November 30, 2012 at 4:05 am
Hi,
What's the best way/tool to get SQL Server workload from a client that's complaining about query performance so we can simulate on our server and analyse it?
I read this article http://www.sqlservercentral.com/articles/SQL+Trace/71841/ and it seems a good choice.
It's the same has using SQL Profiler and store the results in a table, right?
Thanks,
Pedro
November 30, 2012 at 4:16 am
You can use the Performance counter (inbuilt winodws Tool) and Performance related DMV's to get sql server workload .
please refer following PPT , it may help you
-----------------------------------------------------------------------------
संकेत कोकणे
November 30, 2012 at 4:23 am
sanket kokane (11/30/2012)
You can use the Performance counter (inbuilt winodws Tool) and Performance related DMV's to get sql server workload .please refer following PPT , it may help you
I probably didn't make myself clear... I don't want the server "workload" as in counters and memory pressures, .... I want the workload, the data the application is sending to the server so I can simulate the same scenario on my server.
Thanks,
Pedro
November 30, 2012 at 4:40 am
PiMané (11/30/2012)
sanket kokane (11/30/2012)
You can use the Performance counter (inbuilt winodws Tool) and Performance related DMV's to get sql server workload .please refer following PPT , it may help you
I probably didn't make myself clear... I don't want the server "workload" as in counters and memory pressures, .... I want the workload, the data the application is sending to the server so I can simulate the same scenario on my server.
Thanks,
Pedro
Then you will need a trace or some sort, client or server side which can be replayed in profiler on a second system.
November 30, 2012 at 7:48 am
Hi
The best way to capture the workload is to run a 'Server Side Trace' using tsql - you can use the SQL Profiler GUI to generate this . To re-run the workload for analysis you can use RML Utilities (Search in Bing for the download details) which has a command line utility called ostress which enables you to replay the trace file(s) generated from your server side trace. Too perform an analysis of trace I would recomend Qure Analyzer (a briliant Free tool from DBSophic).
My usual approach to this task would be:
1. Backup the database you want to monitor
2. Run server side trace (Trace A)
3. Restore the backup to your server
4. Replay the server side trace using ostress (RML utilities) - while running a server side trace (Trace B).
5. Analyse the results in Qure Analyzer compaing the original trace file with the one you collected during the stress test (Trace A vs Trace B)
Cheers
-Matt
Viewing 5 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply