JMS issue

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
14 messages Options
Reply | Threaded
Open this post in threaded view
|

JMS issue

Filip.L_Ecluse
Hi all,

I am currently using the latest CVS version of Jmeter and it has JMS
support. There is however one big error in the implementation, if I send
JMS messages using multiple threads it occurs as soon as approximately
100messages using 3 threads that JMeter stops the test and reports that
it cannot find the given correlationID. So in fact a multi threaded test
using JMS in JMeter is not possible yet, will this be solved in the
future and if so can you inform me when?

Thanks
Filip

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JMS issue

sebb-2-2
Please can you file a Bugzilla report, and attach the jmeter.log file
to it, plus a copy of a sample test plan that shows the error
(remember to remove any passwords etc first). Remember to provide
versions of OS, JVM, etc.

Can't make any guarantees as to when it will be fixed, but at least
once there is a Bugzilla report it will be easier to keep track of the
problem.

S.
On 5/18/05, [hidden email] <[hidden email]> wrote:

> Hi all,
>
> I am currently using the latest CVS version of Jmeter and it has JMS
> support. There is however one big error in the implementation, if I send
> JMS messages using multiple threads it occurs as soon as approximately
> 100messages using 3 threads that JMeter stops the test and reports that
> it cannot find the given correlationID. So in fact a multi threaded test
> using JMS in JMeter is not possible yet, will this be solved in the
> future and if so can you inform me when?
>
> Thanks
> Filip
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JMS issue

Peter Lin
In reply to this post by Filip.L_Ecluse
you're using the queue sampler right?

the topic sampler should work correctly. I've tested it for 100K,
200K, 300K, 400K, 500K and 1million messages.

peter lin


On 5/18/05, [hidden email] <[hidden email]> wrote:

> Hi all,
>
> I am currently using the latest CVS version of Jmeter and it has JMS
> support. There is however one big error in the implementation, if I send
> JMS messages using multiple threads it occurs as soon as approximately
> 100messages using 3 threads that JMeter stops the test and reports that
> it cannot find the given correlationID. So in fact a multi threaded test
> using JMS in JMeter is not possible yet, will this be solved in the
> future and if so can you inform me when?
>
> Thanks
> Filip
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JMS issue

Filip.L_Ecluse
In reply to this post by sebb-2-2

The bug is filed in bugzilla as


    Bug #34953 <http://issues.apache.org/bugzilla/show_bug.cgi?id=34953>


Greetings
Filip

>Please can you file a Bugzilla report, and attach the jmeter.log file
>to it, plus a copy of a sample test plan that shows the error
>(remember to remove any passwords etc first). Remember to provide
>versions of OS, JVM, etc.
>
>Can't make any guarantees as to when it will be fixed, but at least
>once there is a Bugzilla report it will be easier to keep track of the
>problem.
>
>S.
>On 5/18/05, [hidden email] <[hidden email]> wrote:
>  
>
>>Hi all,
>>
>>I am currently using the latest CVS version of Jmeter and it has JMS
>>support. There is however one big error in the implementation, if I send
>>JMS messages using multiple threads it occurs as soon as approximately
>>100messages using 3 threads that JMeter stops the test and reports that
>>it cannot find the given correlationID. So in fact a multi threaded test
>>using JMS in JMeter is not possible yet, will this be solved in the
>>future and if so can you inform me when?
>>
>>Thanks
>>Filip
>>
>>---------------------------------------------------------------------
>>To unsubscribe, e-mail: [hidden email]
>>For additional commands, e-mail: [hidden email]
>>
>>
>>    
>>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: [hidden email]
>For additional commands, e-mail: [hidden email]
>
>  
>


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JMS issue

Martijn Blankestijn-2
Hi,

I investigated the bug report and the logfile.

The timeout is set to 2000 (milliseconds). This is the (undocumented ;-(
) default when there is no timeout.

2005/05/18 14:00:12 DEBUG - jmeter.protocol.jms.sampler.JMSSampler:
Timeout             = 2000]

This is the flow of the first and the last message.

