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 |
|
DavidChel
Constraint Violating Yak Guru
474 Posts |
Posted - 2008-07-09 : 16:10:29
|
So, I read this article: http://www.sqlteam.com/article/sql-server-2000-performance-tuning-toolsAnd I'm concerned about this: quote: To help identify potential I/O performance problems, the Windows 2000 PhysicalDisk Object: Avg. Disk Queue Length counter is critical to monitor. If the Avg. Disk Queue Length exceeds 2 for continuous periods (over 10 minutes or so) for each disk drive in an array, then you probably have an I/O bottleneck for that array. Ways to remove this bottleneck include: adding drives to an array (if you can), getting faster drives, adding cache memory to the controller card (if you can), using a different version of RAID, getting a faster controller, or reducing the workload on SQL Server.
Exceeds 2 what? The Performance Monitor can have any Y axis value that is required. When I run what I think is the offending process, the Disk Queue Length goes off the scale even at 100 and stays there constantly for the 6 minutes or so the process takes. When I expand the graph to 1000 it even peaks there. So, how does one read this and know if there truly is a problem? |
|
|
Michael Valentine Jones
Yak DBA Kernel (pronounced Colonel)
7020 Posts |
Posted - 2008-07-09 : 16:27:55
|
| Perfmon shows the actual values at the bottom of the Permon display (avg, max, last).The article is saying that if your average disk queue length is greater than 2 times the number of physical disks in the array, it is a problem. For example, if the disk array has 5 drives, an average disk queue length greater than 10 over a long time period would be a problem.CODO ERGO SUM |
 |
|
|
DavidChel
Constraint Violating Yak Guru
474 Posts |
Posted - 2008-07-09 : 16:35:17
|
| I'm sorry, I must've had a brain fart when I read that. Thanks MVJ. |
 |
|
|
|
|
|
|
|