lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Michael McCandless (JIRA)" <j...@apache.org>
Subject [jira] Commented: (LUCENE-735) Simple tool to back-convert from lockless to pre-lockless file format
Date Thu, 30 Nov 2006 11:56:22 GMT
    [ http://issues.apache.org/jira/browse/LUCENE-735?page=comments#action_12454591 ] 
            
Michael McCandless commented on LUCENE-735:
-------------------------------------------

To use this, apply the patch to the Lucene trunk, then ant jar-core then run this:

    java org.apache.lucene.index.ConvertPreLockless <path-to-index>

The conversion is in place, meaning, after this tool runs, your <path-to-index> should
be in 2.0 file format.

> Simple tool to back-convert from lockless to pre-lockless file format
> ---------------------------------------------------------------------
>
>                 Key: LUCENE-735
>                 URL: http://issues.apache.org/jira/browse/LUCENE-735
>             Project: Lucene - Java
>          Issue Type: Improvement
>          Components: Index
>    Affects Versions: 2.1
>            Reporter: Michael McCandless
>         Assigned To: Michael McCandless
>            Priority: Minor
>             Fix For: 2.1
>
>         Attachments: LUCENE-735.patch
>
>
> Simple tool to back-convert from lockless to pre-lockless file format
> The changes for lockless commits (LUCENE-701 ) are fairly minor and so
> creating a tool to convert a lockless format index back to a
> pre-lockless format index is 1) fairly simple, and 2) useful at least
> for brave souls who want to try lockless but have the freedom to roll
> back to Lucene 2.0, using the same index, if anything goes wrong.
> I will attach an initial patch.
> This has not yet received extensive testing so please be extremely
> careful if you use this in production!  I've only done minimal testing
> so far: using IndexFiles to produce an index under lockless,
> converting it to pre-lockless, and then doing searches against that
> index with 2.0.  More testing is clearly needed to ensure separate
> deletes, separate norms, etc, are working correctly.
> The tool prints details of what it did, eg:
>   >> java org.apache.lucene.index.ConvertPreLockless index
>   3 segments in index
>   segment 0: not compound file format
>     has deletions
>     rename _a_2.del to _a.del
>     no separate norms
>   segment 1: not compound file format
>     has deletions
>     rename _b_1.del to _b.del
>     no separate norms
>   segment 2: not compound file format
>     has deletions
>     rename _c_1.del to _c.del
>     no separate norms
>   wrote "segments" file
>   rename segments_8 to segments_8.old
> Caveats:
>   * Tread very carefully!  Test first in a sandox, etc.
>   * Make sure you only run this tool on an index that is not in use by
>     any reader/writers, else you could have problems: the tool
>     currently does not acquire the write lock even though it's
>     modifying the index.
>   * On Windows only: if your index has any un-referenced files (ie,
>     files that should have been deleted but were in use at the time)
>     at the time you run this tool, then they will never be deleted
>     (ie, pre-lockless Lucene won't know to delete them).

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org


Mime
View raw message