_log.
>>> I noticed that a failed case's unlink/read_olh_log requests never reached
>>> osd side.
>>>
>>> How to fix this problem? Any suggestions?
>>>
>>> Thx,
>>>
>>> joseph
>>>
>>> ref:
>>>
>>> [1] How to reproduce the bug:
>>>
>>>      - create bucket
>>>      - put key AAA
>>>      - turn on bucket versioning
>>>      - read AAA // expect 404
>>>      - delete AAA
>>>      - delete bucket
>>>
>>> [2] hammer: rgw: convert plain object to versioned (with null version)
>>> when
>>> removing #8755
>>>
>>> https://github.com/ceph/ceph/pull/8755/commits/12cf255eb2ae666afb29df40d61de754257f7f28?diff=split
>>>
>>> [3] bilog
>>>
>>> # radosgw-admin bilog list --bucket=4e00f488-28a1-11e6-a9e6-002590ae43ca
>>> --cluster rgwltt
>>> [
>>>      {
>>>          "op_id": "00000000001.77.2",
>>>          "op_tag": "rgwltt-rgwltt.6547.113",
>>>          "op": "write",
>>>          "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
>>>          "instance": "",
>>>          "state": "pending",
>>>          "index_ver": 1,
>>>          "timestamp": "0.000000",
>>>          "ver": {
>>>              "pool": -1,
>>>              "epoch": 0
>>>          },
>>>          "versioned": false
>>>      },
>>>      {
>>>          "op_id": "00000000002.78.3",
>>>          "op_tag": "rgwltt-rgwltt.6547.113",
>>>          "op": "write",
>>>          "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
>>>          "instance": "",
>>>          "state": "complete",
>>>          "index_ver": 2,
>>>          "timestamp": "2016-06-02 09:06:34.000000Z",
>>>          "ver": {
>>>              "pool": 15,
>>>              "epoch": 17
>>>          },
>>>          "versioned": false
>>>      },
>>>      {
>>>          "op_id": "00000000003.79.5",
>>>          "op_tag": "00000000574ff71azhrp4zx1y11qcsv2",
>>>          "op": "unlink_instance",
>>>          "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
>>>          "instance": "",
>>>          "state": "complete",
>>>          "index_ver": 3,
>>>          "timestamp": "2016-06-02 09:06:34.787122Z",
>>>          "ver": {
>>>              "pool": -1,
>>>              "epoch": 2
>>>          },
>>>          "versioned": true
>>>      },
>>>      {
>>>          "op_id": "00000000004.80.5",
>>>          "op_tag": "00000000574ff71cgeeopd12arjscnds",
>>>          "op": "unlink_instance",
>>>          "object": "4e1274ec-28a1-11e6-a9e6-002590ae43ca",
>>>          "instance": "",
>>>          "state": "complete",
>>>          "index_ver": 4,
>>>          "timestamp": "2016-06-02 09:06:36.760707Z",
>>>          "ver": {
>>>              "pool": -1,
>>>              "epoch": 3
>>>          },
>>>          "versioned": true
>>>      }
>>>
>>> ]
>>>
>>>
>>> --
>>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
>>> the body of a message to majordomo@vger.kernel.org
>>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
>

^ permalink raw reply	[flat|nested] 7+ messages in thread
* rgw: object listing hangs on versioned bucket
  2016-06-02  9:21 rgw: object null version delete Yang Joseph
  2016-06-02 13:17 ` Orit Wasserman
@ 2016-06-14  3:13 ` KIMURA Osamu
  2016-06-14 10:21   ` Orit Wasserman
  1 sibling, 1 reply; 7+ messages in thread
From: KIMURA Osamu @ 2016-06-14  3:13 UTC (permalink / raw)
  To: ceph-devel, Yehuda Sadeh-Weinraub

[-- Attachment #1: Type: text/plain, Size: 2187 bytes --]

Hello,

An our customer got a problem on RADOSGW.
The customer tried to list objects on a versioned bucket with S3 API,
then they got 500 server error due to FastCGI timeout.
RGW's internal listing process continued over the night (approx 10hours)
until they restarted the RGW. Please see attached radosgw log.

The behavior seems to be similar to bug#12913, but they are using
v0.94.6 of RGW. The bug should be resolved in the version.
http://tracker.ceph.com/issues/12913

As we can find in the log, RGW repeatedly tried to get information
for a specific object (testfile1896.txt).

The object has null version and another version as bellow.
Yes, versioning of the bucket was enabled after some objects had been
put in the bucket.

Similar condition can be found in bug#15243, but the fix only affects
to deletion.
http://tracker.ceph.com/issues/15243

Are there any problem remained in versioned object with null version?


