couchdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefan Kögl (JIRA) <j...@apache.org>
Subject [jira] [Commented] (COUCHDB-1769) Duplicates in View indexes
Date Tue, 28 May 2013 17:31:20 GMT

    [ https://issues.apache.org/jira/browse/COUCHDB-1769?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13668473#comment-13668473
] 

Stefan Kögl commented on COUCHDB-1769:
--------------------------------------

Three views currently have this particular issue. In the log I see entries like the following
for every failed compaction:

{noformat}
[Tue, 28 May 2013 04:30:13 GMT] [error] [<0.27130.82>] ** Generic server <0.27130.82>
terminating
** Last message in was {'EXIT',<0.30247.82>,
                           {view_duplicate_id,
                               <<"05033b78521e97e9cfb06a4bce8c28f7">>}}
** When Server state == {st,<0.27125.82>,couch_mrview_index,<0.30247.82>}
** Reason for termination ==
** unknown_info
{noformat}

Querying the document with

{noformat}
curl http://127.0.0.1:5984/mygpo/05033b78521e97e9cfb06a4bce8c28f7
{noformat}

correctly returns the document.

When checking {{_all_docs}} I see no problem either

{noformat}
curl "http://127.0.0.1:5984/mygpo/_all_docs?key=%2205033b78521e97e9cfb06a4bce8c28f7%22"
{"total_rows":23987690,"offset":654595,"rows":[
{"id":"05033b78521e97e9cfb06a4bce8c28f7","key":"05033b78521e97e9cfb06a4bce8c28f7","value":{"rev":"2-9508e8d08b0b09921e81fb4f1b7925cf"}}
]}
{noformat}

The {{_changes}} looks good as well

{noformat}
$ curl http://127.0.0.1:5984/mygpo/_changes | grep 05033b78521e97e9cfb06a4bce8c28f7
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 5506M    0 5506M    0     0  1237k      0 --:--:--  1:15:56 --:--:-- 1084k{"seq":58690971,"id":"05033b78521e97e9cfb06a4bce8c28f7","changes":[{"rev":"2-9508e8d08b0b09921e81fb4f1b7925cf"}]},
100 5759M    0 5759M    0     0  1130k      0 --:--:--  1:26:54 --:--:--  671k
{noformat}

This particular document contains a timestamp. It looks like the document was added *after*
the last time I fixed such a {{view_duplicate_id}} issue by 0-byte-compaction.
                
> Duplicates in View indexes
> --------------------------
>
>                 Key: COUCHDB-1769
>                 URL: https://issues.apache.org/jira/browse/COUCHDB-1769
>             Project: CouchDB
>          Issue Type: Bug
>          Components: JavaScript View Server
>    Affects Versions: 1.3
>            Reporter: Stefan Kögl
>         Attachments: duplicate.log
>
>
> The compaction of a view (created with CouchDB 1.3) failed due to view_duplicate_id.
The database to which the view belongs was created with 1.2.x but compacted (with the 0-byte-compact-file
trick) with 1.3.
> A full log of the failed compaction can be found at
> https://friendpaste.com/1AL050bzwyLCa9KJUKJw1u
> I experienced the same behaviour already with previous versions but assumed it might
be gone with 1.3, but this seems not to be the case. 
> When such duplicates happen (and appear during compaction), they seem to be at the beginning
of the view index. The view in question has the following _info:
> {noformat}
> {"name":"listeners","view_index":{"signature":"854a8065b438b3e53fcdbbe0a79a8e80","language":"javascript","disk_size":2679632207,"data_size":1026614011,"update_seq":58353111,"purge_seq":0,"updater_running":false,"compact_running":false,"waiting_commit":false,"waiting_clients":0}}
> {noformat}
> Compaction stopped with a 39M large .compact.view file.
> Also the problem seems to happen only one some machines.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message