6

I have a 512GB Samsung SSD in a 2014 MacBook Pro that I purchased new around 2 months ago. I installed Smartmontools via Homebrew to take a look at some of the SSD health / SMART data. Everything looks fine for the most part, but one thing that I'm not understanding is the Wear_Leveling_Count which currently has a value of 12895125514.

This really isn't my area of expertise but I thought Wear_Leveling_Count was typically reported as either a percentage (indicating the life remaining according to the endurance specifications of the manufacturer), or as a raw P/E cycle count. In either case, the value I'm getting seems extremely high.

Does anybody know what's going on here? Am I misinterpreting / misunderstanding something here? Could this potentially be a bug in Smartmontools or the SSD firmware? Smartmontools is also reporting the Wear_Leveling_Count ID as 173 when it should be 177 shouldn't it?

As I mentioned, I'm a bit of a novice in this particular area so any info / suggestions are greatly appreciated.

Update: I came across this reddit thread regarding Wear_Leveling_Count: https://www.reddit.com/r/linux/comments/31btmz/what_is_your_177_wear_leveling_count/

About halfway down the page there's a Samsung XP941 user (which I believe is the same SSD used in the 2014 MacBook Pro). He reports a Wear_Leveling_Count value of 197, which is much more in line with what I'd expect. It also looks like the Wear_Leveling_Count ID should indeed be 177 (as I mentioned, I get an ID of 173). This makes me think it might be some sort of OS X / Apple specific bug.

Does anybody know if Samsung does something different to the SSDs they manufacture for Apple? Smartmontools reports the SSD as an APPLE SSD SM0512F (which as I mentioned, is supposedly just a Samsung XP941). Could something unique to Apple SSDs be confusing Smartmontools?

Thanks

Improve this question
2
  • It has to do with Apple configuration though. Did you enable TRIM on your SSD? Mac computers, by default, do not have it enabled for third-party SSD. You have to update to OS X 10.10.4, then go to the Terminal app and run sudo trimforce enable. After entering your password (it's normal that you don't see what you type), it should enable TRIM on your device, so when deleting a file, the OS doesn't "delete" the file itself, but delegates that to the SSD with its own implementation. That reduces the OS overhead over the SSD a lot. – Alejandro Iván Aug 9 '15 at 3:11
  • Thanks for the suggestion. Sorry I didn't make it clear but the SSD I'm referring to is the stock Apple SSD (model: APPLE SSD SM0512F). I believe Apple uses both Samsung and SanDisk SSDs in the 2014 MacBook Pro so I was just mentioning that I had the Samsung model in case that was relevant to the issue. So TRIM is already enabled by default. – froglegs Aug 9 '15 at 4:43
3

I don't know either how exactly to interpret the RAW_VALUE column (the big number in the final column of the line starting with "173 Wear_Leveling_Count"), but the most significant values in that line are columns VALUE and THRESH.

As long as VALUE stays well above THRESH, your disk should not be approaching its specified end of life yet. I believe the initial value on a fresh disk is 200. When it hits 100 (my value for THRESH), it will have performed as many erase cycles as guaranteed by the manufacturer. After that you're running on borrowed time. On my (relatively lightly used) MacBook Pro, VALUE is 0197, meaning I am perhaps 3% into the life of the drive.

For what it's worth, my RAW_VALUE is 12,891,062,312, a little lower than your value of 12,895,125,514 and about 1/1000 of 12,567,284,681,697 (VALUE: 093) and 13,031,154,781,293 (VALUE: 87) reported in one of the other answers.

For information, the file system seems have written 5.4 TB on my machine so far:

175 Host_Writes_MiB 0x0022 099 099 000 Old_age Always - 5445917

What matters for the actual life of the drive is not the data logically written by the file system but the actual internal write cycles, which can be significantly different. However, if what I have written so far is about 1/30 of its lifetime limit, that would be about 160 TB which sounds about right for a 256 GB drive like mine.

Improve this answer
2

I am running Windows 8 on a MacPro with a Apple SSD SM0512F disk. Lately I've received the following error, which seems to be the same as yours (SMART-value AD):

Windows detected a hard disk problem

I've roamed the net a bit, trying to determine if this is a serious error or just misinterpretation/incompatibility between Windows and the Apple disk (this forum post hints the same), but I am starting to believe that it's in fact the disk that is being worn out.

A thread at Tom's Hardware hints the error might be due to either:

  1. Windows prematurely eating up the SSD's life
  2. This being an actual SSD fault, but Apple trying to keep it a secret

Update 2018

I just want to add that this disk actually finally died. It took less than two years from the purchase date. Unfortunately there was no warranty. The Apple ChurchStore wanted $1350 to replace the disk. I fortunately found a used one on Ebay for 1/4th of the price. Works fine till this day.

I can only speculate in why the disk died prematurely. No good advice to pass on unfortunately. However, at least in my case, the result shows that the SMART detection was unto something.

Improve this answer
3
  • Thanks for the info. Have you tried viewing the SMART info under Windows by any chance (you can check it out with a free tool called CrystalDiskInfo)? I'm curious if you receive the extremely high Wear_Leveling_Count value (in addition to the seemingly incorrect Wear_Leveling_Count ID# of 173) under Windows. If you do receive those erroneous values under Windows, that would suggest that there is indeed something different with the Samsung XP941 SSDs used by Apple (evidenced by the SMART info produced by the standard / stock Samsung XP941 in that reddit thread I linked). – froglegs Sep 7 '15 at 21:39
  • 1
    Yes, I've run CrystalDiskInfo on Windows and it coincides with your observations. To me, the Weal_Leveling_Count simply seems like misinterpreted data. If it was true, I'd suspect my disk would have stopped working a long time ago. – nitech Sep 8 '15 at 8:38
  • Thanks, I agree then that it's being misinterpreted. The fact that the Samsung XP941 that didn't come from a MacBook Pro (i.e. the one from the reddit thread I linked), correctly displays the Wear_Leveling_Count value and ID, makes me think Apple is using their own, unique firmware or something. The fact that smartmontools displays the SSD model as "APPLE SSD SM0512F" also seems to support this. Anyway, thanks again for sharing your experience. The fact that you're receiving the same output (i.e. it's unlikely to be some isolated issue with my SSD) makes me feel better. – froglegs Sep 9 '15 at 3:55
0

I'm having this problem with two identical iMacs bought on the same day. They're showing:

173 Wear_Leveling_Count 0x0036 093 093 100 Old_age Always FAILING_NOW 12567284681697

173 Wear_Leveling_Count 0x0036 087 087 100 Old_age Always FAILING_NOW 13031154781293

I'm quite convinced it's a FW bug, since the bottom one has been working for almost a year now with that message. The second one just started reporting it 2 weeks ago. The thing is I can't do any OS X upgrades any more due to SMART errors...

Also, both have also a Windows partition managed with Bootcamp.

Improve this answer
1
  • 1
    Welcome to Ask Different! When answering a question, take care to directly answer the OP. This answer as written doesn't directly answer the question and instead postulates a guess. Try to find additional sources to reference for a higher quality answer. – Allan Apr 25 '16 at 14:59

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .