[jira] [Commented] (JCR-3146) Text extraction may congest thread pool in the repository

["Marcel Reutegger (Commented) (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/JCR-3146?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13150424#comment-13150424 ] 

Marcel Reutegger commented on JCR-3146:
---------------------------------------

In general looks good to me.

I'm not sure the communication between the threads in JackrabbitThreadPool
is 100% correct. E.g. the first statement in RetryLowPriorityTask.run()
checks if the queue is empty. To me it seems like this should never
happen, right?

Style:
Should we rather keep the JackrabbitThreadPool package private and
only expose the marker as public interface? How about renaming
the LOW_PRIORITY_MARKER to LowPriorityTask and extend it from
Runnable? That way a client wouldn't have to implement Runnable
and the marker interface.

Minor:
 The method waitForTextExtractionTasksToFinish() already does an
index flush at the end. Aren't the additional index flush calls
in IndexingQueueTest now obsolete?

                
> Text extraction may congest thread pool in the repository
> ---------------------------------------------------------
>
>                 Key: JCR-3146
>                 URL: https://issues.apache.org/jira/browse/JCR-3146
>             Project: Jackrabbit Content Repository
>          Issue Type: Improvement
>          Components: jackrabbit-core
>            Reporter: Alex Parvulescu
>            Priority: Minor
>         Attachments: JCR-3146.patch
>
>
> Text extraction congests the thread pool in the repository when e.g. many PDFs are loaded into the workspace. Tasks submitted by the index merger are delayed because of that and will result in many index segment folders.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira