"Windows detected a hard disk problem"

[Yuriman]

I'm running an Intel x-25m G2 as my main drive. Earlier today I managed to fubar my windows installation with too many BSODs from overclocking and had to reinstall. I did a full format and everything is perfectly stable, but Windows keeps popping up with a message telling me to back up my data.

I'm skeptical about my ability to actually damage an SSD due to repeated bluescreens - seems more likely to me that SMART has just accumulated a bunch of errors and Windows is mistaken. How do I make Windows quit bugging me? Can I reset the failed status of the drive, or maybe disable SMART?

 

[mikeymikec]

Is there an Intel SSD checker program?

 

[Nothinman]

SMART is maintained by the firmware on the drive so if it's accumulating errors you still have a problem.

But the delayed write popup you're likely seeing isn't related to SMART, it means that Windows really lost some data during a write or detected a CRC mismatch during a read. In otherwords you almost assuredly have some hardware problem.

Have you looked in the eventlogs to see what the errors contain?

 

[mikeymikec]

+1 Nothinman's comments. I would also use a program (such as SpeedFan) to take a look at the SMART readings to get a better idea of what it might be complaining about.

 

[Yuriman]

Windows event ID 52: "The driver has detected that device \Device\Harddisk1\DR1 has predicted that it will fail. Immediately back up your data and replace your hard disk drive. A failure may be imminent."

Opening up HD Tune, I have:

Marked in yellow:
SMART ID 05 (Reallocated Sector Count) Current: 100, Worst: 100, Threshhold: 0, Data: 78, Status: Ok

Marked in red:
SMART ID B8 (Unknown Attribute) Current: 1, Worst: 1, Threshhold: 99, Data: 438, Status: Failed


Can unexpected power-offs really cause physical damage to an SSD? Or is the timing just really bad?


Wikipedia's SMART article has this to say about 0xB8 (not sure if it's the same error):

"This attribute is a part of Hewlett-Packard's SMART IV technology, as well as part of other vendors' IO Error Detection and Correction schemas, and it contains a count of parity errors which occur in the data path to the media via the drive's cache RAM."

 

[Nothinman]

Windows event ID 52: "The driver has detected that device \Device\Harddisk1\DR1 has predicted that it will fail. Immediately back up your data and replace your hard disk drive. A failure may be imminent."

Opening up HD Tune, I have:

Marked in yellow:
SMART ID 05 (Reallocated Sector Count) Current: 100, Worst: 100, Threshhold: 0, Data: 78, Status: Ok

Marked in red:
SMART ID B8 (Unknown Attribute) Current: 1, Worst: 1, Threshhold: 99, Data: 438, Status: Failed


Can unexpected power-offs really cause physical damage to an SSD? Or is the timing just really bad?


Wikipedia's SMART article has this to say about 0xB8 (not sure if it's the same error):

"This attribute is a part of Hewlett-Packard's SMART IV technology, as well as part of other vendors' IO Error Detection and Correction schemas, and it contains a count of parity errors which occur in the data path to the media via the drive's cache RAM."

If Wikipedia is right you may have a cabling issue. But given that it's a non-standard attribute you should probably check with Intel to see what it really means.

 

[razel]

As others have stated, windows is giving you a heads up about your drive. So to determine the drive's health, it's probably best that you run the Intel's very own SSD Toolbox instead of third party utilities. Intel would know best how to interpret their own SMART values.

What I would do? Reboot, write down to take a picture of your BIOS settings. Reset your BIOS to default, run everything back to stock then turn the machine off, unplug, swap SATA cables to force yourself to reseat the SATA data and power cables, then run the Intel SSD Toolbox. You need a stable environment before moving on.

 

[Yuriman]

Couldn't figure out why I couldn't get the Intel Toolbox running - I had my SSD on the ASMedia SATA3 controller on my board.

Moved it over to the Intel controller and loaded up the SSD Toolbox.

I ran a full diagnostic scan which it says checks every LBA for a hard error, and also checks for data integrity, and it found no errors. I updated the firmware (it was still running its 2009 firmware) and then ran the full diagnostic which again completed successfully.

Google'ing around, someone mentioned that End-to-End Error could be the number of NAND cells that have died. I don't know if 438 is a lot but the drive claims 98% life remaining.

