hard drive problem

[tviceman]

Computer wouldn't boot this morning. I got an error message before windows 7 starts loading saying disk cannot be read. Please press ctrl + alt + del. Bios recognizes hard drive, and switching sata ports did not help.

I inserted the windows 7 disk and ran the repair install util, it said one of the windows directories was corrupt and to run the chkdsk command. Well, after that I told it to repair the install, and it went BSOD on me. So I formatted the drive, as it was a fairly new install anyway, and right now I'm reinstalling Windows 7. What is the best way to check and find out if the drive is going bad? The hard drive itself is less than 6 months old, fyi.

 

[Dahak]

i would run the manufactures diagnostics on it first. if the drive is really starting to fail that should show it

 

[tviceman]

I reinstalled windows 7 successfully and am now running a diagnostic on it. Is there a chance that overclocking your CPU could lead to data corruption errors? Every PC I've owned in the past I had overclocked to high hell and had very few issues.

 

[corkyg]

It most likely depends on the degree of over clocking. Thgt usually does not cause data corruption. That is more likely caused by erratic voltage or power surges, etc. Anything that interrupts a write can cause corruption. Also, the drive could be going South.

 

[Russwinters]

Dont run the drive as your primary drive, run it as a slave/secondary.


Then run www.hddscan.com <- that software


perform a linear scan test first, and watch the "map view" for delays.


A new drive should have no sectors that read over 10ms, let alone 5ms.

 

[tviceman]

It most likely depends on the degree of over clocking. Thgt usually does not cause data corruption. That is more likely caused by erratic voltage or power surges, etc. Anything that interrupts a write can cause corruption. Also, the drive could be going South.

Could faulty RAM lead to corrupt hard drive files? I bought new geil RAM for it and was getting BSOD - so I just RMA'ed it and am overnighting some corsair memory now.

 

[pcgeek11]

Could faulty RAM lead to corrupt hard drive files? I bought new geil RAM for it and was getting BSOD - I just RMA'ed it and am overnighting some corsair memory now.

Yes it can. Corrupted memory when written to the HD = corrupted HD Data.

 

[tviceman]

Yes it can. Corrupted memory when written to the HD = corrupted HD Data.

Thank you for this. Hopefully the new ram will correct this problem.

 

[RebateMonger]

A new drive should have no sectors that read over 10ms, let alone 5ms.

Russ,

I hate to start a new thread for this, but I finally tried out HDDScan and I'm curious about the results. I'm curous what you think of the results.

I have three 1 TB Hitachis (non-Enterprise) in a server. They are nine months to 1.5 years old and run 24/7, all in non-RAID mode.

About nine months ago, the server fell, running, from the seat of a chair onto the carpeted floor. I expected all three disks would be trashed. But after the fall, I ran "chkdsk /r" on all three disks and no errors were found. SMART still gives no warnings and no reallocated sectors.

I'm not sure the best way to post my HDDScan results. But I'll start with the Report table that seems to show all reads longer than 50 ms. The first disk, "Disk A", is a later three-platter model that was installed AFTER the accident. The other two disks, "Disk B" and Disk C", are original five-platter models that were running in the server when it hit the floor.

Disk A
Block start at 179888640 time 203ms
Block start at 179888896 time 217ms
Block start at 948414464 time 195ms
Block start at 948414720 time 214ms
Block start at 1587256576 time 203ms
Block start at 1587256832 time 210ms
Block start at 1842479872 time 112ms


Disk B
Block start at 456236288 time 180ms
Block start at 456236544 time 167ms
Block start at 456343040 time 195ms
Block start at 456343296 time 194ms
Block start at 1026616064 time 177ms
Block start at 1026616320 time 169ms
Block start at 1026711552 time 195ms
Block start at 1026711808 time 195ms
Block start at 1484609280 time 86ms
Block start at 1512924160 time 181ms
Block start at 1512924416 time 162ms
Block start at 1513003776 time 187ms
Block start at 1513004032 time 195ms
Block start at 1900054528 time 182ms
Block start at 1900054784 time 168ms
Block start at 1900114944 time 196ms
Block start at 1900115200 time 203ms


