HDD Connection Issue

othelios

Honorable
Feb 23, 2012
4
0
10,510
Hi,

I have a OCZ Vertex 2 SSD, which has Windows 7 64-bit installed on it, and 3x 1TB HDD's for storage. One of the HDD's, a WD Caviar Black, is used for my Windows user accounts, temp folders and My Documents (I used this guide: http://www.overclock.net/t/1156654/seans-windows-7-install-optimization-guide-for-ssds-hdds) as constantly writing to the SSD wears it out more and I want the SSD to last for as long as possible.

Now the problem I have is that the WD Caviar Black HDD will occasionally (about once a week) lose its connection. Usually I first notice it when I'm playing some music or watching a video. What'll happen is that the media will stop playing and either the program will crash/freeze or it'll give me an error saying it can't find the files. Then Windows will start lagging as some of it's files (e.g. the user folder) are located on that drive. The connection issue seems to stop and start. I turn off the computer (which takes a while as it freezes during the process) and I usually open up the case and un-plug the HDD sata cable and then plug it back it. This seems to fix the problem, but I'm unsure (I've also tried using a different SATA cable and port). Sometimes when booting up Windows will perform a chkdsk and will find several corrupt files. These files are usually related to what I was doing at the time.

I've also run WD's HDD diagnostic tool and whilst it does lag sometimes, it gives the HDD a pass. The only failure that I have witnessed is when I first ran the diagnostic tool and it said that SMART had failed. I ran the test again and it passed. I assume that it failed due to the connection problem where the drive is unaccessible for a few seconds.

The drive is quite new (bought and installed in August 2011) and replaced a Seagate drive which was having the same problem (which I thought was due to the age of the drive). I was able to retrieve my files using a Ubuntu Live CD with no problems.

What I'm asking is does anyone know what I'm on about and can they offer a possible solution? It seems unlikely that a 2nd HDD has become faulty with exactly the same problem. My other 2x 1TB HDD's don't seem to have this problem (possibly because they are powered down most of the time and not in use). Could it be the SSD/HDD setup that I have used? I've tried Google but none of the results match my problem.

Thanks and apologies for the wall of text.
 

othelios

Honorable
Feb 23, 2012
4
0
10,510
Don't know if this helps, but when chkdsk is running it deletes a lot of indexes and recovers a lot of orphaned files.
 
It sounds like your drive may have head or media problems. Data LifeGuard is essentially a pass-or-fail diagnostic. Try a more comprehensive SMART diagnostic. Look for reallocated, pending, or uncorrectable sectors.

HD Sentinel (DOS / Windows / Linux):
http://www.hdsentinel.com/

HDDScan for Windows:
http://hddscan.com/

See this article for SMART info:
http://en.wikipedia.org/wiki/S.M.A.R.T.

Also try a read benchmark in HD Tune:
http://www.hdtune.com

Look for large dips in the data rate, and excessive scatter in the access time graph.
 

othelios

Honorable
Feb 23, 2012
4
0
10,510
Thanks for replying.

I used HD Tune and under the Health tab everything was fine except there was a warning for Ultra DMA CRC Error Count. in the Data column it read '4'. I then performed an error scan which completed successfully with 0 errors. After the scan the Ultra DMA CRC Error Count was at '45'. I googled the error and several posts suggested that this was sometimes normal and was only a concern if it went up. Since it did then this error is apparently due to a faulty cable SATA port/controller. I've replaced the SATA cable and used a different SATA port.

I've since run several tests using HD Sentinel, with 0 errors and it says my drive is fine (no warnings for Ultra DMA CRC Error Count like HD Tune was reporting).

HD Scan did show an error for Ultra DMA which was "0000000000-002D" compared to straight 0's for the other drives.

I also did a benchmark using HD Tune. I ran 3 tests and uploaded the results here:

http://imgur.com/a/I2ijl#0

Only the first result shows a large dip. If oyu could offer your opinion on the results, that'd be great.