FIRST MESSAGE ID
2005/05/18 14:00:17 DEBUG - jmeter.protocol.jms.sampler.MessageAdmin:
put request id ID:414d512051756575654d616e61676572632d8b4220000601
2005/05/18 14:00:19 DEBUG - jmeter.protocol.jms.sampler.MessageAdmin:
get reply for ID:414d512051756575654d616e61676572632d8b4220000601 for
org.apache.jmeter.protocol.jms.sampler.PlaceHolder@1be4713
2005/05/18 14:00:19 DEBUG - jmeter.protocol.jms.sampler.MessageAdmin:
Message ID:414d512051756575654d616e61676572632d8b4220000601 not found.
2005/05/18 14:00:29 DEBUG - jmeter.protocol.jms.sampler.Receiver:
CorrId:ID:414d512051756575654d616e61676572632d8b4220000601
2005/05/18 14:00:29 DEBUG - jmeter.protocol.jms.sampler.MessageAdmin:
put reply id ID:414d512051756575654d616e61676572632d8b4220000601 for
holder null

LAST MESSAGE ID
2005/05/18 14:00:17 DEBUG - jmeter.protocol.jms.sampler.MessageAdmin:
put request id ID:414d512051756575654d616e61676572632d8b4220001401
2005/05/18 14:00:19 DEBUG - jmeter.protocol.jms.sampler.MessageAdmin:
get reply for ID:414d512051756575654d616e61676572632d8b4220001401 for
org.apache.jmeter.protocol.jms.sampler.PlaceHolder@4421ab
2005/05/18 14:00:19 DEBUG - jmeter.protocol.jms.sampler.MessageAdmin:
Message ID:414d512051756575654d616e61676572632d8b4220001401 not found.
2005/05/18 14:00:30 DEBUG - jmeter.protocol.jms.sampler.Receiver:
CorrId:ID:414d512051756575654d616e61676572632d8b4220001401
2005/05/18 14:00:30 DEBUG - jmeter.protocol.jms.sampler.MessageAdmin:
put reply id ID:414d512051756575654d616e61676572632d8b4220001401 for
holder null

    The first line indicates that the message has been send.
2 seconds later
    The second indicates that a notify (arrival of message) or a timeout
has occurred.
    The third line indicates that it is a timeout has occurred (no
message in message administration)
13 seconds later
    The fourth line indicates that a message has been received
    The fifth line administers this message in the message administration

Try to set the timeout to a higher level. With this information it seems
ok that the testcases fail, because an (implicit) timeout of 2 seconds
is used.

Hope this helps.



Greetings,

Martijn

P.S. I will update the bugzilla report as soon as I have my account for
Bugzilla.

[hidden email] wrote:

>
> The bug is filed in bugzilla as
>
>
>    Bug #34953 <http://issues.apache.org/bugzilla/show_bug.cgi?id=34953>
>
>
> Greetings
> Filip
>
>> Please can you file a Bugzilla report, and attach the jmeter.log file
>> to it, plus a copy of a sample test plan that shows the error
>> (remember to remove any passwords etc first). Remember to provide
>> versions of OS, JVM, etc.
>>
>> Can't make any guarantees as to when it will be fixed, but at least
>> once there is a Bugzilla report it will be easier to keep track of the
>> problem.
>>
>> S.
>> On 5/18/05, [hidden email] <[hidden email]> wrote:
>>  
>>
>>> Hi all,
>>>
>>> I am currently using the latest CVS version of Jmeter and it has JMS
>>> support. There is however one big error in the implementation, if I
>>> send
>>> JMS messages using multiple threads it occurs as soon as approximately
>>> 100messages using 3 threads that JMeter stops the test and reports that
>>> it cannot find the given correlationID. So in fact a multi threaded
>>> test
>>> using JMS in JMeter is not possible yet, will this be solved in the
>>> future and if so can you inform me when?
>>>
>>> Thanks
>>> Filip
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: [hidden email]
>>> For additional commands, e-mail: [hidden email]
>>>
>>>
>>>  
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>>  
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JMS issue

sebb-2-2
On 5/18/05, Martijn Blankestijn <[hidden email]> wrote:
> Hi,
>
> I investigated the bug report and the logfile.

Thanks!

> The timeout is set to 2000 (milliseconds). This is the (undocumented ;-(
> ) default when there is no timeout.
>

Might be more intuitive if no timeout meant just that - i.e. wait forever.

S.

