I received a warning from TechTool Protection that one of my clones is failing. It is a 480 GB OWC Mercury Electra. The warranty expired a year ago, unfortunately.

SMART reports that the Relocated Sectors Count is 5 and that the Extended Health is in the middle between Pass and Fail. All other parameters are at the optimum level. Fortunately, I have two more clones (both OWC SSDs) and they are fine.

My question: If I reformat the drive and make a new, pristine clone, will that suffice? Or, should I bite the bullet and simply replace it?

Regardless, whether it is an SSD or HD reformatting can have no positive effect on a drive's SMART values and could possibly have a negative effect. It certainly sounds as if the SSD may be failing or has already failed. 😨 I would not trust anything to that drive I was not willing to lose.

By-the-way, the values reported by your drive reveal that it is an older drive. The newer faster NVMe drives have their own unique SMART values that are different than we are accustomed to. Micromat has a new product out called Drive Scope that is specifically for reporting SMART values whether on an HD, an SSD, or an NVMe. It may end up replacing Techtool Pro in my toolkit.

You mentioned Drive Scope in another thread, and after looking at it I wrote it off as an awfully pricey - $50 - one trick pony, but a bit of digging revealed that my old TTP 6 serial # enabled me to buy it for ONLY $30, so I reclassified it from written-off to back-burnered, and now Jon's mention of a drive having gone south - the first such mention in a considerable period of time - has got me scratching my head.

While sorting through the dandruff flakes, though, I found DriveDx for $20, but 10 bucks is 10 bucks, and unlike Drive Scope, it offers a free trial, which reported the following partial results (...only an extract, but it seems to be something of a summary. I'll post the entire report if you're interested.):

=== DRIVE HEALTH INDICATORS ===
ID | NAME | TYPE | UPDATE | RAW VALUE | VALUE | THRESHOLD | WORST | LAST MODIFIED | STATUS
1 Raw Read Error Rate Life-span online 0x0 200 0 200 - 100% OK
5 Retired Block Count Pre-fail online 0 100 0 100 - 100% OK
9 Power On Hours Life-span online 28,018 94 0 94 - 94.0% OK
12 Power Cycle Count Life-span online 951 99 0 99 - 99.0% OK
169 Total Bad Block Count Pre-fail online 0x3390A800780 242 10 242 - 100% OK
173 Wear Leveling Count Life-span online 0x800900048 196 100 196 - 96.0% OK
174 Host Reads MiB Life-span online 17,522,525 (16.7 TB) 99 0 99 - 99.0% OK
175 Host Writes MiB Life-span online 13,761,726 (13.1 TB) 99 0 99 - 99.0% OK
192 Unsafe Shutdown Count Life-span online 98 99 0 99 - 99.0% OK
194 Temperature (Celsius) Life-span online 35 65 30 19 - 50.0% OK
197 Current Pending Block Count Life-span online 0 100 0 100 - 100% OK
199 UDMA CRC Error Count Life-span online 0 200 0 200 - 100% OK

Does that ring any kind of bell with you?

I also found SSD reliability in the real world: Google's experience | ZDNet, which is comprehensible, at least to a degree, even by me.

The "Drive Health Indicators" are noting more and nothing less than the specific SMART parameters the manufacturer of the drive chose to have the drive report. The ID number and the parameter name are specified in the SMART standards and as you can see there are at least 199 possible parameters in the standard. Every manufacturer has the latitude to choose which parameters to collect and report and more importantly what the failure level is for each parameter. The problem with that is manufacturers tend to set the limits such that by the time SMART reports a problem the drive has already failed.

With the advent of SSDs that connect to the system and report SMART values through the SATA bus there were additional parameters added that gave a somewhat better picture if SSD health. Newer NVMe (Non-Volatile Memory express) drives appear to have settled on a standardized set of parameters that to seem to me to tell the tale pretty succinctly..

NOTE: Numbered categories are mine and not part of the standard.

1. Drive Identification
   • PCI Vendor ID
   • Subsystem Vendor ID
   • Controller ID
   • Namespace ID
   • NVMe Interface Revision
2. Statistics
   • Unsafe Shutdowns
   • Controller Busy Times
   • Max Data Transfer Size
3. Early warning data
   • % of available spares
   • The available spare threshold — set by vendor
   • % used
   • % available
4. Ultimate lifetime indicators
   • Data Units Read
   • Data Units Written
   • Host read commands
   • Host Write commands
5. Has the drive failed?
   • Available spare space below threshold: Y/N
   • Over Temperature threshold: Y/N
   • NVM Subsystem Reliability degraded: Y/N
   • Media in Read-Only Mode: Y/N
   • Volatile Memory Backup Device Failure: Y/N

As I see it if any of the last five items is "Y" the drive should be replaced ASAP if not sooner.

Looks like I'll be replacing the drive. Checkmate just notified me that 7 files are corrupted. Aside from OWC, do you have any recommendations for SSD brands?

I don't know of any studies or statistics on SSD reliability — mainly because they are not used in big data centers as HDs are — but Crucial comes immediately to mind as a store and a brand.

If I were shopping for brands I would start with (in order)

1. OWC
2. Crucial
3. Samsung

But that is just me and I cannot scientifically justify that list or that order other than by my perceived reputation.

I think that means that the sum total of of DriveDx's report is nothing more than a meaninglessly embellished (I'll spare you from the smoke and mirrors.) Disk Utility S.M.A.R.T. status Pass/Fail report, in which case, paying $20 for it is a colossal waste of money?

