Hard Drive Reliability

Do any of you still use hard drives? I’m sure some of you do, and I’m also sure plenty of you aren’t sure if you’re using HDDs or SSDs. For some of us, the hardware is just a set of resources with a certain quantity, capacity and performance. While we used to spec, purchase, and assemble database servers in the past, these days may of us just specify cores, RAM, and IOPs or some storage capacity, not worrying about how our resources are actually implemented. I know that I assume SQLServerCentral is using SSDs in our hosted systems, but I honestly don’t know, much less care about the type of drives in use.

I saw Backblaze release their 2018 hard drive stats recently, looking at the reliability of their storage. Since BackBlaze is in the business of storing data, they need to ensure that they have quality storage. I’m a customer, and I appreciate that they’re not just buying and using devices, but also tracking how well the different brands and models work. Since they have over 100,000 drives, this means they can provide some interesting information on reliability.

As our storage needs grow, many of us won’t use SSDs for all our data. Even all the data in our database servers might not be SSD as the cost is high. I’m sure HDDs are well suited for backups, and potentially even still useful in many servers for data storage when workloads might not be as high. Even at home, I like to use SSDs as much as possible, but for some types of storage, especially pictures, audio, and video, HDDs are much more cost effective with their higher density.

The Backblaze stats are interesting in a few ways. Certainly it’s amazing to see how larger and smaller drives perform. It seems that most drives have a very low failure rate, though there is an order of magnitude difference between different models of similar sizes. It doesn’t seem that the 10-14TB drives are worse than the 4TB drives, which makes me think that vendors are doing a great job with quality control in manufacturing. I used to worry about larger sizes being released and having lower reliability, but this doesn’t seem to be the case.

Some drives have very low failures, but these are averages, so I’m not sure they’re necessarily better, but worth keeping an eye on. I also think that since many of these drives fail at < 1% annualized rate, they’re good choices for your data center. Certainly the HGST ones seem to have very low failure rates. I know I may pick up a couple of these for more photo storage as I’m running out. It seems that I never have enough space for all the pictures I take.

Steve Jones

The Voice of the DBA Podcast

Listen to the MP3 Audio ( 3.6MB) podcast or subscribe to the feed at iTunes and Libsyn.

About way0utwest

Editor, SQLServerCentral
This entry was posted in Editorial and tagged . Bookmark the permalink.