jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Adrian Speteanu <asp.ad...@gmail.com>
Subject Re: java.net.BindException returned from http sampler
Date Thu, 30 Jul 2009 17:54:14 GMT
On Thu, Jul 30, 2009 at 7:26 PM, Deepak Shetty<shettyd@gmail.com> wrote:
> hi
> ok, I assume you've tried out the other things recommended by sebb (and i
> think emailed by adrian).
> Do you get any errors on tomcat when you start getting this error on the
> JMeter client
>
> Can you monitor the sockets on both client and server ? are there too many
> open sockets with TIME_WAIT ? There was some tuning parameter related to the
> time that you close the socket and the time that the system actually closes
> which used to affect the server .
> Googling I found this
> http://publib.boulder.ibm.com/infocenter/cmgmt/v8r3m0/index.jsp?topic=/com.ibm.eclient.doc/trs40019.htm
>
> Adrian
> Ive never seen this happen on light loads. And Im sure this problem wouldve
> been reported more often on light loads in jmeter. in which case this seems
> to be OS/machine specific problem?
Basically this is what I assume. But I haven't got the time  yet to
retest the same application / server with same client PC and there is
still a slight possibility that it might be some other problem. I'll
get back on these two threads when I do re-test.

Momentarily, I've checked if the connections started from jmeter are
closed properly ( netstat -an -o -p TCP ).
Both normal connection that ended properly or those that ended
abruptly (by any inventive means) DO CLOSE after a short period of
time.
So it is definitely not a jvm / client-os problem.
However, at the rate of the test, it is possible to run out available
connections for a large number of threads - and that would be a
test-environment issue (but it must be an exceptional case). Thanks
for the link, good to know how to do that.


>
> regards
> deepak
>
> On Wed, Jul 29, 2009 at 11:40 PM, shaoxianyang <ysxsusan@hotmail.com> wrote:
>
>>
>> I then lower the load to 200(threads)x200(loops), no error happens.
>>
>> Then I change the setup to be 200(threads)x400(loops).  The only error is
>> BindException (Address already in use).  I don't see any connection refused
>> exception.
>>
>> Therefore, we still don't know what cause for BindException.  The log shows
>> some BindException takes long to happen, such as 3071(ms).  From how long
>> it
>> takes, it is more of exception that server throw, instead of client side
>> socket problem.
>>
>> Can anyone make comment on this?  At least that BindException.  Now I am
>> totally confused.  Changing to another host does not eliminate this
>> problem.
>>
>> Thanks.
>>
>>
>>
>> Shaoxian
>> --
>> View this message in context:
>> http://www.nabble.com/java.net.BindException-returned-from-http-sampler-tp24716020p24732916.html
>> Sent from the JMeter - User mailing list archive at Nabble.com.
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
>> For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
>>
>>
>

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


Mime
View raw message