Not quite! Disk Utility's, and other apps that report only the consolidated Pass/Fail are limited to reporting the manufacturer's smoke & mirrors as you called it. I have no idea what DriveDX reports, but apps such as Micromat's TechTool Pro and Drive scope that report the ACTUAL attributes and their values provide information that can enable a reasonably informed user to draw their own conclusion.

If I am considering a rotating rust (HD) drive I would look first at…

• 1. Raw Read Error Rate
• 3. Spin Up Time
• 6. Reallocated Sector Count
• 7. Seek error Rate
• 10. Spin Retry Count

as leading indicators of impending failure. A non-zero value for any of these is an early warning sign, and the closer it is to the manufacturer's threshold setting the sooner it will need to be replaced. If it is over the Manufacturer's threshold the drive has failed. Most of the other attributes are indicators of age.

I just checked to see what SMART attributes are reported for an OWC SATA attached SSD and there were NO pre-fail attributes reported 😡

If I am looking at a NVMe SSD then the leading indicators are:

• Available Spares
• Percentage used
• Percentage Remaining

Once again if any of the maufacturer's thresholds are exceeded, consider the drive not just toast but burnt toast.

NOTE Although both Micromat's TechTool Pro 11 and Drive Scope 1.1.2 report the full set of SMART attributes reported by a drive, in the course of this thread I have come to realize Drive Scope provides some additional information and even a test that IMHO substantially enhances the my ability to analyze what SMART is really telling me:

• SMART logs
• Self Test results
• the ability to trigger both short and long offline self tests
• complete analysis of drive capabilities
• additional information about SMART attributes (pre-fail, old-age)

Given it has been years since I found it necessary to resort to an industrial strength utility for volume repairs, and the limitations of third party apps vis-a-vis. APFS formatted SSDs I would have to say at this point Drive Scope is the more useful app.

I'm not sure what "Not quite!" means, but I saved the DriveDX report because I thought it might come to this, particularly as respects "smoke and mirrors," which are not what you took them for, rather they're specifications and other irrelevant info apparently meant to flesh out the report to make it look more substantial and informative than it really is.

### SYSTEM INFORMATION ###
Report Timestamp : August 13, 2019 12:49:40 PM EDT
Report Timestamp (ISO 8601 format) : 2019-08-13T12:49:40

Application Name : DriveDx
Application Version : 1.8.2.610
Application SubBuild : 1
Application Edition : Standalone
Application Website : https://binaryfruit.com/drivedx
DriveDx Knowledge Base Revision : 9/9

Computer Name : Artie’s MacBook Pro
Host Name : Artie's-MacBook-Pro-2
Computer Model : MacBookPro11,4

OS Boot Time : 2019-08-08T02:25:24
Time Since Boot : 5 days 10h 24m 16s
OS Name : macOS
OS Version : 10.14.6
OS Build : 18G87
OS Kernel Version : Darwin 18.7.0

SAT SMART Driver Version : N/A
ATA Command Support Tolerance : verypermissive
N of drives with S.M.A.R.T support : 1



### DRIVE 1 OF 1 ###
Last Checked : August 13, 2019 12:48:46 PM EDT
Last Checked (ISO 8601 format) : 2019-08-13T12:48:46

Advanced SMART Status : OK
Overall Health Rating : GOOD 100%
SSD Lifetime Left Indicator : GOOD 96.0%
Issues found : 0

Serial Number : S29ANYAH406432
WWN Id : 5 002538 900000000
Volumes : HD, HD2, HDx, HDy
Device Path : /dev/disk0
Total Capacity : 500.3 GB (500,277,790,720 Bytes)
Model Family : Apple (Samsung-based) SSDs
Model : APPLE SSD SM0512G
Firmware Version : BXW1SA0Q
Drive Type : SSD

Power On Time : 28,028 hours (38 months 27 days 20 hours)
Power Cycles Count : 951
Current Power Cycle Time : 130.4 hours



=== DEVICE CAPABILITIES ===
S.M.A.R.T. support enabled : yes
DriveDx Active Diagnostic Config : Apple (Samsung-based) g-series SSDs config [ssd.apple.samsung.g]
Sector Logical Size : 512
Sector Physical Size : 4096
Physical Interconnect : PCI
Logical Protocol : SATA
Removable : no
Ejectable : no
ATA Version : ATA8-ACS T13/1699-D revision 4c
SATA Version : SATA 3.0, 6.0 Gb/s (current: 6.0 Gb/s)
I/O Path : IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEG0@1/IOPP/SSD0@0/AppleAHCI/PRT0@0/IOAHCIDevice@0/AppleAHCIDiskDriver/IOAHCIBlockStorageDevice
Attributes Data Structure Revision : 1
SMART Command Transport (SCT) flags : 0x0
SCT Status supported : no
SCT Feature Control supported : no
SCT Data Table supported : no
Error logging capabilities : 0x1
Self-tests supported : yes
Offline Data Collection capabilities : 0x53
Offline Data Collection status : 0x0
Auto Offline Data Collection flags : 0x0
[Known device ]: yes
[Drive State Flags ]: 0x0


=== CURRENT POWER CYCLE STATISTICS ===
Data Read : 48.2 GB
Data Written : 38.2 GB
Data Read/Write Ratio : 1.26
Average Throughput (Read) : 60.6 MB/s
Average Throughput (Write) : 77.8 MB/s

Operations (Read) : 4,967,179
Operations (Write) : 2,275,016
Operations Read/Write Ratio : 2
Throughput per operation (Read) : 10.2 KB/Op
Throughput per operation (Write) : 17.6 KB/Op