I'll keep it on the Intel controller and keep an eye on that value, but does anyone know if there's a chance that I might still be eligible for RMA?

 

[tweakboy]

What percent does Crystal show ? Also SSD's can not die but can degrade and stop working..,, its being recognized by BIOS right ? Theres hope. But that last error in red is no good. Def you need to backup your stuff my friend. then try chkdsk.. let us know gl

 

[Yuriman]

What percent does Crystal show ? Also SSD's can not die but can degrade and stop working..,, its being recognized by BIOS right ? Theres hope. But that last error in red is no good. Def you need to backup your stuff my friend. then try chkdsk.. let us know gl

I'm currently running Windows on it and aside from Windows popping up with messages that say that the drive is at risk, I haven't had any data problems yet. Scan disk and the Intel toolbox check haven't found any bad sectors or corrupted data.

By "Crystal" do you mean Crystal Disk Info?

It shows drive status as "bad" but with 100% reserve space still remaining.

EDIT: Ran Crystal Disk Mark to see if there might be any slowdown, and I think it's actually benching better now with the new firmware than when it was new. Remember, this is an 80GB drive.

Sequential read: 263MB/s
Sequential write: 79MB/s
512k read: 201MB/s
512k write: 87MB/s
4k read: 25MB/s
4k write: 45MB/s
4k QD32 read: 181MB/s
4k QD32 write: 70MB/s

 

[razel]

Depending on how good you are at ignoring things then I'd continue to ignore the Windows warnings. I know that for Win7 after a few warnings it will give you a choice to stop bothering you.

The good is you passed the tests. Estimated drive life is still close to 100%. What shocked me is the writes and bad shutdowns. You wrote almost 9TB and had over 600 unsafe shutdowns (like pulling the power plug). My sister's laptop with a bad battery and the same Intel G2 barely has 10 unsafe shutdowns. Unless your system is beyond unstable, I'd just chalk it up as a lesson to do the overclocks off an HDD you care less about until you find stable settings.

I suppose you can try to claim warranty, if you are good at crying, but I think you'd be taking advantage of Intel.

 

[tweakboy]

I'm running an Intel x-25m G2 as my main drive. Earlier today I managed to fubar my windows installation with too many BSODs from overclocking and had to reinstall. I did a full format and everything is perfectly stable, but Windows keeps popping up with a message telling me to back up my data.

I'm skeptical about my ability to actually damage an SSD due to repeated bluescreens - seems more likely to me that SMART has just accumulated a bunch of errors and Windows is mistaken. How do I make Windows quit bugging me? Can I reset the failed status of the drive, or maybe disable SMART?

Many times its just saying that.

I never knew Winblows would pop up and say back up your stuff. Must be serious,, I would listen to uncle billy. What does it say in Crystal ? that will tell us a lot.

 

[Yuriman]

Many times its just saying that.

I never knew Winblows would pop up and say back up your stuff. Must be serious,, I would listen to uncle billy. What does it say in Crystal ? that will tell us a lot.

By "Crystal" do you mean Crystal Disk Info?

Not sure which Crystal you'd like me to run.

As for the writes, I've had the drive for 3 years now and it has been my OS and game storage. For the first 2 and a half years of its life I used it in RAID with a second X25m, which is now in my wife's laptop. The were on average about 60-75% full, and I did shuffle things through them a lot, and t has my page file and hyberfile.sys. Since I got the drive I haven't really turned the PC off, any shutdowns were either for updates, or were from power outages or tinkering with overclock settings - the total power-on hours add up to ~2.96 years. I'm surprised it still reports to have 98% of its life left.

Since you put it that way, I won't try to claim warranty. A new SSD is pennies these days, I may well end up needing more space before it finally dies.

EDIT: Looked up the specs, I'm at ~26,000 hours of approximately 1.2 million hours of expected lifetime. That strikes me as funny. This thing is rated for a century and a half of on-time.

Also assuming the flash is good for 10k erase/flash cycles, it would theoretically last up to 800TB of total writes, so I'm at around 1/100 of that.

 

[Insert_Nickname]

If Windows reports a SMART error, replace that drive ASAP. SMART does not report an error unless its serious. Don't even bother trying to figure out what is wrong, the drive is most likely on its last legs...