keep-alive causing threads to hang on shutdown

["Derek Gill (dergill)" <de...@cisco.com>]

Hi,
 
I am using the new functionality provided in 2.3.1 where SOAP/XML
sampler now supports the keep-alive option.  However I seem to be having
a problem where Jmeter hangs on stopping any threads which are running
with the connection set to keep-alive.  So the typical instance I have
is 1000 threads running with keep-alive.  The test will run fine, until
the test is due to stop.  At which point the threads will start to stop,
but after a couple of threads stopping the shutdown hangs.  The only way
I can get the threads to stop is to either kill Jmeter or restart the
apache server which my requests were directed to.  Is this a known issue
?  The server is configured with Keep Alive ON.  If I run the same test
with keep alive turned off on all my Jmeter requests, the shutdown
completes successfully.  I also tried this by backing off to 100
threads, but the same results.
 
Derek

 

Re: keep-alive causing threads to hang on shutdown

[sebb <se...@gmail.com>]

On 06/03/2008, Derek Gill (dergill) <de...@cisco.com> wrote:
> Hi,
>
>  I am using the new functionality provided in 2.3.1 where SOAP/XML
>  sampler now supports the keep-alive option.  However I seem to be having
>  a problem where Jmeter hangs on stopping any threads which are running
>  with the connection set to keep-alive.  So the typical instance I have
>  is 1000 threads running with keep-alive.  The test will run fine, until
>  the test is due to stop.  At which point the threads will start to stop,
>  but after a couple of threads stopping the shutdown hangs.  The only way
>  I can get the threads to stop is to either kill Jmeter or restart the
>  apache server which my requests were directed to.  Is this a known issue

I don't think this has been reported before.

>  ?  The server is configured with Keep Alive ON.  If I run the same test
>  with keep alive turned off on all my Jmeter requests, the shutdown
>  completes successfully.  I also tried this by backing off to 100
>  threads, but the same results.

And does it happen with fewer threads? Just curious.

It may be that your test scenario allows you to remove Keep-Alive from
the last sampler; if so it might be worth trying that.

>
>  Derek
>
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org