Latency Time (Read) : 0 ns
Latency Time (Write) : 0 ns
Retries (Read) : 0
Retries (Write) : 0
Errors (Read) : 0
Errors (Write) : 0


=== PROBLEMS SUMMARY ===
Failed Indicators (life-span / pre-fail) : 0 (0 / 0)
Failing Indicators (life-span / pre-fail) : 0 (0 / 0)
Warnings (life-span / pre-fail) : 0 (0 / 0)
Recently failed Self-tests (Short / Full) : 0 (0 / 0)
I/O Error Count : 0 (0 / 0)


=== IMPORTANT HEALTH INDICATORS ===
ID NAME RAW VALUE STATUS
5 Retired Block Count 0 100% OK
173 Wear Leveling Count 0x800900048 96.0% OK
175 Host Writes MiB 13,764,312 (13.1 TB) 99.0% OK
192 Unsafe Shutdown Count 98 99.0% OK
197 Current Pending Block Count 0 100% OK
199 UDMA CRC Error Count 0 100% OK


=== TEMPERATURE INFORMATION (CELSIUS) ===
Current Temperature : 32
Power Cycle Min Temperature : 32
Power Cycle Max Temperature : 37
Lifetime Min Temperature : 32
Lifetime Max Temperature : 37
Recommended Min Temperature : 5
Recommended Max Temperature : 65
Temperature Min Limit : 5
Temperature Max Limit : 70


=== DRIVE HEALTH INDICATORS ===
ID | NAME | TYPE | UPDATE | RAW VALUE | VALUE | THRESHOLD | WORST | LAST MODIFIED | STATUS
1 Raw Read Error Rate Life-span online 0x0 200 0 200 - 100% OK
5 Retired Block Count Pre-fail online 0 100 0 100 - 100% OK
9 Power On Hours Life-span online 28,028 94 0 94 8/13/19 12:48 PM 94.0% OK
12 Power Cycle Count Life-span online 951 99 0 99 - 99.0% OK
169 Total Bad Block Count Pre-fail online 0x3390A800780 242 10 242 - 100% OK
173 Wear Leveling Count Life-span online 0x800900048 196 100 196 - 96.0% OK
174 Host Reads MiB Life-span online 17,528,594 (16.7 TB) 99 0 99 8/13/19 12:48 PM 99.0% OK
175 Host Writes MiB Life-span online 13,764,312 (13.1 TB) 99 0 99 8/13/19 12:48 PM 99.0% OK
192 Unsafe Shutdown Count Life-span online 98 99 0 99 - 99.0% OK
194 Temperature (Celsius) Life-span online 32 68 30 19 8/13/19 12:48 PM 54.3% OK
197 Current Pending Block Count Life-span online 0 100 0 100 - 100% OK
199 UDMA CRC Error Count Life-span online 0 200 0 200 - 100% OK



=== DRIVE ERROR LOG ===
error log is empty


=== DRIVE SELF-TEST LOG ===
# | LIFETIME (H) | TEST TYPE | PROGRESS | STATUS | LBA of 1st error
1 28018 Short offline 100% Completed without error -

As I said, "I have no idea what DriveX reports". Thanks for the information.

You and I have different interpretations of what is meant by "smoke and mirrors". To my way of thinking the specifications can be relevant in some situations. A friend wrote his doctoral dissertation on how much information was needed to make a sound decision and his careful research and testing revealed that the best decisions were more often made with minimal information, which would support your definition of "smoke and mirrors". On the other hand his research concluded that more information increased the decision maker's confidence in their decision and increased the likelihood of fully prosecuting any decision.

Disk Utility is an example of minimal information, but only tells you the disk has failed or not. I like seeing the additional parameters because on at least two occasions those have given me advance warning of impending failure. In one case complete failure was three days later and the other never completely failed, but over time its performance degraded to the point it was no longer satisfactorily usable. So in this case I prefer as much information as I can get.

I would like a SMART utility that automatically collected periodic SMART data and would notify the user when those values reached a user defined level or the rate of change increased. But AFIK that does not exist.

That would be a blessing!

The real question here, though, is in your expert - and I'm not using the word at all facetiously - opinion, does the extra info in the DriveDx report contribute anything useful or is it just smoke and mirrors as I"ve labeled it?

Even after having read all your posts on the subject of disk failure, I'm still at a loss.

Useful or "Smoke and Mirrors" is a matter of environment. Just taking the DriveX major categories:

• System Information — for an individual relatively pointless, but in a corporate or academic environment where there are lots of different machines this could be invaluable for record keeping purposes.
• ### DRIVE 1 OF 1 ### — In either the individual or organizational situation the identification section is at least interesting and contains data that can be difficult to ferret out
• === DEVICE CAPABILITIES === — not particularly useful to most users
• === CURRENT POWER CYCLE STATISTICS === — an indicator of age but without MTBF (Mean Time Between Failures) data not particularly informative, but as long as the data is collected it doesn't hurt to report it.
• === PROBLEMS SUMMARY === — ie. is the drive well on its way to failure or has it already failed.
• === IMPORTANT HEALTH INDICATORS === — pre-failure indicators
• === TEMPERATURE INFORMATION (CELSIUS) === — Heat was once thought to be a major factor in hard drive failure, but the famous Google study pretty well discounted that theory. I have no idea if there is even any data for SSDs. I wouldn't pay too much attention to this unless you were seeing consistently high temperatures which could be an indicator of ventilation problems in the device, fan failure, bearing failure in a rotating rust drive, or a replacement drive that is not properly sized for the computer.
• === DRIVE HEALTH INDICATORS === — at least half of this information has already been directly reported previously and the rest at least alluded to in the Problems Summary
• === DRIVE ERROR LOG === — Details of previously reported numbers
• === DRIVE SELF-TEST LOG === — Could be useful

CONCLUSION: For the individual user the meat of this report is contained in Problems Summary and Important Health Indicators which are repeated in Drive Health Indicators. The rest might be useful to some people in specific situations or would fit into the nice to know category.

