You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@river.apache.org by "Gregg Wonderly (JIRA)" <ji...@apache.org> on 2007/10/16 22:46:50 UTC

[jira] Commented: (RIVER-271) Threads Assigned to Socket I/O Tasks should include Host in Name

    [ https://issues.apache.org/jira/browse/RIVER-271?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12535338 ] 

Gregg Wonderly commented on RIVER-271:
--------------------------------------

One of the issues with ThreadPoolExecutor is that there are versions of the JDK that do not include a RuntimeException check for lifecycle management.  Further, threads which die, do not affect the count of active threads.  This means that given a recurring error, of some nature, which causes RuntimeExceptions to occur, eventually there will be no threads executing any of the tasks.

If we make changes to use ThreadPoolExecutor, it will be important to have a custom thread factory, which will keep threads from dieing due to RuntimeException at a minimum.  Further analysis of the existing implementations in JDK5, 6 and 7 would be appropriate to decide on an appropriate implementation.


> Threads Assigned to Socket I/O Tasks should include Host in Name
> ----------------------------------------------------------------
>
>                 Key: RIVER-271
>                 URL: https://issues.apache.org/jira/browse/RIVER-271
>             Project: River
>          Issue Type: Improvement
>    Affects Versions: AR2
>            Reporter: Gregg Wonderly
>            Priority: Minor
>
> Bob Scheifler wrote:
> >Gregg Wonderly wrote:
> >> I guess the next question would be does this seem like a change useful
> >> for the codebase for river?  I.e. should someone open an issue and
> >> attach the diffs?
> >
> > If you wanted this functionality integrated in, I think you'd want
> > to do it at a deeper level, e.g. by changing the task name in
> > MuxServer.dispatchNewRequest.  You'd also want to make similar
> > changes to other task names in that mux code.  (Changes here would
> > also relate to larger changes to take advantage of 1.5-and-later
> > ThreadPoolExecutor). 

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