Curious S.M.A.R.T. issue

[ubuysa]

I keep a fairly regular check on the SMART data on my drives to try and predict failures ahead of time. A couple of days ago I came across something unusual which I thought might be worth sharing to see whether any SMART people (see what I did there?) might know more than I about what might be going on.

The drive in question is a 3.5" 1.5TB HDD that I've had for probably close on 8 years (at a guess). It used to be an external drive but has been an internal drive in my PCS PC for the last 12 months. I use it as a media drive but there's nothing on there that isn't also in at least two other places. CrystalDiskInfo shows a caution for this drive because of the Current Pending Sector Count Value.



Current Pending Sector Count is a measure of how unstable the drive has become and is one of the key indicators of drive health. These sectors (and I appear to currently have 6) have apparently become unstable and need to be remapped into the reserved sectors. They are apparently 'pending' until they are next read, a successful read should decrease the Current Pending Sector Count (and presumably a failed read should trigger a remapping?).

According to the SMART data no sectors have yet been remapped, the Reallocated Sectors Count is zero (the normalised worst value is 200), and interestingly the Read Error Count is also zero (the normalised worst value is 200), which is bizarre if I have 6 'flaky' sectors? However, the Write Error Count has been positive in the past (the worst normalised value is 149) and the Uncorrectable Sector Count has also been positive in the past (the worst normalised value is 198) but for an old drive I don't see those as a particular problem? But for the CrystalDiskInfo Caution indicator I don't see anything of major concern with this drive, especially given its age?

Still concerned I ran a wmic /namespace:\\root\wmi path MSStorageDriver_FailurePredictStatus command at the Command Prompt, it reported all drives as 'OK' and a System Diagnostics run via the Performance Monitor reported no disk errors either.

My first instinct was to run a chkdsk /r on the drive, which I did. It found no errors at all. Still not happy I did a full (ie. not quick) format of the drive, this zeroes every sector and thus tests the drive surface to some degree. It formatted cleanly. Still not entirely happy I ran the Ccleaner drive wipe over the whole drive using 7 passes (yes it did take a couple of days!). That reported no errors.

At no time did the Current Pending Sector Count change and yet the drive appears to be working normally.

I'm not one for taking chances so a new HDD to replace this one is already on order, but I'm still curious to understand why the Current Pending Sector Count stubbornly remains at 6 no matter what I do to the drive?

 

[SpyderTracks]

I’m old school. I’ll rely on smart for a warning, but I don’t do anything unless a chkdsk shows an error.

As to why it’s flagging, no idea unfortunately, I’m not very aufait with SMART measurements

 

[NoddyPirate]

That’s interesting @ubuysa - some sources say that if Check Disk fixed any errors then a system restart should reset the SMART data. Others say the SMART data might not reset depending on the manufacturers programming.

Another suggests that the error log is retained for 2^32 ms - what the relevance of that number is I have no idea but it seems beautifully binary! So you might have to wait 47 days or so before the error log overwrites? 🙂

But I guess SMART info is an indication of historical health in many ways. So the existence of an issue in the past dramatically increases the likelihood of failure in the shorter term. With that in mind I can see a logic to some attributes not resetting even if the problem is resolved? For example, if your drives triggers a high temperature alert, it wouldn’t make any sense to me to have that SMART attribute reset once the temperature improves. You might never know it happened otherwise.

Nor would it make sense to be able to clear that alert log. If overtemp happened just once, the chances of failure have increased and the SMART flag should be permanently retained?

Given the total lack of standardisation across manufacturers, perhaps your manufacturer has simply taken the latter view?

 

[ubuysa]

That’s interesting @ubuysa - some sources say that if Check Disk fixed any errors then a system restart should reset the SMART data. Others say the SMART data might not reset depending on the manufacturers programming.

Another suggests that the error log is retained for 2^32 ms - what the relevance of that number is I have no idea but it seems beautifully binary! So you might have to wait 47 days or so before the error log overwrites? 🙂

But I guess SMART info is an indication of historical health in many ways. So the existence of an issue in the past dramatically increases the likelihood of failure in the shorter term. With that in mind I can see a logic to some attributes not resetting even if the problem is resolved? For example, if your drives triggers a high temperature alert, it wouldn’t make any sense to me to have that SMART attribute reset once the temperature improves. You might never know it happened otherwise.

Nor would it make sense to be able to clear that alert log. If overtemp happened just once, the chances of failure have increased and the SMART flag should be permanently retained?

Given the total lack of standardisation across manufacturers, perhaps your manufacturer has simply taken the latter view?

Indeed, that lack of standardisation is a big issue, especially where the RAW data is concerned. Clearly nothing is resetting that Current Pending Sectors Count and it's impossible to discover whether that is by intent or whether it's an indication of a problem.

In any case, in 47 days the replacement drive will be here and installed.

IMO there is no point in having failure indicators like SMART data if you're not going to pay attention to them - even if that does mean being a bit over-cautious.

 

[NoddyPirate]

Indeed, that lack of standardisation is a big issue, especially where the RAW data is concerned. Clearly nothing is resetting that Current Pending Sectors Count and it's impossible to discover whether that is by intent or whether it's an indication of a problem.

In any case, in 47 days the replacement drive will be here and installed.

IMO there is no point in having failure indicators like SMART data if you're not going to pay attention to them - even if that does mean being a bit over-cautious.

