Number of Reported Uncorrectable Errors does not match SMART

[Haole Boy]

Aloha again. I seem to get strange hard drive "errors" on a regular basis. Here's another one...

NOTE: I am asking this question to learn more about hard drives from the experts. I've asked questions like this a few times in the past and I get told multiple times to just replace the drive. Please don't do that with this thread, I'm trying to learn something here. Thanx

So, here's what is going on. Customer complaining about apps crashing. I run the Malwarebytes Toolset and it tells me that a drive has 2749 reported uncorrectable errors! This is an external (USB connected) Seagate drive. I run gSmartControl to take a look at the drive and the pertinent SMART values look like this:

ID# ATTRIBUTE_NAME FLAGS VALUE WORST THRESH FAIL RAW_VALUE
5 Reallocated_Sector_Ct PO--CK 100 100 010 - 0
187 Reported_Uncorrect -O--CK 001 001 000 - 458
197 Current_Pending_Sector -O--C- 100 092 000 - 0
198 Offline_Uncorrectable ----C- 100 092 000 - 0

And then the Statistics section looks like this:

Page Offset Size Value Flags Description
0x04 ===== = = === == General Errors Statistics (rev 1) ==
0x04 0x008 4 2749 --- Number of Reported Uncorrectable Errors

So, I'm confused. How can there be 2749 uncorrectable errors, but the 'Reported Uncorrectable' (SMART 187) shows 458? If the 458 is hex, then that's 1112 decimal which is still less than 50% of what's showing up in the statistics section.

And why are the other 'bad sector' indicators not showing any problems (they are all zero)?

I've attached the full gSmartControl output to this post if anyone wants to look at it.

Mahalo,

Harry Z

 

[Porthos]

 

[phaZed]

First, SMART is not accurate - more often than not. SMART should only be used as an indicator that something is wrong, not gospel.

Also, pending the manufacturer and model... 1 error in SMART doesn't need to correlate to one error. Maybe it's 6 errors for every SMART reported 1 error, etc. Really, like I said, SMART doesn't have to make sense... the drive is bad.

Here's your biggest "other" issue:

Raw_Read_Error_Rate     POSR--   082   046   006    -    143398926
Hardware_ECC_Recovered  -O-RC-   082   064   000    -    143398926
Seek_Error_Rate         POSR--   082   060   045    -    171096619

Those should be Zero. Drive is bad, replace. I know you don't want to hear that ;-)

Remember, it's the drive's firmware that "writes" to SMART, so it's up to that firmware to be "correct" or "incorrect" - SMART is simply the "interface" that software can check on the firmware's reporting.

 

[sapphirescales]

I'm trying to learn something here.

Where are your threads asking about dot-matrix printers and Betamax players? After all, if you're so curious about obsolete technology, why not go all the way?

Using a hard drive these days is like using a dot-matrix printer.

 

[Computer Bloke]

There are only two interesting values for the number of uncorrectable errors or reallocated sectors - zero and anything else.

Anything but zero, replace the drive. With an SSD if it's internal.

Easy.

 

[glennd]

There are only two interesting values for the number of uncorrectable errors or reallocated sectors - zero and anything else.

Anything but zero, replace the drive. With an SSD if it's internal.

Easy.

Mr Boole would be proud of you

 

[backwoodsman]

NOTE: I am asking this question to learn more about hard drives from the experts. I've asked questions like this a few times in the past and I get told multiple times to just replace the drive. Please don't do that with this thread, I'm trying to learn something here. Thanx

I can identify with that. You don't need a reason to pursue this, other than that you want to do it.

Not knowing anything concrete about it, I'd suspect a possible electronic failure is causing the discrepancy, and wouldn't want to leave the drive in service anywhere that failure would matter.

Here are a couple pages that have some relevant info. Bottom line seems to be, that error isn't an automatic deal breaker, but it does greatly increase the chance of failure.

https://kb.acronis.com/content/9122
https://www.backblaze.com/blog/what-smart-stats-indicate-hard-drive-failures/

 

[Haole Boy]

Where are your threads asking about dot-matrix printers and Betamax players? After all, if you're so curious about obsolete technology, why not go all the way?

Using a hard drive these days is like using a dot-matrix printer.

Well, I've looked for a magic wand to wave over all of my customer's machines to convert them from HDDs to SSDs, but I'm not finding one on Amazon Prime (shipping to Hawaii is outlandish). Do you have a better source?

 

[Haole Boy]

Thanx for all the replies!

(And yes, all files on the drive have been moved elsewhere and the drive taken out of service).

Harry Z

 

[Porthos]

Amazon Prime (shipping to Hawaii is outlandish)

Prime members.
Addresses in Alaska, Hawaii, and Puerto Rico

Shipping Speed Amazon Prime Member Price
Standard Shipping (3-7 business days) Free
Expedited Shipping (2-5 business days) Price varies by item size and weight - as low as $5.99 per item
Priority Shipping (1-4 business days) Price varies by item size and weight - as low as $11.99 per item

 

[Haole Boy]

Prime is pretty much a requirement here in HI for just everyday stuff, not just computer parts.

 

[Porthos]

Prime is pretty much a requirement here in HI for just everyday stuff, not just computer parts.

But you said shipping is expensive. Looks like you could have free shipping if you have a few HD's in stock and order another one each time you sell one. Then you have some in reserve in case you get busy.

 

[Markverhyden]

Well, I've looked for a magic wand to wave over all of my customer's machines to convert them from HDDs to SSDs, but I'm not finding one on Amazon Prime (shipping to Hawaii is outlandish). Do you have a better source?

Contact Hogwarts. Rumor has it they've setup a distance learning program because of the COVID-19 pandemic. Supposedly you even get to keep the wand and broom when you graduate!