Regards,
KIMURA

----------------------
<?xml version="1.0" encoding="UTF-8"?>
<ListVersionsResult xmlns="http://s3.amazonaws.com/doc/2006-03-01/">
  <Name>hojo-bucket-02</Name>
  <Prefix>testfile1896.txt</Prefix>
  <KeyMarker></KeyMarker>
  <MaxKeys>1000</MaxKeys>
  <IsTruncated>false</IsTruncated>
  <Version>
   <Key>testfile1896.txt</Key>
   <VersionId>2MSeTpZoDSqt2nTk1oYHZNmF1e84C.1</VersionId>
   <IsLatest>true</IsLatest>
   <LastModified>2016-05-30T08:39:52.000Z</LastModified>
   <ETag>&quot;1d24c7924b9798bb9064dcb043b3d989&quot;</ETag>
   <Size>3152</Size>
   <StorageClass>STANDARD</StorageClass>
   <Owner>
    <ID>XXXXXXXX</ID>
    <DisplayName>XXXXXXXX</DisplayName>
   </Owner>
  </Version>
  <Version>
   <Key>testfile1896.txt</Key>
   <VersionId>null</VersionId>
   <IsLatest>false</IsLatest>
   <LastModified>2016-05-30T02:43:22.000Z</LastModified>
   <ETag>&quot;20a4fc4c12598089a8937496a5eba67e&quot;</ETag>
   <Size>3052</Size>
   <StorageClass>STANDARD</StorageClass>
   <Owner>
    <ID>XXXXXXXX</ID>
    <DisplayName>XXXXXXXX</DisplayName>
   </Owner>
  </Version>
</ListVersionsResult>


-- 
KIMURA Osamu / 木村 修
Engineering Department, Storage Development Division,
Data Center Platform Business Unit, FUJITSU LIMITED

[-- Attachment #2: radosgw.log.gz --]
[-- Type: application/gzip, Size: 55019 bytes --]

^ permalink raw reply	[flat|nested] 7+ messages in thread
* Re: rgw: object listing hangs on versioned bucket
  2016-06-14  3:13 ` rgw: object listing hangs on versioned bucket KIMURA Osamu
@ 2016-06-14 10:21   ` Orit Wasserman
  2016-06-15  5:26     ` KIMURA Osamu
  0 siblings, 1 reply; 7+ messages in thread
From: Orit Wasserman @ 2016-06-14 10:21 UTC (permalink / raw)
  To: KIMURA Osamu; +Cc: ceph-devel, Yehuda Sadeh-Weinraub

It could be a different problem, can you open a new issue?

Thanks,
Orit

On Tue, Jun 14, 2016 at 5:13 AM, KIMURA Osamu
<kimura.osamu@jp.fujitsu.com> wrote:
> Hello,
>
> An our customer got a problem on RADOSGW.
> The customer tried to list objects on a versioned bucket with S3 API,
> then they got 500 server error due to FastCGI timeout.
> RGW's internal listing process continued over the night (approx 10hours)
> until they restarted the RGW. Please see attached radosgw log.
>
> The behavior seems to be similar to bug#12913, but they are using
> v0.94.6 of RGW. The bug should be resolved in the version.
> http://tracker.ceph.com/issues/12913
>
> As we can find in the log, RGW repeatedly tried to get information
> for a specific object (testfile1896.txt).
>
> The object has null version and another version as bellow.
> Yes, versioning of the bucket was enabled after some objects had been
> put in the bucket.
>
> Similar condition can be found in bug#15243, but the fix only affects
> to deletion.
> http://tracker.ceph.com/issues/15243
>
> Are there any problem remained in versioned object with null version?
>
>
> Regards,
> KIMURA
>
> ----------------------
> <?xml version="1.0" encoding="UTF-8"?>
> <ListVersionsResult xmlns="http://s3.amazonaws.com/doc/2006-03-01/">
>  <Name>hojo-bucket-02</Name>
>  <Prefix>testfile1896.txt</Prefix>
>  <KeyMarker></KeyMarker>
>  <MaxKeys>1000</MaxKeys>
>  <IsTruncated>false</IsTruncated>
>  <Version>
>   <Key>testfile1896.txt</Key>
>   <VersionId>2MSeTpZoDSqt2nTk1oYHZNmF1e84C.1</VersionId>
>   <IsLatest>true</IsLatest>
>   <LastModified>2016-05-30T08:39:52.000Z</LastModified>
>   <ETag>&quot;1d24c7924b9798bb9064dcb043b3d989&quot;</ETag>
>   <Size>3152</Size>
>   <StorageClass>STANDARD</StorageClass>
>   <Owner>
>    <ID>XXXXXXXX</ID>
>    <DisplayName>XXXXXXXX</DisplayName>
>   </Owner>
>  </Version>
>  <Version>
>   <Key>testfile1896.txt</Key>
>   <VersionId>null</VersionId>
>   <IsLatest>false</IsLatest>
>   <LastModified>2016-05-30T02:43:22.000Z</LastModified>
>   <ETag>&quot;20a4fc4c12598089a8937496a5eba67e&quot;</ETag>
>   <Size>3052</Size>
>   <StorageClass>STANDARD</StorageClass>
>   <Owner>
>    <ID>XXXXXXXX</ID>
>    <DisplayName>XXXXXXXX</DisplayName>
>   </Owner>
>  </Version>
> </ListVersionsResult>
>
>
> --
> KIMURA Osamu / 木村 修
> Engineering Department, Storage Development Division,
> Data Center Platform Business Unit, FUJITSU LIMITED
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

