Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

We have a Windows 2008 R2 server with SQL installed. We are trying to measure if the disk performance is causing any bottle necks. I've found many suggestions regarding which monitors to use, but am having trouble understanding the results. Each counter uses a different scale and I don't know how that could effect the analysis. This server is hosted and storage is provided by a back end SAN. Are there any good articles that give you a table of what is a good or bad result and what scale the counter should be using for those results?

Thanks You!

share|improve this question

2 Answers 2

Here's one on diskperf, its counters and a description.

And here's its parent chapter on the general topic. It's got some really good sub-sections, for example, Resolving Disk Bottlenecks.

Dropping some general terms (or the names of the counter in question) into Google also yields a plethora of official tehcnnet documentation (and unofficial discussions, as well), so you should have no trouble finding what you're looking for on your own, honestly. I think this one's one of my preferred pages when I need quick run down on precisely what that performance counter actually means, though.

share|improve this answer
    
Quickly looking over the last link, I have found similar articles, but I will look the rest over in detail. Do you know if they address the topic of scale for the counters? From what I've read earlier, Perf Mon choose the scale automatically but is not always correct. I just want to make sure I am comparing thing with equal metrics. –  seag33k Feb 7 at 1:02
    
@seag33k scale is always going to depend on your particular environment, and vary widely between environments. Start with the defaults, and adjust as needed. Remember, you can always log the data so you have the raw numbers, which you can re-graph and review later. –  HopelessN00b Feb 7 at 12:14

Windows Performance Monitor can be a good solution. Here are some of the recommended values for disk performance metrics:

Average Disk sec/Read The recommended Average Disk sec/Read value is below 8ms. Maximum peaks during excessive I/O operations can be up to 25 milliseconds, but values constantly higher than 20 milliseconds indicate poor performance

Average Disk sec/Write Values higher than 4 milliseconds indicate poor performance, while the values less than 1 milliseconds indicate the best performance

Average Disk sec/Transfer The recommended value is less than 8 ms

Disk Reads/sec and Disk Writes/sec If the values are low, they indicate slow disk I/O operation processing. There is no specific threshold, as it depends on disk specification and your server configuration. You can monitor these metrics for a while, determine trends and set a baseline. Watch out for peaks

Current Disk Queue Length The value should be less than 2 per disk spindle

Average Disk Queue Length Up to 2 per disk. For disk systems, the recommended value is less than 2 per individual disk drive in an array. For example, in a 6 disk array the Current Disk Queue Length value of 12 means that the queue is 2 per disk

As for the graph ratio, it's best to tweak it manually, as by default, the graph shows the values up to 100. To change the graph ratio:

  1. Right click the graph
  2. Select Properties
  3. Select the Graph tab
  4. Modify the Vertical scale Maximum and Minimal values

enter image description here

Another option that you can use to make the lines easier to read is to change the line ratio. A default line scale is set for each counter, but in case the values are very low or high, you see only a flat line close to the bottom or top of the graph. In that case, modifying a counter line scale is necessary

enter image description here

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.