Could a hard drive's partition "spontaneously" corrupt?

[Coldblackice]

I have a WD Red 3TB drive that's recently been causing problems, with event log saying it's not responding within the timeout period (when I was using Intel's controller driver), and now "The driver detected a controller error on \Device\Ide\IdePort1." (now using Microsoft's controller driver). However, SMART values look completely fine. I also did a chkdsk on it and found no errors or problems. It just hangs for a while, then suddenly "goes" just fine.

I've swapped SATA cables, tried different controller ports, and removed the Intel ICH10R controller driver and reverted back to Microsoft's -- all to no avail. It doesn't seem like "typical" dying hard drive platter symptoms -- it'll hang for a while trying to access it, but then it'll begin transferring data, and doing so just fine, at ~100MB/s.

Could there be any funniness happening with its partitioning, given it's a GPT partition on a non-UEFI system? Could alignment be an issue? Any particular software I could use to check for problems?

 

[drescherjm]

If this is the same drive as your other thread make a copy of your disk then test the original with a badblocks 4 pass read/write test. In that test badblocks will write a pattern to every sector then see if it can read them back then repeat for 3 other patterns. If the drive is dieing it is highly likely to have problems in this test.

 

[Nenu]

At some point a problem will be detected if there is one, this isnt spontaneous, its the way problems are detected.

Do as advised or run the the WD diagnostics tool with a view to RMAing the drive.

 

[dangerouseddy]

a few years ago I had an hd that did the same thing, was ok for a few weeks just freezing for a bit. then it just died completely, I would back up your data. any drive that acts in such a way surely has a platter or head/servo defect that's getting worse imho.

 

[patrickdk]

This is normal behavure when you ask the disk to read a sector, and it has a bad crc. It will keep trying for awhile, before it gives up. On the red's, it is suppost to give up quickly, so like 1second. On normal desktop drives this can be well past a minute of retries as it attempts to read a good copy of that sector.

 

[Coldblackice]

If this is the same drive as your other thread make a copy of your disk then test the original with a badblocks 4 pass read/write test. In that test badblocks will write a pattern to every sector then see if it can read them back then repeat for 3 other patterns. If the drive is dieing it is highly likely to have problems in this test.

Good to know, will do.

At some point a problem will be detected if there is one, this isnt spontaneous, its the way problems are detected.

Do as advised or run the the WD diagnostics tool with a view to RMAing the drive.

As mentioned below, what if it's something like a servo/head defect? Short of a clean-room platter extraction, is there any recourse for recovery? Any tips for stepping through this if it is that kind of problem?

The reason I ask is because it doesn't seem like the "standard" dying hard drive problem, with the inevitable clicks of death. Although maybe that's on account of it being a Red drive. I just don't get how a well-cooled, moderately-used drive suddenly starts dying after 4-5 months.

a few years ago I had an hd that did the same thing, was ok for a few weeks just freezing for a bit. then it just died completely, I would back up your data. any drive that acts in such a way surely has a platter or head/servo defect that's getting worse imho.

Shoot :/ I fear that I may be too late. I've thought it was a controller error, until I finally saw that the drive had been switch to "PIO Mode" in Windows Device Manager, which signaled that it was the problem. What's strange is that SMART was giving it thumbs up, time and again. I realize SMART isn't foolproof, but I thought surely there would've been some number of relocated sectors if there was a problem.

On that note, what did seem to be increasing was "Raw Read Error Rate" -- even though my research seemed to show that this isn't a SMART value to worry about. But given it's the value that stood out from the other drives, I have a feeling it's somehow connected to it, even if secondarily.

This is normal behavure when you ask the disk to read a sector, and it has a bad crc. It will keep trying for awhile, before it gives up. On the red's, it is suppost to give up quickly, so like 1second. On normal desktop drives this can be well past a minute of retries as it attempts to read a good copy of that sector.

Ah, so is that why desktop drives get the click of death, but the Red doesn't? That's what was curious -- there were no clicks -- which led me to believe it must be some atypical error, like the motor or arm servo.

 

[tonyyy]

Good to know, will do.
Ah, so is that why desktop drives get the click of death, but the Red doesn't? That's what was curious -- there were no clicks -- which led me to believe it must be some atypical error, like the motor or arm servo.

All drives can have different types of issues. Could be bad sector, could be motor failure, needle failure, bad board etc.

Parts will just fail.

if you care about your data then always backup.

 

[phis6]

If you're not feeling comfortable with the drive's behavior then back up the data and make a low level diagnostic of your drive using MHDD or HDDscan.