Disk C
Block start at 95370752 time 154ms
Block start at 95503616 time 78ms
Block start at 95661568 time 74ms
Block start at 95661824 time 60ms
Block start at 95711488 time 82ms
Block start at 96627456 time 179ms
Block start at 97051904 time 76ms
Block start at 97052160 time 69ms
Block start at 97113856 time 75ms
Block start at 97189632 time 71ms
Block start at 97419776 time 82ms
Block start at 98270208 time 65ms
Block start at 141269248 time 107ms
Block start at 161586688 time 65ms
Block start at 215345152 time 211ms
Block start at 215345408 time 204ms
Block start at 215445248 time 171ms
Block start at 215445504 time 169ms
Block start at 310911744 time 57ms
Block start at 351888384 time 57ms
Block start at 351891456 time 74ms
Block start at 351892224 time 57ms
Block start at 461315584 time 57ms
Block start at 608157696 time 74ms
Block start at 751046912 time 67ms
Block start at 810802688 time 205ms
Block start at 810802944 time 194ms
Block start at 810895616 time 167ms
Block start at 810895872 time 167ms
Block start at 889400576 time 57ms
Block start at 889400832 time 54ms
Block start at 889401088 time 62ms
Block start at 889401600 time 57ms
Block start at 1022143744 time 58ms
Block start at 1022144512 time 66ms
Block start at 1022145280 time 58ms
Block start at 1151734272 time 59ms
Block start at 1151737344 time 58ms
Block start at 1151738112 time 58ms
Block start at 1151740672 time 66ms
Block start at 1273979136 time 58ms
Block start at 1273979904 time 66ms
Block start at 1273980160 time 55ms
Block start at 1274648320 time 50ms
Block start at 1274650624 time 50ms
Block start at 1276204800 time 74ms
Block start at 1331735040 time 211ms
Block start at 1331735296 time 194ms
Block start at 1331812096 time 171ms
Block start at 1331812352 time 168ms
Block start at 1391401472 time 58ms
Block start at 1391403520 time 74ms
Block start at 1391404288 time 58ms
Block start at 1391406592 time 50ms
Block start at 1502076416 time 50ms
Block start at 1502077952 time 58ms
Block start at 1502079232 time 50ms
Block start at 1502080512 time 66ms
Block start at 1502081536 time 66ms
Block start at 1607334400 time 66ms
Block start at 1607335680 time 50ms
Block start at 1607336448 time 74ms
Block start at 1607337216 time 66ms
Block start at 1706969856 time 51ms
Block start at 1706971392 time 50ms
Block start at 1706972416 time 50ms
Block start at 1706973696 time 74ms
Block start at 1706974208 time 60ms
Block start at 1706977280 time 50ms
Block start at 1706980096 time 66ms
Block start at 1706980352 time 57ms
Block start at 1708773888 time 82ms
Block start at 1758333440 time 236ms
Block start at 1758333696 time 195ms
Block start at 1758393600 time 166ms
Block start at 1758393856 time 162ms
Block start at 1800162816 time 58ms
Block start at 1800164352 time 52ms
Block start at 1800166400 time 74ms
Block start at 1800166656 time 51ms
Block start at 1800169472 time 50ms
Block start at 1876086272 time 50ms
Block start at 1876087552 time 58ms
Block start at 1886732288 time 84ms
Block start at 1886734080 time 60ms
Block start at 1886735360 time 58ms
Block start at 1886736640 time 83ms
Block start at 1886737664 time 60ms
Block start at 1886738944 time 50ms
Block start at 1886739712 time 58ms

My question is: Are these "normal" results for a set of working 1 TB disks? Or should I be worried? One of these three disks (either "Disk B" or "Disk C") is the boot disk, running Win Server 2008 and Hyper-V". I don't know if that affects the HDDScan results.

Thanks,
RebateMonger

 

[Emulex]

i don't really trust stats when in hypervisors - the measurement tools may not get an allocated time slot if other vcpu's are running.

 

[Russwinters]

Results may not be telling the truth, if you are using those drives currently in the system they are probably accessing other content.


The best way to use HDD scan is to make sure the drives are not going to be used by the system, so they are just sitting idle.

 

[RebateMonger]

Results may not be telling the truth, if you are using those drives currently in the system they are probably accessing other content.

For these tests, I turned off all the virtual machines. The only files on the two non-System disks are virtual disks for Hyper-V machines (which were all turned off). So only the host OS disk (running Server 2008) should have been doing anything.

 

[Russwinters]

If thats the case it looks like those areas of the disk are starting to slow down a lot.


I wouldn't suspect heads though, since it is not consistent throughout the WHOLE disk, only in isolated areas.

This looks just like standard media degradation, many of those sectors should be getting reallocated pretty soon.


Keep your backups updated =)

 

[RebateMonger]

I figured that when that server hit the carpet, those five platters each took a hit on at least one side. I throw away disks if they show any reallocated sectors, so hopefully they'll hang in there.

The third Hitachi 1 TB disk that was in the server when it hit the floor had a problem when I turned it on after several months of non-use. A virtual machine image on it got corrupted. The disk passed a disk-wide "chkdsk /r" and a HDTune scan, but the virtual disk's partitioning informaiton apparently got damaged. I was able to grab data from it (maybe I ran GetDataBack, I don't recall). I didn't look too deeply, since I had a full backup of the virtual machine.

The corruption might have been a physical result from the accident. Or it might have just been a software issue. I really couldn't detect anything physically wrong. But I think I'm pretty lucky if I dropped a server with three 1 TB disks and got away with it. I'm thinking maybe I wasn't as fortunate as it seemed at first. That's OK. It's pretty dumb to drop a running server.

Thanks for your comments, Russwinters and Emulex.