HDD unmounting itself (windows 7)

E

EdZ

Gawd
Joined
Jan 28, 2008
Messages
771
This one has me stumped. My system has been running fine in it's current configuration for over a year now. This evening, during a file copy (newly created file), two of my partitions (on the same drive) disappeared (from explorer and from drive manager) causing the copy, and explorer, to hang. On reboot, the drive was back, but attempting to copy the file cause it to suddenly unmount again (listening closely, it sounded like the heads parked and the drive shut down). But this time, a simple reboot did not bring the drive back, but a power-off power-on cycle does. I ran checkdisc on the drive, and it reported no errors. I can copy some files off the drive, but others will cause the drive to disappear on access. This is the state the system is currently in: a fresh boot will have the drive show up fine. Hitting certain files will cause the drive the drive top suddenly shut itself down. If a warm restart is performed, the bios will wait at post where the bios would normally say "SATA 4 (name of drive)", never proceeding, and with no error message (as if it knows a drive IS there, and is expecting the drive to respond). A power cycle will bring the drive back up.


My two vague guesses are:
1) the drive is physically failing in an extremely odd manner. If so, i need a program that can copy off as much data as possible without hitting whatever bad areas that force the drive to suddenly unmount
2) the drive is physically fine, but some quirk caused at least one file to be written to a bad sector, and any time a read is attempted on that sector the drive for some reason shuts itself down.


I can think of no reason why the drive would physically power down so suddenly on reading only certain files, why checkdisk would be able to scan the drive without any errors (or triggering a powerdown), or why the bios would pause when it reaches the drive rather than failing with an error message.


Has anyone experienced a similar problem before?
 
Has anyone experienced a similar problem before?
Click to expand...

Yes. I have seen this dozens of times at work.

My bet is the drive is going bad.

Install CrystalDiskInfo and see if it says Caution or Bad for the drive.

Note: CrystalDiskInfo reads the drives SMART data. In this data there are a few statistics that the drive records that indicate physical problems with the drive.
 
Last edited:

Note that the HDD with the 'caution' light is not the drive failing.

::EDIT::
And now the problem has gotten worse. A regular boot will bluescreen about 20 seconds after reaching the desktop. A safe mode boot will run for longer before locking up (but not bluescreening).

And even worse, physically disabling (unplug sata & power) causes a "disc read error" on attempting to boot.
 
Last edited:
Note that the HDD with the 'caution' light is not the drive failing.
Click to expand...

Please post the pic of that drive if you can get it.

If not I can tell you a method if you have a recent linux livecd.
 
cdi2.png

I am now booting from the 'caution' drive, which has an old but still functioning copy of Vista on it. The previous failing drive is now disconnected. The Windows 7 drive (an SSD) is reading as 'RAW' in Disc Management. Attempting to boot from the SSD directly results in the "disc read error". Attempting to boot from the old BCD on the vista drive (which still has an entry for windows 7) gives a 0xc000000f Required device is Inaccessible error. Using the repair installation tool on the Windows 7 DVD has no apparent effect.
I've tried running testdisk (from the vista install), which recognises an NTFS partition. However, it cannot access it. I ran a a 'deeper scan' and tried to write to the partition table, but got a "could not write to partition table" error. Then the drive unmounted itself and disappeared.
 
For D:

The C5 and C6 raw error counts represent sectors that the disk lost or corrupted data on.

05 represents 14 hex (20 decimal) bad sectors that were successfully reallocated to good sectors. I would only worry about this number if the number increased daily or something.


So the problem drive is a SSD?
 
The second drive, yes. I just can't think why removing the initial failing drive (a regular HDD) would cause the SSD to suddenly fail too. In a desperation bid I tried reconnecting the failing HDD, with no effect on the SSD.
 
I just can't think why removing the initial failing drive (a regular HDD) would cause the SSD to suddenly fail too.
Click to expand...

I do not think the problems are related. Can you try the manufacturer tools on the SSD drive? I can no help much with that since I only have an Intel SSD and I have never used the utilities being that I use my SSD under linux and there is not a linux version of Intel toolbox.
 
Crucial do not appear to have any data recovery programs available on their site. The timing of 'remove failing drive -> other drive immediately fails' seems too perfect for it to be mere coincidence, especially as I've had no problems with the SSD until now.
 
Performed a CMOS reset: no effect. I dug out an old IBM desktop that thankfully had a SATA port. Put the SSD in there, still shows as RAW. Ran testdisk with a simple analyse, it successfully saw an NTFS partition, but could not view any files "can't open filesystem, filesystem seems damaged". Ran a 'deeper analyse', it again found an NTFS partition, but still couldn't view files. Wrote the partition table to disk, restarted, still no change.
testdisk.png

If it was the drive controller on one motherboard causing screwy effects with SATA drives this all could almost make sense, but on a different machine too? Something did a real number to this drive, and did it at the exact moment another failing drive was removed. And I know for certain it wasn't static discharge from me: my hand's didn't even go near the same drive cage.
 
Ah, progress of sorts. Loaded the initial 'failing' drive into the other computer, and it appears to work perfectly. The SSD is still as totally hosed as ever. My suspicions are definitely on the motherboard (or at last the SATA controller) doing something very screwy. And damaging the partition table of my SSD in a way that appears to prohibit recovery.
 
EdZ said:
Note that the HDD with the 'caution' light is not the drive failing.

::EDIT::
And now the problem has gotten worse. A regular boot will bluescreen about 20 seconds after reaching the desktop. A safe mode boot will run for longer before locking up (but not bluescreening).

And even worse, physically disabling (unplug sata & power) causes a "disc read error" on attempting to boot.
Click to expand...

This really sounds like bad RAM. Random crashes with BSODs normally your memory is going bad. Run Memtest outside of the OS and find out which stick(s) of RAM are bad.

In fact, I'm 99.9% sure this is your memory going out on you, not the HDD or the controller.
 
You must log in or register to reply here.
Back
Top