I agree with the last paragraph personally. But I also wonder if the first paragraph is kept deliberately ambiguous by manufacturers to ensure that the last paragraph works in their favour? (Does that makes sense? Cos even I don’t know what I just said!)

Weirdly cynical I know - but I have been shocked to see that the failure rates of HDD’s, particularly high capacity ones, is still terrible in this day and age. The Seagate ‘scandal’ a few years ago of hidden SMR drives added in to it - makes me feel like manufacturers are fighting a price war that has no future in it if drives don’t need to be replaced ‘relatively’ frequently.

That weirdly cynical view would give an incentive to ensure that SMART flags never reset too I guess. 🤔

 

[ubuysa]

I agree with the last paragraph personally. But I also wonder if the first paragraph is kept deliberately ambiguous by manufacturers to ensure that the last paragraph works in their favour? (Does that makes sense? Cos even I don’t know what I just said!)

Weirdly cynical I know - but I have been shocked to see that the failure rates of HDD’s, particularly high capacity ones, is still terrible in this day and age. The Seagate ‘scandal’ a few years ago of hidden SMR drives added in to it - makes me feel like manufacturers are fighting a price war that has no future in it if drives don’t need to be replaced ‘relatively’ frequently.

That weirdly cynical view would give an incentive to ensure that SMART flags never reset too I guess. [emoji848]

I think if I was running a data center with thousands of drives then I might be as cynical as you. But for one 8 year old drive? Nah, scrap it to be safe - even if that's what they want me to do!

 

[NoddyPirate]

I think if I was running a data center with thousands of drives then I might be as cynical as you. But for one 8 year old drive? Nah, scrap it to be safe - even if that's what they want me to do!

See? Their plot is working perfectly! 😀

When your new drive arrives you should leave that old one plugged in for fun and see what happens in 47 days. A forum experiment!

The reference I saw was to time stamps rather than run time so I don’t think it has to be powered on permanently. But then you leave your system on full time anyway I think?

 

[ubuysa]

See? Their plot is working perfectly! [emoji3]

When your new drive arrives you should leave that old one plugged in for fun and see what happens in 47 days. A forum experiment!

The reference I saw was to time stamps rather than run time so I don’t think it has to be powered on permanently. But then you leave your system on full time anyway I think?

I don't have enough SATA ports to leave it in. I do leave it on 24 x 7, you can see that in the CrystalDisk Info power on time.

Although the data on there is backed up (twice) a new 2TB drive was around €60 delivered. My peace of mind, not to mention my data, is worth a lot more than that.

Don't buy a dog and then bark yourself.

 

[NoddyPirate]

I don't have enough SATA ports to leave it in. I do leave it on 24 x 7, you can see that in the CrystalDisk Info power on time.

Although the data on there is backed up (twice) a new 2TB drive was around €60 delivered. My peace of mind, not to mention my data, is worth a lot more than that.

Don't buy a dog and then bark yourself.

Agreed! (Woof woof!)

(Just plug it back in briefly on June 24th, actually make it the 25th, and see what it says. 🙂)

Id be interested to know if you ever get to bottom of it though. I replaced my Mother’s laptop HDD with an SSD - it was on its last legs and the SMART status showed over 1,000 bad sectors - but my brothers HDD is heading down the same road performance wise - less then half the read/write speed you’d expect and 100% disk usage in Task Manager for extended periods - yet his SMART info looks OK.

I’d have to wonder if the SMART features are being handled totally differently between the two....

 

[ubuysa]

Agreed! (Woof woof!)

(Just plug it back in briefly on June 24th, actually make it the 25th, and see what it says. [emoji846])

Id be interested to know if you ever get to bottom of it though. I replaced my Mother’s laptop HDD with an SSD - it was on its last legs and the SMART status showed over 1,000 bad sectors - but my brothers HDD is heading down the same road performance wise - less then half the read/write speed you’d expect and 100% disk usage in Task Manager for extended periods - yet his SMART info looks OK.

I’d have to wonder if the SMART features are being handled totally differently between the two....

I hope your brother's data is all backed up?!

Entropy always increases.

 

[NoddyPirate]

I hope your brother's data is all backed up?!

Entropy always increases.

Yes - SSD replacement on the way also along with some extra RAM (currently 8GB and hovering near 90% usage a lot).

It’ll be a fun hour or two with some screwdrivers and a beer! No wait - make that A screwdriver and SOME beer. 👍

 

[Bigfoot]

Pr

Yes - SSD replacement on the way also along with some extra RAM (currently 8GB and hovering near 90% usage a lot).

It’ll be a fun hour or two with some screwdrivers and a beer! No wait - make that A screwdriver and SOME beer. 👍

Probably best to resort to the screwdriver first and then the beers, rather than the other way around.

 

[NoddyPirate]

Pr

Probably best to resort to the screwdriver first and then the beers, rather than the other way around.

Good advice!

Last time around I broke the retaining clip on the keyboard rubbon cable connector - I think the beer hindered my ability to see that my fix - just shove the cable back in and hope for the best - was unlikely to work.

After installing the new drive I could not get into BIOS to boot from the USB key for the Windows Install. In fact I couldn't get any hot key functions to work at all. After about 90 minutes of swearing it dawned on me that the keyboard wasn't working.

Thankfully I found the most fantastic fix - a piece of electrical tape applied under the cable to make it a snug fit in the connector and all has been well since.