SSD reliability in the real world: Google's experience

From ZDNet this morning:
=================
Using data from millions of drive days in Google datacenters, a new paper offers production lifecycle data on SSD reliability. Surprise! SSDs fail differently than disks - and in a dangerous way. Here's what you need to know.
 
SSDs are a new phenomenon in the datacenter. We have theories about how they should perform, but until now, little data. That's just changed.
 
The FAST 2016 paper Flash Reliability in Production: The Expected and the Unexpected, (the paper is not available online until Friday) by Professor Bianca Schroeder of the University of Toronto, and Raghav Lagisetty and Arif Merchant of Google, covers:

• Millions of drive days over 6 years
• 10 different drive models
• 3 different flash types: MLC, eMLC and SLC
• Enterprise and consumer drives

KEY CONCLUSIONS

• Ignore Uncorrectable Bit Error Rate (UBER) specs. A meaningless number.
• Good news: Raw Bit Error Rate (RBER) increases slower than expected from wearout and is not correlated with UBER or other failures.
• High-end SLC drives are no more reliable than MLC drives.
• Bad news: SSDs fail at a lower rate than disks, but UBER rate is higher (see below for what this means).
• SSD age, not usage, affects reliability.
• Bad blocks in new SSDs are common, and drives with a large number of bad blocks are much more likely to lose hundreds of other blocks, most likely due to die or chip failure.
• 30-80 percent of SSDs develop at least one bad block and 2-7 percent develop at least one bad chip in the first four years of deployment.

THE STORAGE BITS TAKE

Two standout conclusions from the study. First, that MLC drives are as reliable as the more costly SLC "enterprise" drives. This mirrors hard drive experience, where consumer SATA drives have been found to be as reliable as expensive SAS and Fibre Channel drives.
 
One of the major reasons that "enterprise" SSDs are more expensive is due to greater over-provisioning. SSDs are over-provisioned for two main reasons: to allow for ample bad block replacement caused by flash wearout; and, to ensure that garbage collection does not cause write slowdowns.
 
The paper's second major conclusion, that age, not use, correlates with increasing error rates, means that over-provisioning for fear of flash wearout is not needed. None of the drives in the study came anywhere near their write limits, even the 3,000 writes specified for the MLC drives.
 
But it isn't all good news. SSD UBER rates are higher than disk rates, which means that backing up SSDs is even more important than it is with disks. The SSD is less likely to fail during its normal life, but more likely to lose data.

Personally I regard SSDs as semi-consumables. At least until a format with a more assured life expectancy is developed. I don't regard HDDs as especially reliable either. Between 2011 and 2013 I had three new Seagate Barracudas fail without warning, the second two being warranty replacements for the first.

Which means using a backup strategy that's effective. Unfortunately there are millions and millions of people with laptops and other devices containing only SSDs who never think of backing data up until they learn the hard way why it's a good idea.

As for the swap file, in performance terms putting that on an SSD can make a huge difference. Just be prepared to replace the drive once every few years.

I spent BIG money on large SSD's a few years back, maybe 2012, when they first came out at the 512 GB size.  I think they were each around $700-$800, and one failed miserably within the first few MONTHS.
 
It freaked me out so bad that a high-end drive touting its reliability and all that would fail so quickly, that I rethought my approach to building my monster machine (32 GB memory and liquid cooling and 2 high end video cards and such).  Anyways, I returned the piece of junk failed drive and got store credit, and got 2 separate 2 TB SATA III 7,200 drives and a 43" HDTV with the store credit from returning the bad SSD.
 
Things have gotten MUCH better with them, but I am still quite careful to backup things of value to me from any SSD drive FAR more often than my 7,200 drives.
 
Bob Bone
 

robert_e_bone
It freaked me out so bad that a high-end drive touting its reliability and all that would fail so quickly, that I rethought my approach to building my monster machine (32 GB memory and liquid cooling and 2 high end video cards and such).  Anyways, I returned the piece of junk failed drive and got store credit, and got 2 separate 2 TB SATA III 7,200 drives and a 43" HDTV with the store credit from returning the bad SSD.

"Moster" is relative. After 2 years, I am still lucky with our recent servers, 256GB of ram, OS on 2 mirrored SSDs and 60TB in RAID6, no bad drives so far