MonitorMemory ReportingTask in HDF-2.0

[#1]

Hi All,

I am not able to generate any Notification while configuring MonitorMemory RepostingTask in HDF-2.0.( Tested in HDF-1.2 and was working) Does anyone faced similar issue or am i missing something here?

I even tested with 1% threshold values for all pools in the list, but no luck. Any pointers?

Thanks,

Jobin George

[#2 · Accepted solution]

@Jobin George

Are you seeing anything in the nifi-app.log?

I configured this reporting task on a instance of HDF 2.0 and it appears to be working.

Thank you,

Matt

[#3]

Hi @Matt

Except below lines nothing else related to this

2016-11-17 14:56:34,873 INFO [StandardProcessScheduler Thread-3] o.a.n.c.s.TimerDrivenSchedulingAgent MonitorMemory[id=01581011-0307-1aad-d6e9-7701f73fe2a3] started.

Any possible configuration I might be missing?

Thanks,

Jobin

[#4]

hey @Jobin George,

Confirm that G1GC is still set as the garbage collector in your bootstrap.conf.

There really isn't much more to this reporting task.

How long did you leave it running? Do you have an active dataflow that is constantly using JVM memory?

By default this reporting task only runs every 5 minutes. So at the time it runs the threshold would need to be exceeded. Could it be possible that your memory usage is going up and down but overall is low enough to just not trigger? Try running a constant flow of data that is allowed to queue on some connections. Since FlowFile attributes live in memory up to the swap threshold, there would be a constant heap memory usage. See if that does not cause it to trigger.

Thanks,

Matt

[#5]

@Matt Clarke G1GC was set as garbage collector and Issue was fixed in the next version of HDF. This one went unnoticed, accepting the answer.

Thanks,

Jobin George