[jira] Created: (LUCENE-1168) TermVectors index files can become corrupt when autoCommit=false

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

[jira] Created: (LUCENE-1168) TermVectors index files can become corrupt when autoCommit=false

JIRA jira@apache.org
TermVectors index files can become corrupt when autoCommit=false
----------------------------------------------------------------

                 Key: LUCENE-1168
                 URL: https://issues.apache.org/jira/browse/LUCENE-1168
             Project: Lucene - Java
          Issue Type: Bug
          Components: Index
    Affects Versions: 2.3
            Reporter: Michael McCandless
            Assignee: Michael McCandless
             Fix For: 2.4


Spinoff from this thread:

  http://www.gossamer-threads.com/lists/lucene/java-dev/55951

There are actually 2 separate cases here, both only happening when
autoCommit=false:

  * First issue was caused by LUCENE-843 (sigh): if you add a bunch of
    docs with no term vectors, such that 1 or more flushes happen;
    then you add docs that do have term vectors, the tvx file will not
    have enough entries (= corruption).

  * Second issue was caused by bulk merging of term vectors
    (LUCENE-1120 -- only in trunk) and bulk merging of stored fields
    (LUCENE-1043, in 2.3), and only shows when autoCommit=false, and,
    the bulk merging optimization runs.  In this case, the code that
    reads the rawDocs tries to read too far in the tvx/fdx files (it's
    not really index corruption but rather a bug in the rawDocs
    reading).



--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Updated: (LUCENE-1168) TermVectors index files can become corrupt when autoCommit=false

JIRA jira@apache.org

     [ https://issues.apache.org/jira/browse/LUCENE-1168?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Michael McCandless updated LUCENE-1168:
---------------------------------------

    Attachment: LUCENE-1168.patch

Patch attached. I added the original test case as a unit test, plus 2
other tests.

The fixes themselves are trivial (a few lines).  I plan to commit end
of today.

Thanks for finding & reporting this Suresh!

> TermVectors index files can become corrupt when autoCommit=false
> ----------------------------------------------------------------
>
>                 Key: LUCENE-1168
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1168
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Index
>    Affects Versions: 2.3
>            Reporter: Michael McCandless
>            Assignee: Michael McCandless
>             Fix For: 2.4
>
>         Attachments: LUCENE-1168.patch
>
>
> Spinoff from this thread:
>   http://www.gossamer-threads.com/lists/lucene/java-dev/55951
> There are actually 2 separate cases here, both only happening when
> autoCommit=false:
>   * First issue was caused by LUCENE-843 (sigh): if you add a bunch of
>     docs with no term vectors, such that 1 or more flushes happen;
>     then you add docs that do have term vectors, the tvx file will not
>     have enough entries (= corruption).
>   * Second issue was caused by bulk merging of term vectors
>     (LUCENE-1120 -- only in trunk) and bulk merging of stored fields
>     (LUCENE-1043, in 2.3), and only shows when autoCommit=false, and,
>     the bulk merging optimization runs.  In this case, the code that
>     reads the rawDocs tries to read too far in the tvx/fdx files (it's
>     not really index corruption but rather a bug in the rawDocs
>     reading).

--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Resolved: (LUCENE-1168) TermVectors index files can become corrupt when autoCommit=false

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

     [ https://issues.apache.org/jira/browse/LUCENE-1168?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Michael McCandless resolved LUCENE-1168.
----------------------------------------