I admit to being prejudiced, but I find Drive Scope's display significantly clearer and more understandable for the mythical average user. But that may be because I am used to it and originally got it to beta test.

Or to put it less politely, the report contains a lot of irrelevant information that pretends to give it weight that it hasn't got, i.e. "smoke and mirrors," and its "meat" isn't much, if anything, more than the details of the Pass/Fail we get from Disk Utility.

And beyond the clarity of its reporting, DriveScope displays critical info that Drive Dx doesn't - more likely can't - reach.

CONCLUSION: If you're going to spend money on a S.M.A.R.T. monitor, DriveScope is your only REAL option, but whether it's worth $50 or even $30...?

Actually the critical information DOES appear in DriveDX but the report requires closer reading or there is a risk that it could be overlooked or misinterpreted.

As to the VALUE of DriveScope I look at it this way. Over the years you, me, and a lot of others bought have purchased one or mire of the three "industrial Strength" volume repair tools for roughly $100 and have spent $30 to $50 or more every year or so since keeping it up to date. And every year we have used them less and less as MacOS has become more and more stable. With the advent of SSDs and APFS those tools have become more and more limited in what they can do and some things we have depended upon them for such as disk and file optimization/defragmentation have become counter productive and pointless. But the question of drive health still remains, particularly anticipating impending drive failure and that is the domain of SMART. But Disk Utility's pass/fail report provides NO WAY to anticipate future failure and given the manufacturer's tendency to be over-optimistic in setting the SMART attribute limits further analysis is needed.

DriveDX, DriveScope, and TechTool Pro provide the information needed for analysis to anticipate that failure. TechTool Pro provides the essential data in what I find to be a highly readable and comprehendible format along with a host of hardware testing tools and HFS+ HD volume repair utilities, DriveDX provides all of the data in text form but not in as obvious a format, Drive Scope provides all the data in what I find to be an easily apprehendable format.

• Would I pay $100 for TechTool Pro again — probably not. It is still a powerful utility but the simple fact is I don't have much, if any, call for its services lately. However, since I already have TTP I will keep it and buy updates if for no other reason than it does a decent job reporting SMART.
• would I buy DriveDX? I might if there were no other better options, but there are IMO better options
• Would I pay $50 for Drive Scope? I might, if I anticipated future updates were free and upgrades reasonably priced but as you said, that is a high price for a one trick pony. I initially got Drive Scope as a beta tester and it was little better than TTP, the new version however is a substantial step up and I didn't have much hesitation coughing up the $30 upgrade price.
• If I had to choose between TTP and Drive Scope it would be a tough decision. I used to feel TTP's eDrive and later their ProToGo drives alone were sufficient incentive to buy the product, but Apple's Recovery Drive/Recovery Assistant has pretty well rendered those TTP utilities non-essential. At this point I would probably go with Drive Scope as having the best value long term. That could easily change if TTP gained more APFS functionality.

Bingo!!!

It wasn't until I read that that it finally got through to me that DriveDx reports the SAME critical info that Drive Scope reports; without going into unnecessary detail, I was under the impression that DS is able to dig deeper into the "guts" of a SSD than DDx is.

And as for your conclusion, yeah, I"ve got to agree with you, so the same impetus that drove me to spend $100 + another $50 for DiskWarrior and (only because I was able to get it for a reasonable price) $20 (+ $10 for a CD, because I"ve got an unwarranted affection for hard media ) for TechTool Pro now impels me to invest once again.

Its complicated format and smoke & mirrors notwithstanding, I'd go with DDx simply because at $20 (and including a test period) it's a far better value than DS is at $50, but since my TechTool Pro serial # enables me to get DS for $30 I'll likely go with Micromat as the more established developer (and hope they maintain the price point, which seems quite sufficient for a one trick pony).

Thanks for your indulgence.

PS: DDx was unable to read S.M.A.R.T. on the SSD in my OWC Mercury Elite Pro mini enclosure, and according to what I"ve read, it's because it's only USB 3.0, so I'm wondering if that's a can't-get-around-it limitation or it can be circumvented?

PPS: Does DS offer functionality that TTP doesn't, i.e. why do you need both?

I've been following this thread with interest.

I have never had a drive fail, although I pulled a 8Gb HDD from my Beige G3 when it started ticking. How long ago was that!?

Anyway, I downloaded DriveDx and ran the trial. I buy HDDs for data and SSDs for OS/apps. I tend to replace a drive at the end of its warranty period and move it into a backup role.

DriveDx reported the SSD I use very rarely for Snow Leopard at 35% of its lifespan. As it's over five years old, that's not surprising. But what did surprise me is that the 80Gb IBM SSD that I keep on the shelf as backup for the SL drive was reported as excellent over the USB connection and it is over eight years old! I guess it is lasting longer because it is not powered up all day?

Anyway, it cost me £40 today for a DriveDx license and a new Crucial 120 Gb SSD.

Cheers!

