Understanding Disk Performance Metrics in perfmon

We're trying to review the disk activity for a client server to see if it is the cause of recent performance issues and we're currently looking at disk activity on Performance Monitor (built into Windows Server 2008, perfmon.exe).

We're looking at the following items:

Logical Disk \ Disk Transfers/sec
Logical Disk \ Disk Bytes /sec
Logical Disk \ Disk Read Bytes /sec
Logical Disk \ Disk Write Bytes /sec

Logical Disk \ Avg. Disk sec /Transfer
Logical Disk \ Avg. Disk sec /Read
Logical Disk \ Avg. Disk sec /Write


The server is running two SATA 7200 RPM hard drives in a RAID 1 environment.

Can someone let me know what I should expect for some of these metrics. Without anything to compare these numbers to, I don't know if they're high or low.

Thanks
LVL 1
AceofTechsAsked:
Who is Participating?
 
SebastianAbbinantiConnect With a Mentor Commented:
What OS is on the server? What applications is the server running? 75% of memory usage seems normal, but it depends on many factors. Obviously applications would take more memory if more memory is available.

You want to look at Pages/Sec and Page Faults per Sec.

Look at this over time. It does spike so the median makes more sense than the average.

Thanks,
S.
0
 
SebastianAbbinantiCommented:
Disk Queue Length will help you understand if the bottle need is the disk drive.

Thanks,
S.
0
 
AceofTechsAuthor Commented:
I've now gone ahead and added the following counters:

Logical Disk \ Avg. Disk Queue Length
Logical Disk \ Current Disk Queue Length.

Does anyone have any recommendations on what these items should be, and/or what any of the counters I previously listed should be?
0
Has Powershell sent you back into the Stone Age?

If managing Active Directory using Windows Powershell® is making you feel like you stepped back in time, you are not alone.  For nearly 20 years, AD admins around the world have used one tool for day-to-day AD management: Hyena. Discover why.

 
SebastianAbbinantiCommented:
You want the queue length to be less than 1. Its a ratio of requested i/o's to completed i/o's.

If its higher than 1, then it shows that your performance bottleneck is your drives. This is when you would consider faster drives, more drives (RAID), etc.

Thanks,
S.
0
 
AceofTechsAuthor Commented:
My current Disk Queue Length is averaging about 0.03. Anything else you'd suggest I look at in these metrics?
0
 
SebastianAbbinantiCommented:
If you are experiencing disk queue length of .03 while suffering performance issues, it not a disk issue. What type of performance issues are you experiencing? Its it application based? Is it a measurable performance decrease?

Thanks,
S.
0
 
AceofTechsAuthor Commented:
We're having users report that things are extremely slow when accessing the fileshares or accessing data over the fileshares. For example, when users attempt to open folders of files on the server, they said there is a noticeable delay. Additionally, a Quickbooks user has mentioned that when he attempts to reconcile Quickbooks is so slow its virtually unusable across the network.

We've already replaced network components (i.e. the Network switch) without any noticeable improvement.

We've begun looking at CPU, Memory and Disk usage to see if there is anything noticeable there.
0
 
SebastianAbbinantiCommented:
Did anything else change in the environment before these issues began. Do they always occur, or just when the server is under load?

What do your CPU and Memory metrics look like?

Thanks,
S
0
 
AceofTechsAuthor Commented:
Nothing to our knowledge changed at the same time. The closest events were the addition of two new workstations.

In the last few weeks, they said everything has simply gotten very slow. We initially went through and reviewed the antivirus settings, to make sure PCs weren't scanning the network drives; and reviewed a few other things on the systems.

CPU Usage on average is staying under 5%, Memory usage is about 75% of physical memory (roughly 3GB of 4GB total physical memory, not including page file in use)
0
 
andyalderCommented:
Everything points to network, since you've replaced the switch I would look at the DNS settings on the server and workstations. If it's a single server the workstations should have the server as DNS resolver, server should have itself and the DNS service shoud have forward and reverse zones for the local network plus use the ISP as forwarder for resolution.
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.