Sent date: 05/21/2008
From: Dan Rosanova <(email address - cut out)>
Message:LoadGen doesn't need / shouldn't be run on the BizTalk server you are
testing. This will only slow it down. Do both BizTalk server receive
from this same file location? Are they part of the same BizTalk
Group? LoadGen is certainly the right tool here. Also I would say if
all you're doing is receiving and sending the file, do you need to use
an orchestration at all? You could use a direct subscription on the
receive port to do the same thing with less overhead.
Kind Regards,
-Dan
Sent date: 05/22/2008
From: Shiva
Message:Hi Dan,
I am new to BizTalk Server, please apologise if I have used wrong
terminology.
I have two BizTalk servers, but “LoadGen†is running only in one server.
(Can move LoadGen to another dedicated server)
Yes both servers are in the same group.
Since I am running “LoadGen†in one BizTalk box, only one server will get
message feed from the file drop-location. But I am running the host instance
in both the servers.
I created two hosts
“LoadGenFileSendâ€
“LoadGenFileReceiveâ€
And then created an instance of the above mentioned host in both BizTalk
servers, and then configured the “File†adapter to use “LoadGenFileSend†&
“LoadGenFileReceive†hosts.
My orchestration runs under the default host.
Q1. Is this approach correct? Or Do I need to make both BizTalk boxes to
receive message from the same location.
Q2. If you think I should have a common file-drop location, is it ok to have
a common file-drop location in the network? Would this cause additional
overheads?
I don’t need an orchestration for this but in future we will have some
business process in the orchestration. All I am trying to do is experiment
and learn how to test and monitor the performance counters.
I am monitoring the following performance counter only in one BizTalk box.
\\BIZ2006-01\BizTalk:FILE Receive Adapter(LoadGenFileReceive)\Messages
received
\\BIZ2006-01\BizTalk:FILE Receive Adapter(LoadGenFileReceive)\Messages
received/Sec
\\BIZ2006-01\BizTalk:Message Box:General
Counters(biztalkmsgboxdb:sql2005-01)\MsgBox Parts Cleanup (Purge Jobs)
\\BIZ2006-01\BizTalk:Message Box:General
Counters(biztalkmsgboxdb:sql2005-01)\Spool Size
Q1. Do I need to monitor in both BizTalk server? If so, how do I put the
results together?
Many thanks for your response.
Thanks
S
"Dan Rosanova" wrote:
Show quoted text
>>Are they part of the same BizTalk Group?
>>Do both BizTalk server receive from this same file location?
>>do you need to use an orchestration at all? You could use a
>>direct subscription on the receive port to do the same thing with less overhead
> LoadGen doesn't need / shouldn't be run on the BizTalk server you are
> testing. This will only slow it down. Do both BizTalk server receive
> from this same file location? Are they part of the same BizTalk
> Group? LoadGen is certainly the right tool here. Also I would say if
> all you're doing is receiving and sending the file, do you need to use
> an orchestration at all? You could use a direct subscription on the
> receive port to do the same thing with less overhead.
>
> Kind Regards,
> -Dan
>
Sent date: 05/30/2008
From: Jay Kinker
Message:Hi Siva,
I have been using PerfMon (Windows Performance Monitor) for doing performace
benchmarking.
performace benchmarking and how to improve performace is very well
documented at MSDN.
I collected and blogged those links at:
http://geekswithblogs.net/jaykinker/archive/2006/12/19/101562.aspx
Hope this helps...
Cheers,
Jay Kinker__
http://geekswithblogs.net/jaykinker
----------------------------------------------------------------------------------------
"Shiva" wrote:
Show quoted text
> My objective is to measure “Maximum throughputâ€.
>
> I have written a simple BizTalk solution which consists of an orchestration
> to receive an xml file from a file-drop location using a file adapter and
> then write to another file-drop location.
>
> I then deployed the solution in an environment which consists of two BizTalk
> servers (BZ1 & BZ2) connected to SQL Server 2005. ****Not load balanced*****
>
> Created Host “LoadGenTest†and then Host Instances for BZ1 & BZ2 and then
> configured the File adapter to run under “LoadGenTestâ€.
>
> 3. Installed LoadGen Tool in one BizTalk Servers and fired few messages and
> noticed messages are going through successfully.
>
> 4. During the test run monitored “Message Receiveâ€,â€Message Recived Per
> Secondâ€,
> “Spool table†in one of the BizTalk box.
>
> Question
> My objective is to measure maximum throughput. Please could someone let me
> know whether I’m heading the right direction?
>
> Do I need to monitor in the both BizTalk boxes, at the moment I am
> monitoring only in one box.
>
> BZ1 & BZ2 are not load balanced, would this affect my test results?
>
> If my approach is not correct please could someone let me know perform this
> test?
>
> Thanks
>
> S
>
Email TopXML