I just ordered 1.0TB Mercury Extreme Pro 6G 2.5-inch 7mm SATA 6.0Gb/s Solid-State Drive from OWC to replace the failing 480 GB SSD.

I have three of those same enclosures and the limitation is the USB standard which does not provide for carrying S.M.A.R.T. data and so is not included in the USB bridge chips. Either Thunderbolt 2 or 3 provides for reporting S.M.A.R.T. in the standard.


It depends on whether the drive itself uses an ATA (SATA or PATA) or NVMe attachment.

• With an ATA attached HDs or SSDs DS not only reports the drive's attributes but additionally in separate screens details about the drive itself, its capabilities, can trigger shut or long self-tests, and a log of previous self tests including SMART pass/fail.
• With NVMe drives DS only reports the NVMe data and SMART logs. I believe this is a limitation of the NBVMe standard.

Only the attributes are essential and those are reported by TTP and DriveX as well, but the other information could be informative, not to mention the ability to trigger a self test. By the way a Short self test on a 1TB 7200 rpm HD takes 2 minutes, the long test takes 3 hours and the drive will be off-line during the test. (Not sure how that works with the boot drive??)

FOR MORE INFORMATION

• S.M.A.R.T.
• SATA
• NVMe

❓❓the 80Gb IBM SSD that I keep on the shelf as backup fo the SL drive was reported as excellent over the USB connection❓❓
Please elucidate! The USB standard does not support reporting SMART values. How were you able to accomplish this feat of legerdemain?

SSDs have no moving parts — rotating disks, moving read/write heads, bearings, etc. — that wear when the drive is powered up. The biggest wear factor for SSDs is the number of Write operations to the drive. A data segment on an SSD can only be written/re-written a finite number of times before it will no longer accept a write operation. There are a number of techniques to even out the wear on the drive (for details see this Wikipedia article) and there are spare sectors built into every SSD that can be mapped in when a well used sector reaches its limit. "Wear" is therefore measured by the percentage of spares that have been used, and the % remaining available spares. As long as an SSD is primarily being read and seldom, if ever, written to there is nothing creating wear and so it theoretically could last generations.

WARNING: While an SSD could theoretically last generations, that is not true of the data written on the SSD. All electronic data storage deteriorates over time and SSD data deteriorates faster than most. Experience and tests have shown that data on industrial grade SSDs begins to deteriorate in as little as two to three years and for whatever reason consumer grade SSDs can go as long as four years and still be reliably readable. The point is, if you are using SSDs for long term historical or archival storage of data, the data needs to be "refreshed" (re-written) every few years to maintain its integrity. Data on magnetic media such as Hard Drives is longer lived but HDs are not immune to data loss over time either. I suspect your eight year old backup drive has lost some, perhaps a significant amount of its data integrity over the years. The drive is still perfectly usable but the data would be suspect. That is why I don't recommend SSDs for Time Machine and other backup media. HDs are still a LOT less expensive, retain data integrity longer, and speed is not a significant factor for backup purposes.

Please elucidate! The USB standard does not support reporting SMART values. How were you able to accomplish this feat of legerdemain?

I had the SSD mounted in a USB caddy. DriveDx said if I wanted to get a report about the drive, I had to install another app, which I did. A reboot later, I got more info about the SSD.

The SSD in question is actually two 80Gb SSDs in a caddy which allows a RAID configuration, so it reads as a single drive.

Anyway, I've saved the report for you:

https://www.dropbox.com/s/d2hvur4uhzd3snx/DriveDxReport2019-08-19_20-17-17-294.txt?dl=0

Would using Carbon Copy Cloner every month constitute a "refreshing?"

As far as "cheaper" goes, this 120 Gb SSD only cost under £20! I think I paid £120 when it was new. I know it's not a data backup solution and 4Tb SSDs are wildly expensive, but for what I needed, £20 was a bargain.

What was the "other app"? There are others here who might like to know.

I notice the DriveDX report does not show the individual drives in your array and the actual SMART attribute values are not included in this report so in essence this becomes an elaborate SMART pass/fail similar to Disk Utility and other apps. What is missed by the lack of the individual attributes is any means of anticipating future drive failure. (I have had two drives fail completely yet Disk Utility still reported SMART passed.)

I have a four drive Thunderbolt 2 enclosure configured RAID 5 and Drive Scope reports each drive individually with all of their characteristics and reported SMART attributes, so I see the health of each drive independently. Since it is RAID 5 if any one of the four drives fails or is failing it can be "hot swapped" for a new drive while the array is in use and the system will automatically re-build/re-populate the failed drive so it is important that I know the health of each individual drive.


Maybe. Remember in CCC...

...so it is possible some files may grow quite old and never be rewritten.

This page on BinaryFruit's site will explain:
https://binaryfruit.com/drivedx/usb-drive-support

I downloaded the DriveDx installer from their site and just followed the instructions when it noticed the external (UBS-mounted) drive. And, as you say, the report was not nearly as detailed as for the internal drives.

I only attached the report for the external drive. All internal drives had rather extensive reports, impressive enough for me to part with my cash.

Just noting that DriveDx does all of the same.

There's not much of it, but isn't this the S.M.A.R.T. attributes section of freelance's report?

=== DRIVE HEALTH INDICATORS ===
ID | NAME | TYPE | UPDATE | RAW VALUE | VALUE | THRESHOLD | WORST | LAST MODIFIED | STATUS
194 Temperature (Celsius) Life-span online 44 56 30 50 - 37.1% OK

I kinda screwed up my earlier reporting, but, happily, not beyond redemption. Thanks for nudging me in the right direction.

