Hi Deepak,
Thank you for your response.
After I submitted my question, I started analyzing the logs, and as you
suspected, the problem was java out of heap memory
I increased my JVM_ARG to -Xms=4096- Xmx=8000, and was able to get more files
download to complete. But still, I only got 20 out of 50 total download that I
had hoped for.
Is there some timer or timing arrangement with my sampler arrangement under the
Thread Group that I can play with to overcome this heap memory issue?
How about instead of a single Thread Group with 5 HTTPS samplers under it (the
first 4 leading up to the 5th which is the actual file downloading request),
right now I run this Thread Group with User = 50, Ramp = 100 sec, cycle = 1 ...
instead of that, what if I make them into 5 separate Thread Group, each thread
group with User = 1, Ramp=1, cycle =1. Will I be able to launch all 50
Thread Groups sequentially but quickly, one after another? For each Thread
Group, like I said, 5 samplers are needed to cause the download to begin.
All 5 executes quickly, but the 5th request will take time to finish (for 310
MB to download). But as soon as Group 1-5 is done (but downloading now in
progress), will Group 2-1 ~ 2-5 start to execute?
Will doing the above help to release the heap memory? In other words, as
Thread 1-5 finish dumping data to local file, that memory will be freed for
following succession of Thread 2-5, 3-5, 4-5, 5-5, 6-5, etc to use while they
also finish up dumping data to file?
Thank you
Charles
2015/11/07 18:43:25 DEBUG - jmeter.threads.ThreadGroup: Ending thread Thread
Group 1-1
2015/11/07 18:43:30 ERROR - jmeter.threads.JMeterThread: Test failed!
java.lang.OutOfMemoryError: Java heap space
2015/11/07 18:43:31 ERROR - jmeter.threads.JMeterThread: Test failed!
java.lang.OutOfMemoryError: Java heap space
2015/11/07 18:43:31 INFO - jmeter.threads.JMeterThread: Thread finished:
Thread Group 1-40
2015/11/07 18:43:30 ERROR - jmeter.threads.JMeterThread: Test failed!
java.lang.OutOfMemoryError: Java heap space
at java.io.ByteArrayOutputStream.<init>(Unknown Source)
at
org.apache.jmeter.protocol.http.sampler.HTTPSamplerBase.readResponse(HTTPSamplerBase.java:1741)
at
org.apache.jmeter.protocol.http.sampler.HTTPAbstractImpl.readResponse(HTTPAbstractImpl.java:412)
at
org.apache.jmeter.protocol.http.sampler.HTTPHC4Impl.sample(HTTPHC4Impl.java:352)
at
org.apache.jmeter.protocol.http.sampler.HTTPSamplerProxy.sample(HTTPSamplerProxy.java:74)
at
org.apache.jmeter.protocol.http.sampler.HTTPSamplerBase.sample(HTTPSamplerBase.java:1146)
at
org.apache.jmeter.protocol.http.sampler.HTTPSamplerBase.sample(HTTPSamplerBase.java:1135)
at
org.apache.jmeter.threads.JMeterThread.process_sampler(JMeterThread.java:434)
at org.apache.jmeter.threads.JMeterThread.run(JMeterThread.java:261)
at java.lang.Thread.run(Unknown Source)
-----Original Message-----
From: Deepak Shetty [mailto:[email protected]]
Sent: Saturday, November 07, 2015 8:11 PM
To: JMeter Users List
Subject: Re: not all 50 http samplers to download file are executed
Did you check jmeter.log for errors like out of memory ?
On Nov 7, 2015 1:11 PM, "Charles Lin (charlin)" <[email protected]> wrote:
>
>
> I have a script to download a file. It works fine when the file size
> is below certain size. All 50 samplers are executed.
>
>
>
> But when file size is large – in this particular test, 330 MB file –
> when the test plan execution is completed, I see that only 10 or so
> out of total
> 50 samplers show up in the View Result Tree. Does that mean 40
> download did not occur?
>
>
>
> At first I thought 50 files were not saved onto my test machine when
> test plan finished was due to the fact that test plan exited before all the
> data
> could be written to disk files. However, when looking at the View Results
> Tree and only seeing 10+ of the total 50 samplers, it now looks like many
> of the samplers were not even executed. But why? Thread group is set up
> for 50 user, ramp period = 2 (I was mistaken here … thinking 2 sec
> between each user but it’s actually 2 sec for total of 50 user
> requests, so I have readjusted it to 50 seconds duration to achieve 1
> sec/user rate, will
> retest to see …). I put a Constant Timer which appears to be delay of
> submit between each # sampler, so #1 submitted (and timer start counting 60
> sec) and during this 60 sec, all 50 #1 samplers are submitted. Then #75
> gets submitted (all 50 of them), and 50 sec later, #17 samplers (50 times)
> are submitted, etc. So that means when we get to #24 samplers, I should
> expect to see all 50 of the #24 samplers to be there.
>
>
>
> So why did I see only 10+ #24 samplers, and end up with only 10+
> actual downloaded files instead of all 50?
>
>
>
> Thank you
>
> Charles
>
>
>
>
>
>
>
>
