[
https://issues.apache.org/jira/browse/NIFI-7566?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17140871#comment-17140871
]
ASF subversion and git services commented on NIFI-7566:
-------------------------------------------------------
Commit 57c7883f647348aba181440950f01cf1d62846c6 in nifi's branch
refs/heads/master from Mark Payne
[ https://gitbox.apache.org/repos/asf?p=nifi.git;h=57c7883 ]
NIFI-7566: Avoid using Thread.sleep() to wait for Site-to-Site connection to be
handled. Instead, use TimeUnit.timedWait and use Object.notifyAll when setting
the beingServiced flag. This significantly reduces latency and improves
throughput for small-batch site-to-site communications
This closes #4353.
Signed-off-by: Andy LoPresto <[email protected]>
> Site-to-Site uses Thread.sleep() to wait for background thread to complete
> its job. It shouldn't.
> -------------------------------------------------------------------------------------------------
>
> Key: NIFI-7566
> URL: https://issues.apache.org/jira/browse/NIFI-7566
> Project: Apache NiFi
> Issue Type: Improvement
> Components: Core Framework
> Reporter: Mark Payne
> Assignee: Mark Payne
> Priority: Major
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Currently, the StandardPublicPort sits in a loop, calling
> `Thread.sleep(100L);` while it waits for a background thread to complete its
> job. This can add huge latency to the process of servicing a Site-to-Site
> request/connection. Instead, it should use `Object.wait(100);` or similar and
> be notified when the other thread completes its task.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
