[ 
https://issues.apache.org/jira/browse/NIFI-6065?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16775259#comment-16775259
 ] 

ASF subversion and git services commented on NIFI-6065:
-------------------------------------------------------

Commit e26bacf69799fc98ce9b4674551ecb520488a6e6 in nifi's branch 
refs/heads/master from Mark Payne
[ https://gitbox.apache.org/repos/asf?p=nifi.git;h=e26bacf ]

NIFI-6068, NIFI-6065: Updated StandardFunnel to avoid looping indefinitely and 
instead transfer no more than 10,000 FlowFiles before returning from onTrigger. 
Updated Local Ports to behavior in the same way. Updated Root Group Ports so 
that instead of blocking for up to 100 milliseconds for an incoming request, it 
blocks for up to 1 millisecond and if nothing is available yields for the 
'bored yield duration'

Signed-off-by: Brandon Devries <[email protected]>

This closes #3328.


> Root Group Ports "waiting" too long
> -----------------------------------
>
>                 Key: NIFI-6065
>                 URL: https://issues.apache.org/jira/browse/NIFI-6065
>             Project: Apache NiFi
>          Issue Type: Improvement
>          Components: Core Framework
>    Affects Versions: 1.9.0
>            Reporter: Brandon DeVries
>            Assignee: Mark Payne
>            Priority: Major
>
> In NIFI-408, a change was made to StandardRootGroupPort to poll for 100 
> millis instead of doing a non-blocking poll, and yielding if there were no 
> results.  While I don't know the initial motivation for this change 
> (thoughts, [~markap14]?), the current implementation is resulting in an 
> excessive number of Timer-Driven threads stuck in the TIMED_WAITING state.  
> Essentially, if you have a high number of low traffic remote ports, you lose 
> a substantial chunk of your timer driven thread pool to simply "waiting".
> My first thought would be to revert the change, possibly even without the 
> context.yield.  Second would be to drastically reduce the poll timeout... 10 
> or 5 millis, although any choice is somewhat arbitrary and flow-dependent. 
> EDIT: which probably makes it a candidate for a nifi.property 
> Any other thoughts or suggestions would be appreciated.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to