    Resolution: Fixed

> TermVectors index files can become corrupt when autoCommit=false
> ----------------------------------------------------------------
>
>                 Key: LUCENE-1168
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1168
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Index
>    Affects Versions: 2.3
>            Reporter: Michael McCandless
>            Assignee: Michael McCandless
>             Fix For: 2.4
>
>         Attachments: LUCENE-1168.patch
>
>
> Spinoff from this thread:
>   http://www.gossamer-threads.com/lists/lucene/java-dev/55951
> There are actually 2 separate cases here, both only happening when
> autoCommit=false:
>   * First issue was caused by LUCENE-843 (sigh): if you add a bunch of
>     docs with no term vectors, such that 1 or more flushes happen;
>     then you add docs that do have term vectors, the tvx file will not
>     have enough entries (= corruption).
>   * Second issue was caused by bulk merging of term vectors
>     (LUCENE-1120 -- only in trunk) and bulk merging of stored fields
>     (LUCENE-1043, in 2.3), and only shows when autoCommit=false, and,
>     the bulk merging optimization runs.  In this case, the code that
>     reads the rawDocs tries to read too far in the tvx/fdx files (it's
>     not really index corruption but rather a bug in the rawDocs
>     reading).

--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Commented: (LUCENE-1168) TermVectors index files can become corrupt when autoCommit=false

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/LUCENE-1168?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12568048#action_12568048 ]

Michael McCandless commented on LUCENE-1168:
--------------------------------------------

Backported to 2.3 branch.

> TermVectors index files can become corrupt when autoCommit=false
> ----------------------------------------------------------------
>
>                 Key: LUCENE-1168
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1168
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Index
>    Affects Versions: 2.3
>            Reporter: Michael McCandless
>            Assignee: Michael McCandless
>             Fix For: 2.4
>
>         Attachments: LUCENE-1168.patch
>
>
> Spinoff from this thread:
>   http://www.gossamer-threads.com/lists/lucene/java-dev/55951
> There are actually 2 separate cases here, both only happening when
> autoCommit=false:
>   * First issue was caused by LUCENE-843 (sigh): if you add a bunch of
>     docs with no term vectors, such that 1 or more flushes happen;
>     then you add docs that do have term vectors, the tvx file will not
>     have enough entries (= corruption).
>   * Second issue was caused by bulk merging of term vectors
>     (LUCENE-1120 -- only in trunk) and bulk merging of stored fields
>     (LUCENE-1043, in 2.3), and only shows when autoCommit=false, and,
>     the bulk merging optimization runs.  In this case, the code that
>     reads the rawDocs tries to read too far in the tvx/fdx files (it's
>     not really index corruption but rather a bug in the rawDocs
>     reading).

--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

[jira] Updated: (LUCENE-1168) TermVectors index files can become corrupt when autoCommit=false

JIRA jira@apache.org
In reply to this post by JIRA jira@apache.org

     [ https://issues.apache.org/jira/browse/LUCENE-1168?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Hoss Man updated LUCENE-1168:
-----------------------------

    Fix Version/s: 2.3.1

> TermVectors index files can become corrupt when autoCommit=false
> ----------------------------------------------------------------
>
>                 Key: LUCENE-1168
>                 URL: https://issues.apache.org/jira/browse/LUCENE-1168
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: Index
>    Affects Versions: 2.3
>            Reporter: Michael McCandless
>            Assignee: Michael McCandless
>             Fix For: 2.3.1, 2.4
>
>         Attachments: LUCENE-1168.patch
>
>
> Spinoff from this thread:
>   http://www.gossamer-threads.com/lists/lucene/java-dev/55951
> There are actually 2 separate cases here, both only happening when
> autoCommit=false:
>   * First issue was caused by LUCENE-843 (sigh): if you add a bunch of
>     docs with no term vectors, such that 1 or more flushes happen;
>     then you add docs that do have term vectors, the tvx file will not
>     have enough entries (= corruption).
>   * Second issue was caused by bulk merging of term vectors
>     (LUCENE-1120 -- only in trunk) and bulk merging of stored fields
>     (LUCENE-1043, in 2.3), and only shows when autoCommit=false, and,
>     the bulk merging optimization runs.  In this case, the code that
>     reads the rawDocs tries to read too far in the tvx/fdx files (it's
>     not really index corruption but rather a bug in the rawDocs
>     reading).

--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]