You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Jason Rutherglen (JIRA)" <ji...@apache.org> on 2009/11/13 20:11:39 UTC
[jira] Created: (LUCENE-2063) Use thread pool in
ConcurrentMergeScheduler
Use thread pool in ConcurrentMergeScheduler
-------------------------------------------
Key: LUCENE-2063
URL: https://issues.apache.org/jira/browse/LUCENE-2063
Project: Lucene - Java
Issue Type: Improvement
Components: Index
Affects Versions: 2.9.1
Reporter: Jason Rutherglen
Priority: Minor
Fix For: 3.1
Currently it looks like CMS creates a new thread object for each
merge, which may not be expensive anymore on Java5+ JVMs,
however we can fairly simply implement the Java5 thread pooling.
Also I'm thinking we may be interested in using thread pools for
other tasks in IndexWriter (such as LUCENE-2047 performing
deletes in the background).
--
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: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
[jira] Commented: (LUCENE-2063) Use thread pool in
ConcurrentMergeScheduler
Posted by "Jason Rutherglen (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2063?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12777666#action_12777666 ]
Jason Rutherglen commented on LUCENE-2063:
------------------------------------------
Right, this isn't the highest priority, however try running the
LUCENE-1313 and you'll see a lot of unique thread objects being
created because of the small segment merges happening quickly on
the RAMDir.
It can also potentially be used for LUCENE-2047, or a follow on
async patch.
> Use thread pool in ConcurrentMergeScheduler
> -------------------------------------------
>
> Key: LUCENE-2063
> URL: https://issues.apache.org/jira/browse/LUCENE-2063
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Affects Versions: 2.9.1
> Reporter: Jason Rutherglen
> Priority: Minor
> Fix For: 3.1
>
>
> Currently it looks like CMS creates a new thread object for each
> merge, which may not be expensive anymore on Java5+ JVMs,
> however we can fairly simply implement the Java5 thread pooling.
> Also I'm thinking we may be interested in using thread pools for
> other tasks in IndexWriter (such as LUCENE-2047 performing
> deletes in the background).
--
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: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
[jira] Commented: (LUCENE-2063) Use thread pool in
ConcurrentMergeScheduler
Posted by "Simon Willnauer (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-2063?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12777645#action_12777645 ]
Simon Willnauer commented on LUCENE-2063:
-----------------------------------------
One of the issues with this is that the code depends on Thread instead of Runnable / Callable. I have looked into this class too and figured that simply exchanging the merger threads with runnables would not work.
A way of doing this is to subclass ExecutorService where we can expose the required functionality on a pooled basis.
Beside this I'm not sure if there is any benefit performance / resource wise as merge threads are not needed as frequently as other threads. Threadpools make a lot of sense once you use threads frequently but keeping those merge threads around might not be required. Not sure if this is really needed though.
simon
> Use thread pool in ConcurrentMergeScheduler
> -------------------------------------------
>
> Key: LUCENE-2063
> URL: https://issues.apache.org/jira/browse/LUCENE-2063
> Project: Lucene - Java
> Issue Type: Improvement
> Components: Index
> Affects Versions: 2.9.1
> Reporter: Jason Rutherglen
> Priority: Minor
> Fix For: 3.1
>
>
> Currently it looks like CMS creates a new thread object for each
> merge, which may not be expensive anymore on Java5+ JVMs,
> however we can fairly simply implement the Java5 thread pooling.
> Also I'm thinking we may be interested in using thread pools for
> other tasks in IndexWriter (such as LUCENE-2047 performing
> deletes in the background).
--
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: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org