Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
Author |
Topic |
karthik_rams@yahoo.com
Starting Member
1 Post |
Posted - 2006-07-19 : 09:39:01
|
Hi All,When we did PT for our application, Full Scans/sec is really huge and it is between 75 to 150. When we did stress testing, it is above 500.Please help me in identifying the problem that causes too much of FT scans.I tried doing this, but I'm not sure whether the approach is correct:Step 1. Create performance log counter against SQLServer:Access Methods\Full Scans/sec (Results as)Time \SQLServer:Access Methods\Full Scans/sec07/19/2006 04:56:03 PM 13.51882392307/19/2006 04:56:13 PM 3.60042370507/19/2006 04:56:23 PM 3007/19/2006 04:56:33 PM 10.099895061Step 2. Create Trace (using SQL Profiler)Start both the Performance log and SQL Profiler at same time.Step 3. Pick a time duration from Performance Log results and identify the queries that got executed during this time frame using Profiler. Step 4. From Profiler results, identify the queries that got executed and evaluate the query plan of these queries. Here I'm not clear with the queries that i pick, because there are different eventclass during this time like RPC:Completed,SQL:BatchStarting, SQL:BatchCompleted. I did analysis on "TextData" only against RPC:Completed and counted on Table/Clustered Index Scans from Query Plan.Thanks & Regards,Karthik R |
|
|
|
|