Dan

I'd add that when it doubt get a thread dump out.  If ever the system
seems to be behavior incorrectly run

bin/nifi.sh dump

wait 30 seconds

bin/nifi.sh dump

And ideally send the full contents of the logs directory in a tar.gz

tar czvf nifilogs.tar.gz logs

Thanks
Joe

On Tue, Jan 30, 2018 at 8:43 AM, dan young <[email protected]> wrote:
> Hello Koji,
>
> I don't see any OOM errors in the logs, I'll keep an eye on the avail.
> thread count.  Thank you.
>
> Regards,
> Dan
>
> On Mon, Jan 29, 2018 at 10:49 PM Koji Kawamura <[email protected]>
> wrote:
>>
>> Hi Dan,
>>
>> If all available Timer Driven Thread are being used (or hang
>> unexpectedly for some reason), then no processor can be scheduled.
>> The number at the left top the NiFi UI under the NiFi logo shows the
>> number of threads currently working.
>> If you see something more than 0, then I'd recommend to take some
>> thread dumps to figure out what running thread is doing.
>>
>> Other than that, I've encountered unexpected behavior with a NiFi
>> cluster if a node encountered OutOfMemory error.
>> The cluster started to behave incorrectly as it can not replicate REST
>> requests among nodes. I'd search any ERR logs in nifi-app.log.
>>
>> Thanks,
>> Koji
>>
>> On Tue, Jan 30, 2018 at 1:10 PM, dan young <[email protected]> wrote:
>> > Hello,
>> >
>> > We're running a secure 3 node 1.4 cluster.  Has anyone seen any
>> > behaviour
>> > where the cluster just stops scheduling the running of flowfiles/tasks?
>> > i.e. cron/timer, just don't run when they're suppose to.  I've tried to
>> > stop
>> > and restart a processor that is say set to run ever 900sec, but nothing
>> > happens.  Then only thing I can do is to cycle through restarting each
>> > node
>> > in the cluster and then we're good for a few days....this is something
>> > that
>> > just started happening and has occurred twice in the last week or
>> > so.....
>> > Anything I should keep an eye out for or look for in the logs?
>> >
>> > Regards,
>> >
>> > Dan

Reply via email to