How to Test Drive Spead Tom's Hardware

How to Examination Your Storage with CrystalDiskMark

Last Updated December v, 2021

I've written nigh how to test your SAN'due south performance with SQLIO, only I'll be honest with you lot: that'south the hard mode.  It takes knowledge and fourth dimension, and you only take i of those.  (I'll be charitable and non tell you lot which one.) Instead, let's get seat-of-the-pants numbers for your storage.

Go to the CrystalDiskMark download page, but PAY CLOSE ATTENTION. This is tricky. In that location are multiple download links – you want CRYSTALDISKMARK, not CrystalDiskInfo, and yous want the Standard Edition, Nix Version. The nix version doesn't require an installer, which is cool considering I'm not a fan of installing things on production servers. The zip version can be just saved to a network share, then anybody can run information technology from there.

Later extracting the zip file'due south contents, run DiskMark64.exe on an idle server or desktop first (not your live SQL Server, because information technology'll dull things down while it runs.)  It'll look like this:

Across the top, set up the first iii dropdowns to:

• 1 – the number of exam passes yous want to run.  If y'all want a fast seat-of-the-pants guess, do one, but keep in mind it can be wildly variant betwixt passes if something else happens to be going on in the SAN.
• 1GiB – the test file size.  If you're nether the gun, do a quick 1GiB test, but for real become-alive prep, I like using 32GB to reduce the chances that I'thou but striking enshroud and getting artificially fast numbers.  Smaller test file sizes may await fast merely don't really reflect how a large database volition work. Only know that the bigger the test file, the longer it takes to generate.
• Thou: – the drive letter to examination.  Proceed an eye on the free space in that location – you don't want to create a test file that tin can run your server out of drive space. You want to test where your information, log, and TempDB files live, and for fun, also examination the C drive and your desktop or laptop for comparison.

After making your choices, click the All push.  While it runs, here'south an caption of each row's results:

• SEQ1M Q8T1 – lots of long, sequential operations. For SQL Server, this is somewhat akin to doing backups or doing table scans of perfectly defragmented data, like a data warehouse.
• SEQ1M Q1T1 – ignore, SQL Server doesn't work like this.
• RND4K Q32T16 – random tiny operations, but many done at a time.  This is somewhat akin to an active OLTP server, or a TempDB drive.
• RND4K Q1T1 – Ignore, SQL Server doesn't work like this.

The more astute readers (and by that I hateful y'all, you good-looking charmer) will notice that 4K operations don't actually measure out SQL Server's IO.  SQL Server stores stuff on disk in 8K pages, and zooming out a little, groups of viii 8KB pages (64K extents).  We're not looking to become an exact representation of SQL Server'due south IO patterns here – we're just trying to become a fast, one-button-click-easy measurement of how storage performs.  Commonly I find that during the first round of storage tests, it's not performing well catamenia – and it doesn't make sense to bring SQL Server into the game but all the same.

Sample CrystalDiskMark Results

Hither's a sample set of results from a 335GB general purpose SSD book in Amazon EBS:

Note how you can type in the bottom box of CrystalDiskMark's results – see how I typed Amazon General Purpose SSD? That's great for making notes that will be visible in the screen shots to aid you determine which test results came from which machine.

And here is a set from an imperceptible SSD locally attached to that same EC2 VM:

Discover how the ephemeral SSD is 10x-30x faster on reads, and 4x-18x faster on writes? Non to mention that the ephemeral drive is completelyfree with your VM. You can see why people are tempted to store databases on there, merely that's a word for another mean solar day.

You tin get IOPs, latency, and throughput numbers from CrystalDiskMark too past clicking File, Save Text, so become into a text editor and open up the results. The text version of the results has more details:

Then what'southward a adept or bad number?  If your server boots from a mirrored pair of local drives, and stores its SQL Server data somewhere else (like on a larger array or on a SAN), and so exam the local mirrored pair also.  Compare the numbers for where y'all're storing the valuable, high-performance data to where you're storing the OS, and you might be surprised.  Often I find that the Os's drives perform even meliorate because nosotros only oasis't configured and tuned our storage.

Keep these original CrystalDiskMark screenshots in a shared folder for the group to access, and then challenge anybody involved to exercise better.  Simple tuning techniques similar tweaking the read/write bias on the RAID controller's cache, right-sizing the NTFS resource allotment units, and working with different stripe sizes can usually yield double the storage functioning without spending a dime.

Share this post