First, the app joemike inquired about is SAT SMART Driver, and these are the DriveDx and Micromat pages on the subject; the DDx page includes a d/l link.

The first time I installed it I couldn't find anything to do in System Prefs, it didn't work, and I somehow got a "You're out of luck!" pop-up.

After reading your post, though, I installed it again, and again couldn't find anything to do in System Prefs, but this time it occurred to me to unmount/mount my external (DOH!), and VOILA!

Second, both of us have confused the issue a bit by posting only the complete reports that joemike and I find so cumbersome, so here are the actual GUI aspects of DDx's reports on my internal and external SSDs. (The small shots are expanded views of the collapsed General Information sections of the large shots above them.)

It now appears to me that DDx and DS are almost identical apps, although with obviously necessarily differing GUI aspects, and unless joemike sees any real down side to its functionality I'm going to buy it for $20 rather than SD for $30, not only to save myself $10, but to stick it to Micromat for what may easily be considered extreme overcharging, i.e. $50 if you haven't got a Micromat serial #.

And finally, in a different vein, it's interesting to note the difference between the DRIVE HEALTH INDICATORS and IMPORTANT HEALTH INDICATORS reported by DDx for my internal (Apple (Samsung-based) g-series SSD) and external (SanDisk Marvell-based SSD) drives:

Internal:

=== DRIVE HEALTH INDICATORS ===
ID | NAME | TYPE | UPDATE | RAW VALUE | VALUE | THRESHOLD | WORST | LAST MODIFIED | STATUS
1 Raw Read Error Rate Life-span online 0x0 200 0 200 - 100% OK
5 Retired Block Count Pre-fail online 0 100 0 100 - 100% OK
9 Power On Hours Life-span online 28,181 94 0 94 8/19/19 10:17 PM 94.0% OK
12 Power Cycle Count Life-span online 952 99 0 99 - 99.0% OK
169 Total Bad Block Count Pre-fail online 0x3390A800780 242 10 242 - 100% OK
173 Wear Leveling Count Life-span online 0x800900049 196 100 196 - 96.0% OK
174 Host Reads MiB Life-span online 17,622,696 (16.8 TB) 99 0 99 8/19/19 10:17 PM 99.0% OK
175 Host Writes MiB Life-span online 13,845,735 (13.2 TB) 99 0 99 8/19/19 10:17 PM 99.0% OK
192 Unsafe Shutdown Count Life-span online 98 99 0 99 - 99.0% OK
194 Temperature (Celsius) Life-span online 35 65 30 19 - 50.0% OK
197 Current Pending Block Count Life-span online 0 100 0 100 - 100% OK
199 UDMA CRC Error Count Life-span online 0 200 0 200 - 100% OK

=== IMPORTANT HEALTH INDICATORS ===
ID NAME RAW VALUE STATUS
5 Retired Block Count 0 100% OK
173 Wear Leveling Count 0x800900049 96.0% OK
175 Host Writes MiB 13,845,735 (13.2 TB) 99.0% OK
192 Unsafe Shutdown Count 98 99.0% OK
197 Current Pending Block Count 0 100% OK
199 UDMA CRC Error Count 0 100% OK

External:

=== DRIVE HEALTH INDICATORS ===
ID | NAME | TYPE | UPDATE | RAW VALUE | VALUE | THRESHOLD | WORST | LAST MODIFIED | STATUS
5 Retired Block Count Life-span online 0 100 0 100 - 100% OK
9 Power On Hours Life-span online 5,071 253 0 100 - 100% OK
12 Power Cycle Count Life-span online 638 100 0 100 - 100% OK
166 Min Write/Erase Cycle Life-span online 1 100 0 100 - 100% OK
167 Min Bad Block/Die Life-span online 53 100 0 100 - 100% OK
168 Maximum Write/Erase Cycle Count Life-span online 53 100 0 100 - 100% OK
169 Total Bad Block Count Life-span online 0x3DF 100 0 100 - 100% OK
171 Program Fail Count Total Life-span online 0 100 0 100 - 100% OK
172 Erase Fail Count Total Life-span online 0 100 0 100 - 100% OK
173 Average Write/Erase Count Life-span online 10 100 0 100 - 100% OK
174 Unexpected Power Loss Count Life-span online 298 100 0 100 - 100% OK
184 End-to-End Error Life-span online 0 100 0 100 - 100% OK
187 Reported Uncorrectable Errors Life-span online 0 100 0 100 - 100% OK
188 Command Timeout Life-span online 0 100 0 100 - 100% OK
194 Temperature (Celsius) Life-span online 34 66 0 50 - 66.0% OK
199 SATA CRC Error Life-span online 0 100 0 100 - 100% OK
212 SATA Error Count Handshake Life-span online 0 100 0 100 - 100% OK
230 Percentage Total P/E Count Life-span online 33 100 0 100 - 100% OK
232 Available Reserved Space Pre-fail online 0x64 100 4 100 - 100% OK
233 Total NAND Writes Life-span online 5,725 (5.6 TB) 100 0 100 - 100% OK
241 Lifetime Writes Life-span offline 3,307 (3.2 TB) 253 0 253 - 100% OK
242 Lifetime Reads Life-span offline 3,515 (3.4 TB) 253 0 253 - 100% OK
244 Thermal Throttle Life-span online 0x0 0 0 100 - 100% OK

=== IMPORTANT HEALTH INDICATORS ===
ID NAME RAW VALUE STATUS
5 Retired Block Count 0 100% OK
168 Maximum Write/Erase Cycle Count 53 100% OK
230 Percentage Total P/E Count 33 100% OK
232 Available Reserved Space 0x64 100% OK
233 Total NAND Writes 5,725 (5.6 TB) 100% OK
241 Lifetime Writes 3,307 (3.2 TB) 100% OK

The differences between what the two manufacturers consider important may be telling to joemike, but it's meaningless to me.

