Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations Chriss Miller on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Disk monitor

Status
Not open for further replies.
Feb 1, 2008
17
US
I want to monitor my disk access on the exchange server cause access to the mailboxes are slow. Can anyone tell me what i should monitor to determine a disk bottlneck.
I can use perfmon, but dont knwo what i should monitor and whats good or bad.
 
What's your current disk configuration, RAID level, partitions, etc.? How much physical memory does the system have? Where's your pagefile for the system? Where are your logs and your DBs located? All this is and probably other things I haven't thought of are going to affect your performance.

You may want to just run the Best Practice Analyzer to see if your system has the recommended setup:
In answer to your question, though, you generally want to start with PhysicalDisk: Avg. Read Queue Length, Avg. Write Queue Length, and %Disk Time and drill down further if any of these are high. I think the first 2 should be 2 or fewer, and the % Disk time should be less than 50%.

You also want to look for excessive paging, though. If you're paging a lot, then you need to add physical memory.
 
I have over 4 gigs of physical memory.
I have OS on the first HD, DB on a raid1, and Logs on a regualr HD.
Page file is on the same drive as the OS.
Thank you, i will run best practice ananlyzer again.i ran the older version and it did warn me on the HD access.
THanks again, i am hopeing to find the bottlenech. I know there is something wrong.
The new version is a little different.
 
Also, should the memory usage be set to system catch or programs. my computer propertys>advanced>Performance option?
 
I find that the queue lengths aren't very telling. I'd focus on the physical disk counters reads/sec, writes/sec, seconds/write, and seconds/read. From these, you'll be able te determine your IO load, read:write ration, and IO response times. In order to make sense of the data, it would be helpful to know how your disks are configured as well.

If you find that your writes are longer than 20ms on average or have spikes over 50ms lasting more than a few seconds, then you IO subsystem is performing poorly.

 
Thank you both for your help.
As a last question, what happens if i loos my log files, since i have them on a regular drive?
 
I checked my physical ram, i have 3.5gigs, i have the swithc in my boot.ini.
\WINDOWS="Windows Server 2003, Standard" /fastdetect /3GB /Userva=3030 /NoExecute=OptOut
I got an warning using the analyzer that my page file is incorrect.
it is set to min 4096 max 4096.
What should i have this set to?
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top