^ permalink raw reply	[flat|nested] 7+ messages in thread
* Re: rgw: object listing hangs on versioned bucket
  2016-06-14 10:21   ` Orit Wasserman
@ 2016-06-15  5:26     ` KIMURA Osamu
  0 siblings, 0 replies; 7+ messages in thread
From: KIMURA Osamu @ 2016-06-15  5:26 UTC (permalink / raw)
  To: ceph-devel; +Cc: Orit Wasserman, Yehuda Sadeh-Weinraub

I opened a ticket with earlier log.
http://tracker.ceph.com/issues/16309

Thank you,
KIMURA

On 2016/06/14 19:21, Orit Wasserman wrote:
> It could be a different problem, can you open a new issue?
>
> Thanks,
> Orit
>
> On Tue, Jun 14, 2016 at 5:13 AM, KIMURA Osamu
> <kimura.osamu@jp.fujitsu.com> wrote:
>> Hello,
>>
>> An our customer got a problem on RADOSGW.
>> The customer tried to list objects on a versioned bucket with S3 API,
>> then they got 500 server error due to FastCGI timeout.
>> RGW's internal listing process continued over the night (approx 10hours)
>> until they restarted the RGW. Please see attached radosgw log.
>>
>> The behavior seems to be similar to bug#12913, but they are using
>> v0.94.6 of RGW. The bug should be resolved in the version.
>> http://tracker.ceph.com/issues/12913
>>
>> As we can find in the log, RGW repeatedly tried to get information
>> for a specific object (testfile1896.txt).
>>
>> The object has null version and another version as bellow.
>> Yes, versioning of the bucket was enabled after some objects had been
>> put in the bucket.
>>
>> Similar condition can be found in bug#15243, but the fix only affects
>> to deletion.
>> http://tracker.ceph.com/issues/15243
>>
>> Are there any problem remained in versioned object with null version?
>>
>>
>> Regards,
>> KIMURA
>>
>> ----------------------
>> <?xml version="1.0" encoding="UTF-8"?>
>> <ListVersionsResult xmlns="http://s3.amazonaws.com/doc/2006-03-01/">
>>  <Name>hojo-bucket-02</Name>
>>  <Prefix>testfile1896.txt</Prefix>
>>  <KeyMarker></KeyMarker>
>>  <MaxKeys>1000</MaxKeys>
>>  <IsTruncated>false</IsTruncated>
>>  <Version>
>>   <Key>testfile1896.txt</Key>
>>   <VersionId>2MSeTpZoDSqt2nTk1oYHZNmF1e84C.1</VersionId>
>>   <IsLatest>true</IsLatest>
>>   <LastModified>2016-05-30T08:39:52.000Z</LastModified>
>>   <ETag>&quot;1d24c7924b9798bb9064dcb043b3d989&quot;</ETag>
>>   <Size>3152</Size>
>>   <StorageClass>STANDARD</StorageClass>
>>   <Owner>
>>    <ID>XXXXXXXX</ID>
>>    <DisplayName>XXXXXXXX</DisplayName>
>>   </Owner>
>>  </Version>
>>  <Version>
>>   <Key>testfile1896.txt</Key>
>>   <VersionId>null</VersionId>
>>   <IsLatest>false</IsLatest>
>>   <LastModified>2016-05-30T02:43:22.000Z</LastModified>
>>   <ETag>&quot;20a4fc4c12598089a8937496a5eba67e&quot;</ETag>
>>   <Size>3052</Size>
>>   <StorageClass>STANDARD</StorageClass>
>>   <Owner>
>>    <ID>XXXXXXXX</ID>
>>    <DisplayName>XXXXXXXX</DisplayName>
>>   </Owner>
>>  </Version>
>> </ListVersionsResult>

-- 
KIMURA Osamu / 木村 修
Engineering Department, Storage Development Division,
Data Center Platform Business Unit, FUJITSU LIMITED

^ permalink raw reply	[flat|nested] 7+ messages in thread
end of thread, other threads:[~2016-06-15  5:26 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-06-02  9:21 rgw: object null version delete Yang Joseph
2016-06-02 13:17 ` Orit Wasserman
2016-06-07  2:30   ` Yang Joseph
2016-06-07  8:32     ` Orit Wasserman
2016-06-14  3:13 ` rgw: object listing hangs on versioned bucket KIMURA Osamu
2016-06-14 10:21   ` Orit Wasserman
2016-06-15  5:26     ` KIMURA Osamu

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.