Is more actually more, or is it less?

I expect to receive my new SSD in a few days. In the meantime, I need to erase the defective SSD before discarding it. Is a simple erase sufficient for an SSD or should I use a 7-way erase?

A simple erase erases ONLY the volume directory and leaves the data untouched. Granted the data is not as organized as it would be on an HD and therefore much harder to piece together, but given enoughvdesire, time, and money it is potentially recoverable. A 7 pass erase actually overwrites the data cells with alternating patterns of ones and zeros and I don't know if even an NSA level lab could recover it or not. I have read varying opinions on that but all agree someone would really have to want the data.

A valid question! My antipathy for the mimalist SMART pass/fail is based on two different drives that failed completely, but based on an aggregate SMART score Disk Utility still reported the drives had passed SMART. On the other hand, TechTool Pro showed one or more SMART attributes that had exceeded the limits and were being reported as FAIL.

Even if manufacturers set more realistic pass/fail limits for the SMART attributes the minimalist score still would only tell you if the drive had failed, and you probably already know or at least suspect that is the case. Having all the attributes permits an informed user to anticipate impending failure and perhaps short stopping a McGillicuddy's corollary to Murphy's Law event. The fly in this ointment is the word "informed".

In the final analysis the additional information is useless if the user does not understand what the attributes are telling her/him. It would be possible for an app to apply AI techniques to analyze the data, but then they risk being sued by drive manufacturers for saying a drive had failed or is failing when, as in the case of the two drives I mentioned, it still passes the manufacturer's overly optimistic failure criteria.

In the meantime I wish there were an app that would check the SMART values at a user specified and report not only changes in the attribute values, but the rate of change in selected key values. That would truly indicate impending failure. (I think I will write Micromat and suggest this as a new feature for Drive Scope.

According to Erase a volume using Disk Utility on Mac, a Secure Erase is not possible with an SSD. "With a solid-state drive (SSD), secure erase options are not available in Disk Utility. For more security, consider turning on FileVault encryption when you start using your SSD drive." I suppose that I could turn on FileVault encryption on the disk and not erase it(?)

Note that is SECURE erase that is not available. You can erase/reformat an encrypted SSD, I did it just a few days ago but that was a simple erase that only erased the volume directory. The content of the drive remained as encrypted text which will be overwritten as the drive is used.

NOTES:

1. Be forewarned: Filevault encrypts every sector on the drive, whether there is data in that sector or not, so it can take a long time — even days — to complete. You can continue to use the drive normally while encryption is going on but you will likely experience a slowdown in performance while it is happening.
2. You can format a drive APFS (encrypted) or HFS+ (encrypted) and that will work if it is a data only volume.
3. You cannot install MacOS on an empty volume formatted APFS (encrypted) or HFS+ (encrypted). You must make the initial installation on an unencrypted volume and THEN turn on Filevault to encrypt the volume
4. You can upGRADE or upDATE an existing copy of MacOS on an encrypted volume.

It appears that it is only possible to encrypt the boot drive with FileVault. Does that mean that I would have to boot from the defective SSD (it is a clone, after all) and then encrypt it? I'm thinking that it would be easier to drill several holes in the drive or, better yet, expose it to a nuclear blast.

And an excellent answer; thanks.

Your final point is so obvious, though, that I've got to wonder why such an app hasn't been developed yet; although it obviously couldn't provide definitive failure anticipation, it doesn't seem as if it would be a big deal to enable it to provide at least minimal proactive early warning functionality which would certainly surpass the functionality of Drive Scope and DriveDx.

But... You didn't actually answer the question I (thought I) really asked, namely, do the different mixes of SMART attributes reported for my Samsung and SanDisk SSDs give either report an edge in assisting me in anticipating drive failure?

And in a different vein, doesn't it seem odd that only one SMART attribute is reported for freelance's SSD?

=== DRIVE HEALTH INDICATORS ===
ID | NAME | TYPE | UPDATE | RAW VALUE | VALUE | THRESHOLD | WORST | LAST MODIFIED | STATUS
194 Temperature (Celsius) Life-span online 44 56 30 50 - 37.1% OK

I suspect part of the reason is SMART was playing second fiddle to surface scans and other indicators applicable to HDs. The resurgence of interest in SMART is fairly recent and is just now being fleshed out. 🤷‍♂️


Yes — if you understand enough about what they are reporting to interpret the attributes. There are a whole series of SMART attributes for hard drives and SATA connected SSDs that are intended to indicate a drive is in a "pre-failure" status, and others that indicate the age of the drive and both are so designated in the Drive Scope report. Whether a report indicates a parameter is a "pre-fail" indicator or not that attribute still indicates "pre-fail status". See Drive Scope HD Attributes. Any positive number in a pre-fail attribute is at least a caution flag. If a pre-fail attribute racks up half of the threshold value you should have the drives replacement on order and planning your switchover strategy. Typically the closer you get to a threshold value the faster the rate of deteriration.

Unfortunately NVMe drives do not have as rich a set of SMARY attributes. (See Drive Scope NVMe Attributes) but the Spares statistics are a good surrogate for "pre-fail" attributes and units Read and Written are a decent indicator of age while any NO in the Yes/No attributes indicates failure.