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JMS issue

Peter Lin
I'll take a look at the sampler and see if I can change it to read
from jmeter.properties


peter


On 5/18/05, sebb <[hidden email]> wrote:

> On 5/18/05, Martijn Blankestijn <[hidden email]> wrote:
> > Hi,
> >
> > I investigated the bug report and the logfile.
>
> Thanks!
>
> > The timeout is set to 2000 (milliseconds). This is the (undocumented ;-(
> > ) default when there is no timeout.
> >
>
> Might be more intuitive if no timeout meant just that - i.e. wait forever.
>
> S.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JMS issue

Filip.L_Ecluse
In reply to this post by sebb-2-2
sebb wrote:

>On 5/18/05, Martijn Blankestijn <[hidden email]> wrote:
>  
>
>>Hi,
>>
>>I investigated the bug report and the logfile.
>>    
>>
>
>Thanks!
>
>  
>
>>The timeout is set to 2000 (milliseconds). This is the (undocumented ;-(
>>) default when there is no timeout.
>>
>>    
>>
>
>Might be more intuitive if no timeout meant just that - i.e. wait forever.
>
>S.
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: [hidden email]
>For additional commands, e-mail: [hidden email]
>
>  
>
How do I set the timeout value because I tried to put the JMS property
timeout to a very high value and it did not solve anything. Am I doing
it wrong or might there be another issue?

Greetings
Filip

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JMS issue

sebb-2-2
In reply to this post by Peter Lin
Not sure that's necessary - just need to change it so the code calls

resource.wait(timeout)
or
resource.wait()

as appropriate.

If it is considered desirable to provide a default, this could be used
to set the Timeout field on the Gui if not already set. I guess the
default value could be a property.

S.
On 5/19/05, Peter Lin <[hidden email]> wrote:

> I'll take a look at the sampler and see if I can change it to read
> from jmeter.properties
>
>
> peter
>
>
> On 5/18/05, sebb <[hidden email]> wrote:
> > On 5/18/05, Martijn Blankestijn <[hidden email]> wrote:
> > > Hi,
> > >
> > > I investigated the bug report and the logfile.
> >
> > Thanks!
> >
> > > The timeout is set to 2000 (milliseconds). This is the (undocumented ;-(
> > > ) default when there is no timeout.
> > >
> >
> > Might be more intuitive if no timeout meant just that - i.e. wait forever.
> >
> > S.
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [hidden email]
> > For additional commands, e-mail: [hidden email]
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JMS issue

sebb-2-2
In reply to this post by Filip.L_Ecluse
If you set it very high, it might overflow and become negative?

Try enabling debug for

jmeter.protocol.jms.sampler.JMSSampler

and this should show the value actually used.

S.
On 5/19/05, [hidden email] <[hidden email]> wrote:

> sebb wrote:
>
> >On 5/18/05, Martijn Blankestijn <[hidden email]> wrote:
> >
> >
> >>Hi,
> >>
> >>I investigated the bug report and the logfile.
> >>
> >>
> >
> >Thanks!
> >
> >
> >
> >>The timeout is set to 2000 (milliseconds). This is the (undocumented ;-(
> >>) default when there is no timeout.
> >>
> >>
> >>
> >
> >Might be more intuitive if no timeout meant just that - i.e. wait forever.
> >
> >S.
> >
> >---------------------------------------------------------------------
> >To unsubscribe, e-mail: [hidden email]
> >For additional commands, e-mail: [hidden email]
> >
> >
> >
> How do I set the timeout value because I tried to put the JMS property
> timeout to a very high value and it did not solve anything. Am I doing
> it wrong or might there be another issue?
>
> Greetings
> Filip
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JMS issue

Martijn Blankestijn-2
In reply to this post by Filip.L_Ecluse
I thought the timeout could be set in the userinterface of the sampler. Not as a JMS property, but as a regular property.

Please check the screen, because in the version that I want to commit tonight the timeout version attribute is on the JMS Sampler page (including the milliseconds remark)


Greetings,

Martijn

> sebb wrote:
>
>> On 5/18/05, Martijn Blankestijn <[hidden email]> wrote:
>>
>>
>>
>>> Hi,
>>>
>>>
>>> I investigated the bug report and the logfile.
>>>
>>>
>>>
>>
>> Thanks!
>>
>>
>>
>>
>>> The timeout is set to 2000 (milliseconds). This is the (undocumented
>>> ;-(
>>> ) default when there is no timeout.
>>>
>>>
>>>
>>>
>>
>> Might be more intuitive if no timeout meant just that - i.e. wait
>> forever.
>>
>> S.
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>>
>>
>>
> How do I set the timeout value because I tried to put the JMS property
> timeout to a very high value and it did not solve anything. Am I doing it
> wrong or might there be another issue?
>
> Greetings
> Filip
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>
>


