https://bz.apache.org/bugzilla/show_bug.cgi?id=63565

--- Comment #9 from Philippe Mouawad <[email protected]> ---
(In reply to Mohamed Ibrahim from comment #8)
> We are using 5.1 and I am very sure about it. I guess it's also a bug with
> 5.1.

Then please provide the correct jmeter.log
The one you attached has this:
2019-05-24 05:27:56,395 INFO o.a.j.JMeter: Version 3.2 r1790748
2019-05-24 05:27:56,395 INFO o.a.j.JMeter: java.version=1.8.0_05

> 
> 
> And we are using Java sampler for pushing the data to the couchbase. We have
> internally created the plugin for pushing the data using Java SDK couchbase
> client(v2.7.7)
> 
We would need your code as most probably leak is in this part of code.
To confirm it, once you have run the test and reached the limit, run 

jmeter/bin/heapdump.sh 

and share the hprof generated
> 
> 
> Also, We have executed the same round of test by downgrading the Jmeter to
> Jmeter-3.3.4 .
> 
> 
> 1. We didn't observe the memory re-allocation within the heap after 2.30
> hours (out of 4 hours) and its very stable with memory and TPS as well.
> 
> 2. We could still observe the GC time is exponentially growing for that
> version of Jmeter as well. If you want I can share that information as well

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to