There are several tools/utilities (such as DriveDX) that verify the integrity of spinning drives. But is there a Drive DX for SSDs?

I know one is urged to refrain from erasing SSDs, but at what point should the red flags be raised?

In sum, how does one know when his SSD is failing?

Good question. TechToolPro, Drive Genius, Checkmate, and others will all perform a "surface scan" of an SSD which would report any new bad data blocks, but that leads to some questions…

1. Is a surface scan of an SSD a reliable indicator of its health?
2. Would the additional write operations in a surface scan contribute to shortening the life span of an SSD?
3. What is the normal failure pattern for an SSD, block by block or cataclysmic?

While SMART tests of HDDs are almost notoriously unreliable in predicting future drive failure I ran a SMART test of the SSD and HDD portions of the fusion drive in my Mac mini to compare against one another to see how they compare in what SMART parameters each reports.

• The SSD only reports 10 parameters compared to the HDD's 21).
• Five of the SSDs parameters are also reported by the HDD
• Of the 10 parameters reported by the SSD one (Power-Off Retract Count) seems totally inappropriate to an SSD and two (Total Bad Block and Wear Leveling Count) appear to be potentially useful.
• Interestingly enough of the two HDDs on my system neither report Total Bad Blocks

My initial conclusions are…

• In the mid to short term SSDs will almost totally supplant HDDs for all but bulk data storage.
• Given a tool like TechTool Pro 8 that reports all the SMART parameters there are individual parameters that could potentially warn a knowledgable user of impending drive failure
• Personally I have reservations about the efficacy of a surface scan of an SSD until I get answers to my questions.
• There is no definitive test or tool for checking SSD health or forecasting impending failure
• There is a market for a tool or measurement that can predict SSD failure within a reasonable degree of accuracy.
• I have been known to be wrong more than once in the past and certainly will be wrong in the future

FULL DISCLOSURE TechTool Pro is the only tool I know of that displays all of the SMART parameters reported by the drive. I have no relationship, pecuniary or otherwise, with Micromat (the developer of TechTool Pro) other than that of a satisfied customer.

I have Checkmate enabled and it runs surface scans on my conventional HDDs but omits them on my external SSD and the SSD part of the fusion drive. I can only guess that this is programmed into Checkmate by Micromat. I have not tried to run a surface scan manually with TTP.

DriveDX shows 23 separate "Health Indicators" as well as Overall Health, Overall Performance Rating, and SSD Lifetime Left Indicator (last 3 all in percentages).

But to even get this far, I had to DL & install a special file (via DriveDX) as this type of file is not allowed on the AppStore (and that's where I obtained my DriveDX0.

At least now I have some clues, but much of the data is beyond me. OTOH, one can get automatic alerts when overall health drops below xx percent.

Just to add to my frustration, I recently received an alert that my new SSD's Overall Health Rating was 64.3%. Rebooting my machine magically improved the Health Rating to 100%. Hard not to have confidence in that system...

I bought my first SSD when prices got reasonably low. My intention was to run OS and apps on the SSD and keep data on a spinner, so I bought an Intel 320 80 Gb SSD, in fact two of them to run in RAID, for £270. In two months, they will be five years old and the warranty will expire. Although they still contain my SL system, they have been pretty much retired since I started running Yosemite on a daily basis. The drive still mounts daily, even if it's not accessed.

I used to run TechTool Pro on the Intels to optimize the drives once a month, until common wisdom said it was a bad idea. They've never given any indication of any sort of failure.

I've had a Samsung 840 Pro 256 Gb drive in my MacBook for two months shy of three years. I think this drive came with a five-year warranty. It cost £175. An extra 100 Gb for £100 less two years after my first SSD. Never a moments worry.

My workhorse now is a Samsung 840 EVO 120 Gb that's a month shy of two years old. It's mounted on a PCIe card and is so fast! £65! Still a year to go on its three-year warranty.

I don't worry about them other than to run monthly maintenance. Disk Warrior, Disk Utility. Still, I back them up weekly to an external HHD and an internal HHD.

My data HHDs get a daily backup. All drives will fail at some point, I just seem to retire mine before it happens.

Or, to paraphrase Bobby McFerrin and Meher Baba: Don't worry, be happy and backup.

Checkmate 1.1.5 and previous versions have always performed surface scans of the SSD portion of the fusion drive in my Mac mini. Unlike Checkmate's SMART and other tests, surface scans are only performed once a week and sometimes multiple partial scans of one drive may be logged so you have to scan the logs carefully to find the SSD scan.

After a bit of correspondence with the folks at DriveDx, the bottom line re conflicting SSD reports just minutes apart (such that I understand)…

DriveDx reports drive health according to the data provided by the drive's firmware. [OK, I understand (or at least accept) that.]

In my case, the health indicator #1, "Raw Read Error Rate”, fluctuated, e.g., decreased. And then, for an unknown reason, increased back after my reboot. [I guess that means there is an error in the drive’s firmware or an error in the algorithm that interprets the firmware.]

My (simplistic) conclusion from all this is that SSDs, like spinning drives, may well be victims of a false positive. Ergo, one must bear in mind that SSD health reports only provide “indicators”, not “conclusions”. [not much new there.]

Ah, vagaries, vagaries…

Y'know, Harv, SSDs have been around for more than just a coupl'a years now, and I don't remember ever seeing a post here that had either the possible or definite failure of one of them as its issue, so I'm just gonna take a passive approach, and not get into running software that's more likely to distress me unwarrantedly than realistically and keep my backups current instead.

Those health indicators provided by the drive's firmware are the S.M.A.R.T. (Self Monitoring Analysis and Reporting) we have been discussing and are nothing new. The problem with SMART is the manufacturer chooses which of the 254 parameters specified in the standard its firmware will report and even more importantly what the failure level is. Not surprisingly, manufacturers typically set the failure level so high that it only triggers when the drive has completely failed and therefore provides little or no warning of impending failure.



Because of the manufacturers predilection for setting the failure values very high, SMART is far more likely to yield false negatives (reporting failing drives as good) rather than false positives. However, a knowledgable user looking at the individual parameters can draw reasonable inferences.

If you are interested there is a good Wikipedia article that discusses SMART and has a table describing each of the 254+ possible parameters and indicating which are considered critical but in truth this is all rotating media centric and IMHO there should be other SSD specific parameters but it will probably take years to get any industry consensus on what those should be and even longer to "teach" the firmware to report them.

Many thanks, joemikeb. Your insight is always appreciated, even (perhaps especially), if entails noting my errors.

I can attest to this behavior from personal experience. Internal hard drive on an iMac was making a funny sound. I ran a S.M.A.R.T check and it came back OK (no details shown in whatever software I used at that time). 48 seconds later the drive died. I was amazed at the lack of warning, especially after a check.