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JMS issue

Filip.L_Ecluse
In reply to this post by sebb-2-2
sebb wrote:

>If you set it very high, it might overflow and become negative?
>
>Try enabling debug for
>
>jmeter.protocol.jms.sampler.JMSSampler
>
>and this should show the value actually used.
>
>S.
>On 5/19/05, [hidden email] <[hidden email]> wrote:
>  
>
>>sebb wrote:
>>
>>    
>>
>>>On 5/18/05, Martijn Blankestijn <[hidden email]> wrote:
>>>
>>>
>>>      
>>>
>>>>Hi,
>>>>
>>>>I investigated the bug report and the logfile.
>>>>
>>>>
>>>>        
>>>>
>>>Thanks!
>>>
>>>
>>>
>>>      
>>>
>>>>The timeout is set to 2000 (milliseconds). This is the (undocumented ;-(
>>>>) default when there is no timeout.
>>>>
>>>>
>>>>
>>>>        
>>>>
>>>Might be more intuitive if no timeout meant just that - i.e. wait forever.
>>>
>>>S.
>>>
>>>---------------------------------------------------------------------
>>>To unsubscribe, e-mail: [hidden email]
>>>For additional commands, e-mail: [hidden email]
>>>
>>>
>>>
>>>      
>>>
>>How do I set the timeout value because I tried to put the JMS property
>>timeout to a very high value and it did not solve anything. Am I doing
>>it wrong or might there be another issue?
>>
>>Greetings
>>Filip
>>
>>---------------------------------------------------------------------
>>To unsubscribe, e-mail: [hidden email]
>>For additional commands, e-mail: [hidden email]
>>
>>
>>    
>>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: [hidden email]
>For additional commands, e-mail: [hidden email]
>
>  
>
so if i understand it correctly I have to change it in the source code
of Jmeter to put the default timeout to another value?

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JMS issue

sebb-2-2
On 5/19/05, [hidden email] <[hidden email]> wrote:

> sebb wrote:
>
> >If you set it very high, it might overflow and become negative?
> >
> >Try enabling debug for
> >
> >jmeter.protocol.jms.sampler.JMSSampler
> >
> >and this should show the value actually used.
> >
> >S.
> >On 5/19/05, [hidden email] <[hidden email]> wrote:
> >
> >
> >>sebb wrote:
> >>
> >>
> >>
> >>>On 5/18/05, Martijn Blankestijn <[hidden email]> wrote:
> >>>
> >>>
> >>>
> >>>
> >>>>Hi,
> >>>>
> >>>>I investigated the bug report and the logfile.
> >>>>
> >>>>
> >>>>
> >>>>
> >>>Thanks!
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>>The timeout is set to 2000 (milliseconds). This is the (undocumented ;-(
> >>>>) default when there is no timeout.
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>Might be more intuitive if no timeout meant just that - i.e. wait forever.
> >>>
> >>>S.
> >>>
> >>>---------------------------------------------------------------------
> >>>To unsubscribe, e-mail: [hidden email]
> >>>For additional commands, e-mail: [hidden email]
> >>>
> >>>
> >>>
> >>>
> >>>
> >>How do I set the timeout value because I tried to put the JMS property
> >>timeout to a very high value and it did not solve anything. Am I doing
> >>it wrong or might there be another issue?
> >>
> >>Greetings
> >>Filip
> >>
> >>---------------------------------------------------------------------
> >>To unsubscribe, e-mail: [hidden email]
> >>For additional commands, e-mail: [hidden email]
> >>
> >>
> >>
> >>
> >
> >---------------------------------------------------------------------
> >To unsubscribe, e-mail: [hidden email]
> >For additional commands, e-mail: [hidden email]
> >
> >
> >
> so if i understand it correctly I have to change it in the source code
> of Jmeter to put the default timeout to another value?

I checked the code last night, and the default value is hard-coded in
the source.
The default only applies if the user enters a value of <1, i.e. <=0.

As far as I could see, if any other value is entered on the screen,
the default will not apply, and the value will be taken from the GUI
field.

Changing the default would involve changing the source - however that
should not be needed.

S.

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JMS issue

Peter Lin
setting it in the sampler is definitely the best solution.

peter

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]