Hey Chris, Did this end up working out for you? Also did you by chance get a stack-trace of the thread that was hanging? It helps us to debug and potentially fix a problem when we know what's failing/stuck.
Joe - - - - - - Joseph Percivall linkedin.com/in/Percivall e: [email protected] On Monday, November 23, 2015 9:52 AM, Chris Teoh <[email protected]> wrote: Ah I mean the stuck processor wouldn't let me stop it. Restarting NiFi saw them coming back in stuck state where the process appears to still be running but no right click stop or start option available. On Tue, 24 Nov 2015 at 1:49 AM, Joe Witt <[email protected]> wrote: Good point on stopping just that process and didn't realize the >filename trick was an option - cool > >On Mon, Nov 23, 2015 at 9:30 AM, Mark Payne <[email protected]> wrote: >> Chris, >> >> Just to expand upon what Joe mentioned below. >> >> When you say that it immediately becomes stuck when you restart, I assume >> that means that you're restarting with it running. If you tell the processor >> to stop >> and restart, it should not begin running. This way you wouldn't need to >> change >> the autoResumeState. Unless I'm misunderstanding something? >> >> Also, when you run "bin/nifi.sh dump" it is often easier to specify a >> filename there >> such as "bin/nifi.sh dump thread-dump.txt" so that it writes the thread dump >> to the >> filename specified, rather than writing it to the logs. >> >> Thanks >> -Mark >> >> >>> On Nov 23, 2015, at 9:25 AM, Joe Witt <[email protected]> wrote: >>> >>> Chris, >>> >>> If you run into a case of a stuck thread we'd love to see the stack >>> trace. You can generate one by running 'bin/nifi.sh dump' and sending >>> us the logs. I believe it is bootstrap.log specifically. >>> >>> If you have a processor which immediately gets stuck after startup >>> where it takes a thread and never relinquishes even if told to stop >>> then that does present a tricky situation for a user today. We need >>> to provide a mechanism [1] whereby the user can say 'kill' and we'd >>> take that thread off into some special space and whatever sessions(s) >>> it has are automatically rolled-back no matter what. Right now the >>> approach you'd have to take is to set conf/nifi.properties >>> >>> nifi.flowcontroller.autoResumeState=false >>> >>> Then restart and now you can change/tweak/fix/test that processor >>> because all processes will come up stopped. Just remember to change >>> that setting back :-) >>> >>> Now, perhaps add additional copies of that processor to the graph and >>> tweak settings to find what is ultimately holding it up. Tell it to >>> redirect the error stream, try the command manually, etc.. >>> >>> [1] https://issues.apache.org/jira/browse/NIFI-78 >>> >>> Thanks >>> Joe >>> >>> On Mon, Nov 23, 2015 at 8:45 AM, Chris Teoh <[email protected]> wrote: >>>> Hi folks, >>>> >>>> I have a processor stuck with a process (1 displayed in top right corner). >>>> How do I stop this? I have tried restarting NiFi but it comes back with the >>>> same issue. >>>> >>>> Kind Regards >>>> Chris >> >