LOL!!! Ok. On to the OP. I'm with @phaZed on this. SMART is far from an exact science. I've mentioned before that I've lost count of the number of times I've seen a BIOS POST SMART message, Hard Drive failure imminent, that's been going on for years. That being said it's better than nothing.

When I'm looking for solid hard drive stats the first place I look to is backblaze.com. Those folks have been kind enough to publicly share their data for several years now. Their sample pool is small compared to available drives, all are 3.5", mostly Enterprise class. But you can use any of them in a DT or NAS. In fact that's how I pick my drives for NAS, using their stats.

The link below goes to a discussion about SMART, failures, etc along with link to other great articles. Even has some spiffy looking charts you can show your customers. As it turns out SMART 187 is something they pay close attention to because it does work well as a failure predictor.

https://www.backblaze.com/blog/hard-drive-smart-stats/

 

[Mike McCall]

Excellent reference!

 

[NviGate Systems]

Is this by any chance a Seagate drive? Because Seagate used to have insanely large errors in the SMART data and when pressed for answers Seagate would only say it was "normal". I think it was read errors or uncorrected errors. That was a few years ago, but it's possible the drive you have could be older.

Harry I totally get what you are saying, and I realize that for many they think hard drives are junk and a waste of manufacturing. But SSD's are not the answer to everything and they fail too and often have issues that can be even more annoying than a (slightly) slower HDD.

For USB drives there also exists the very real issue of bad USB bridge chipsets. For big name vendors usually not an issue, but cheapie drives it can be terrible. I recall back in the day (many years ago) that there were some USB IDE chips that were complete crap or couldn't handle above certian capacities properly, even though there were sellers shipping them with big drives. Same thing for SATA, I don't keep up on the new chips but there are some one hung low brands that can cause issues that make it seem like the drive is at fault when in fact it's not.

If you have time and the external drive can be safely taken apart, you might try running more smart tests etc while it's hooked up to internal sata. The USB bridge might be messing with smart.

 

[SlickMinnow]

I can heartily second NviGate's comments. Over the years, I've seen several usb drives apparently fail. Often I would open the case and retrieve the drive, slap it in a machine or another usb adapter or toaster. After a flick and a swish of my Hogwarts wand, the drive worked as normal. Of course WD rained on 'recovery' parade a couple years ago by integrating the usb electronics with the hdd controller electronics on their passport drives - and others. No more separating the actual hdd out for this trick.

 

[Haole Boy]

Is this by any chance a Seagate drive? Because Seagate used to have insanely large errors in the SMART data and when pressed for answers Seagate would only say it was "normal". I think it was read errors or uncorrected errors. That was a few years ago, but it's possible the drive you have could be older.

Harry I totally get what you are saying, and I realize that for many they think hard drives are junk and a waste of manufacturing. But SSD's are not the answer to everything and they fail too and often have issues that can be even more annoying than a (slightly) slower HDD.

For USB drives there also exists the very real issue of bad USB bridge chipsets. For big name vendors usually not an issue, but cheapie drives it can be terrible. I recall back in the day (many years ago) that there were some USB IDE chips that were complete crap or couldn't handle above certian capacities properly, even though there were sellers shipping them with big drives. Same thing for SATA, I don't keep up on the new chips but there are some one hung low brands that can cause issues that make it seem like the drive is at fault when in fact it's not.

If you have time and the external drive can be safely taken apart, you might try running more smart tests etc while it's hooked up to internal sata. The USB bridge might be messing with smart.

Yes, this is a Seagate drive. Interesting point about the chipset possibly being a source of these errors. It would be interesting to take it out of the enclosure and hook it up directly to my bench machine, but I found this when working remotely with the customer. The customer still has the drive someplace on a shelf. I hope they will remember it has issues and not use it again.

 

[labtech]

Would recommend using more than one SMART program to confirm what one says. HDDScan, CrystalDiskInfo, etc
I mentioned this before in your other SMART thread.

FYI: Seagate provides its own crappy tool for testing drives [they use it for warranty check up].
https://www.seagate.com/support/downloads/seatools/seatools-win-master/

They have a YouTube video: https://www.seagate.com/support/downloads/seatools/#knowledge_base

 

[Joep]

Here's your biggest "other" issue:

Raw_Read_Error_Rate     POSR--   082   046   006    -    143398926
Hardware_ECC_Recovered  -O-RC-   082   064   000    -    143398926
Seek_Error_Rate         POSR--   082   060   045    -    171096619

Those should be Zero. Drive is bad, replace. I know you don't want to hear that ;-).

ECC recovery is quite normal in the sense it's happening all the time. In hard drives, SSDs and memory cards. Also this by fzakbar: http://www.users.on.net/~fzabkar/HDD/Seagate_SER_RRER_HEC.html.

Blackblaze monitors these (or used to at least):

SMART 5 Reallocated Sectors Count
SMART 187 Reported Uncorrectable Errors
SMART 188 Command Timeout
SMART 197 Current Pending Sector Count
SMART 198 Uncorrectable Sector Count

If I am not mistaking a RAW value > 0 on any of these and the drive is scheduled for replacement.

 

[Haole Boy]

Would recommend using more than one SMART program to confirm what one says. HDDScan, CrystalDiskInfo, etc
I mentioned this before in your other SMART thread.

FYI: Seagate provides its own crappy tool for testing drives [they use it for warranty check up].
https://www.seagate.com/support/downloads/seatools/seatools-win-master/

They have a YouTube video: https://www.seagate.com/support/downloads/seatools/#knowledge_base

Thanx for the feedback. I have run multiple SMART checking programs in the past against the same (defective) drive, and never received different results. I did not do this here, but will try to remember this for the future.

Harry Z