There are a LOT of reasons for that:

• The drive is connected via USB and the USB standard does not allow for reporting SMART so he is getting the report via what is essentially a hack.
• The enclosure is a two drive hardware RAID which is pretty much guaranteed to be either RAID 0 or RAID 1 and will be "seen" by the computer as if it were a single entity.
• We have no idea what attributes the two drives in the enclosure report or whether the closure attachment is ATA or NVMe and it is doubtful the hack can sort that out.
• I see the individual drives in my RAID array because the connection is Thunderbolt 2 which does provide for reporting SMART and I am using Software RAID so the data is read and written to individual drives and assembled into a single stream in the computer.

I think I"m beginning to understand this.

After looking through your report and my two (Damn design copyrights!), I"m left with these questions:

1. I assume it's manufacturer option, but do you see any logic as to why some attributes are identified as "Pre-Fail" on one drive and "Life-Span/Old Age" on another?
2. "Any positive number" in which column of your report?
3. Would you please explain "If a pre-fail attribute racks up half of the threshold value" graphically? I don't follow it from what I've been looking at.

The report on my SanDisk SSD was generated by the same "hack," and it appears to be complete.

1. Why the difference in interpretation?
   1. There can be honest differences of engineering opinion on whether a given attribute is a reliable pre-fail indicator or merely old age.
   2. Although the attributes number and name are defined by the SMART standard the "tolerance" and what its significance is, is up to the manufacturer's engineers as "tweaked" by their marketing department.
   3. You also have to consider those engineer's first or only language may well not be English so the connotation of the significance title could well come into play here as well.
2. I didn't think that one out too well I probably should have said any positive raw value in a pre-fail indicator but I might not want to be confined to the pre-fail values. (If I had the resources of Google Labs or access to Apple's Data Centers it would be interesting to statistically test and find the optimum set of SMART attributes to choose for SATA HDs and SSDs as well as NVMe drives)
3. Pick an attribute, say attribute number 1. "Raw Read Error Rate" with a [max] value of 100% and a threshold of 62%. That implies that when the Raw Read Error Rate reaches 62% of the attempts the drive has failed that attribute. (Note I got those numbers from an HGST drive and a raw read error rate that high would indicate the drive's performance was in the toilet.) But back to my suggestion if the Raw Read Error Rate reaches 31% I would want to know it because
   1. the drive performance is suffering
   2. it is likely caused by magnetic media flaking off the platter
   3. Once that starts it is likely to continue and accelerate
   4. I would want to look at the other pre-fail attributes to see if they confirm the Raw Read Error Rate. — The drive may well have already failed but not reached the manufacturer's too rosy failure settings.

Remember freelance's drive is a two drive RAID array so it is quite reasonable that even with the hack only one value can be seen and extracted. Reconfiguring the enclosure to another RAID level or if possible turn off RAID so the drives can be used individually and the results might be quite different.

You might want to take a look at the Drive Scope Manual for a clear and concise description of where the attribute data comes from for both SATA and NVME drives and pointers on interpretation. It doesn't take long to read and is the best description I have come across and is a free download.

I just read it and found to my surprise that only the first thirteen attributes for SATA drives are "set in stone" and the rest may or may not be the same from drive to drive. There were also reminders that with SSDs whether they are SATA or NVMe interface when key attributes are exceeded the drive may switch automatically to Read Only mode. From Micromat's forums I discovered that SATA SSDs may report the same attributes but the attribute itself is inferred (my term) from other values more appropriate to SSDs.

Further update. I found that Drive Scope actually has a built-in installer the hack for USB connected drives but it won't install on Catalina

Thanks for the link to the Drive Scope manual; it's a useful read.


That sentence is unclear.

I am glad the Drive Scope manual was helpful. As to my sentence, although it seemed perfectly clear when I wrote it, in retrospect it isn't clear at all. 🤷‍♂️

What I was attempting to say is that although the SMART report for a SATA attached SSD may be reported as a standard SMART attribute value the drive itself may not actually report that value as it does not directly apply to an SSD so the reported value may be "inferred" from values that are reported by the drive.

Although the first 13 attributes are supposedly "set in stone" by the SMART alliance standards the only example I have of a SATA attached SSD, an OWC Mercury Electra 6G SSD, and its reported attribute names only match two of my HGST HD's attributes: 1. "Raw Read Error Rate" and 178. Temperature Celsius. Nine of the SSD's reported attributes have the very informative attribute name "Unknown" and all are of the type "old-age". (Apparently only heaven and OWC know when one of their Mercury Electra 6G SSDs is failing or has failed — which is disappointing.)

Got it, thanks!


This is a compilation of the attributes common to my, yours, and freelance's SSDs. (You posted your attributes in an image so I couldn't delete the extraneous items.)

For whatever it's worth, they're not all identical.

That is exactly my point! Supposedly the first thirteen are supposed to be identical according to the SMART Alliance, but the manufacturers tend to go their own way. Personally the attributes I look for are:

1. Hard Drives & SATA attached SSDs
   1. Raw. Read Error Rate
   2. Spin Up Time
   3. Reallocated Sector Count
   4. Seek Error Rate
2. NVMe
   1. % Available Spares
   2. % Used
   3. % Remaining
   4. Any YES answer to the other items means the drive isn't failing, it has failed.
3. What I would like to have and what I get are NOT always (seldom?) the same thing 🤷‍♂️