Disk I/O for Exchange 2010

Posted on 2011-10-06
Medium Priority
Last Modified: 2012-05-12
I am looking at creating a disaster recovery site for my Exchange 2010 implementation.  The issue that I am running into is picking out a RAID that will be able to handle the I/O from Exchange 2010.  What is the best way to determine how much I/O is being generated by my current system?  I have a couple of different arrays that I am looking at and the only real difference that I am seeing between is cost and I/O capability.  I want to make sure that I am picking the correct system to handle the I/O.  I know in UNIX I would run iostat –xtc 15 > /tmp/io.txt and looking at the information in there, but there are a lot more options in Windows 2008 Enterprise Edition SP2 for disk I/O.  Which counter/counters are going to provide me the best information?
Question by:warewols
  • 2
  • 2
LVL 47

Expert Comment

ID: 36926133
Windows' performance monitor does a good job reporting how much I/O goes on at the LOGICAL level, but if you want to see physical disk I/O, then you need to get something that can drill into your specific raid controller.

Author Comment

ID: 36926294
Any specific counters that you recommend looking at on perfmon?
LVL 47

Accepted Solution

David earned 2000 total points
ID: 36926700
The biggest one you care about is queue depth.  If it is always in the 0-2 range, then bottleneck isn't disk I/O, it is elsewhere.  It will report read and write MB/sec, but those numbers are really a bit fuzzy as they represent what the programs are asking the RAID to provide, and not necessarily accurate enough to make a decision for whatever you are trying to do.  Filesystem settings, load balancing, and other factors can skew the results.

Example,  on RAID1, some controllers do load balancing.  In perfect world, 50% of the reads get serviced by each disk, whatever disk can get to the data fastest, or has it in cache.  Not all controllers do this.   Windows has no idea what is going on.

If you bench, make sure you bench and do data integrity testing and degraded performance.  EVERYTHING is easy and works when hardware is healthy, but what happens when you are doing a file system defrag and you yank out a drive in the middle of it?  Does I/O hang for 1, 2, .. or 60 seconds?  Does performance drop by 10% or by 90%?


Author Comment

ID: 36926991
Performance is fine for the current Exchange 2010 production system.  I am just attempting to spec out a brand new system for a disaster recovery site.  I would like to know what my current I/O is so that I can spec out a disk array for the new system.

Featured Post

Industry Leaders: We Want Your Opinion!

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

How to effectively resolve the number one email related issue received by helpdesks.
Stellar Exchange Toolkit: this 5 in 1 toolkit comes loaded with mega-software tool. Here’s an introduction to tools’ usage and advantages:
This tutorial will walk an individual through the steps necessary to install and configure the Windows Server Backup Utility. Directly connect an external storage device such as a USB drive, or CD\DVD burner: If the device is a USB drive, ensure i…
To add imagery to an HTML email signature, you have two options available to you. You can either add a logo/image by embedding it directly into the signature or hosting it externally and linking to it. The vast majority of email clients display l…
Suggested Courses

839 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question