GitHub user bbende opened a pull request:
https://github.com/apache/nifi/pull/189
NIFI-1436 Fixing race condition in SocketChannelDispatcher
From looking at the test failure that was reported, it was a
ConcurrentModificationException in the close() method of
SocketChannelDispatcher, which happens when looping over selector.keys().
When the processor is unscheduled it calls stop() then close(). The stop
method sets a stopped flag that the run method uses as an exit condition, and
then calls selector.wakeup(). It appears that it would be possible for the run
method to have not fully finished before close() is called, making it possible
for the selector's keys to get modified while close() is happening.
I was never able to produce the test failure locally, so if someone who
encountered it could test this change it would be helpful.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/bbende/nifi NIFI-1436
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/nifi/pull/189.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #189
----
commit a114f582d4578d64d984628567a1d7b6767b4fc5
Author: Bryan Bende <[email protected]>
Date: 2016-01-25T14:44:24Z
NIFI-1436 Fixing race condition in SocketChannelDispatcher where close()
could be called before run() is fully done
----
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
