- This topic has 2 replies, 1 voice, and was last updated 8 years, 4 months ago by
.
-
Posts
-
-
Hey,
we have a custom flat file disasembler that dis-asembles inbound flat file.
the inbound file arrives in a receive folder.
The send port subscribes to the Receive port and sends the file to the outbound folder.the issue is disassembly seems to take a long time in our prod env. I have monitored the prod Biztalk node and it is looking good.
I have seen CPU shoot up on Sql node.Interestingly, when we test on our dev box, it seems to run much much faster,taking only a fraction of time.
The only diff between our dev box and the Sql server node in prod is AWE is enabled.
Could that be the problem? any tips on diagnosing this is appreciated.
Regs,
-
Anon,
The AWE flag enables Addressing Windows Extensions which allows applications such as SQL Server to access more than 4GB of physical memory. If you have more the 4Gb of memory installed in your SQL Server box I would suggest that this flag is enabled along with the /3GB and/or /PAE switches (see http://www.modhul.com/2007/11/10/optimising-system-memory-for-sql-server-part-i/ for more information).
However, I would suggest that if you are experiencing high CPU loads, I would monitor tools such as SQL Profiler to determine whether you are experiencing much higher traffic in your prod environment than you are in dev. If you are, you’re into classic scale up / scale out territory.
Other possible causes:
- Better hardware / CPU / CPU cores in dev?
- Better SQL Server data and log file separation in dev (i.e. high I/O is handled more gracefully)
Diagnosing SQL Server issues is always a long and complicated job, please let us know how you get on.
Rgds, Nick.
-
Hello nick,
thanx for your guidance.
in our prod set up we have nearly 8 biztlak nodes each node have around 6 host instances connecting to a
clustered sql box.The sql box is quad core and having 16 gb ram. we have alloted 4 gb(custom size)
for VM paging file. the sql server is awe enabled and the box was booted with /pae option.coming back to the problem, we detected very HIgh PF usage. In the task manager, it was consistenly on over 14
gb.When we restarted Sql service,the PF usage came back to normal range around 1 gb and processing was very quick.
But everytime we process a
large inbound file via Biztalk, the PF usage on SQl seems to rise. After the processing, it does not seem
to come down.However CPU/PF usage on the 8 biztlak nodes are normal.any ideas?
-
- The forum ‘BizTalk 2004 – BizTalk 2010’ is closed to new topics and replies.