Slow I/O on WD800JD

Joined
Oct 24, 2001
Messages
856
We're working on 3 machines (all built identically several weeks ago) that appear to have some significant problems with I/O. Bart's Stuff Test and HDTach report the max I/O speeds at about 3 mb/sec. When the inital system came in with this problem, I swapped in our test drive (Seagate 7200.7 ST380013A8) and had no problem loading Windows and running the benchmarks. The drive was performing basic I/O at about 50 mb/sec. We went ahead and swapped in a replacement drive in this system and chalked it up to a defective drive. Now we have 2 more systems of the same order that are exhibiting the same symptoms. It appears that this batch of drives might not have DMA enabled on the hardware of the actual drive, but I'm not entirely sure how to check. Here are the system specs:

Intel D915GUXLK
P4 540J
SPI 350W PSU
2x 256MB DDR2-533
XP Pro
WD800JD (I'll get the exact model number if you need it)
ASUS EN6200GE (GeForce 6200 PCI-E)


Let me know if you guys have any ideas. I've scoured the web trying to find out if WD released a bad batch of drives or if there is any aparent compatibility problem with this drive and have had no luck finding anything. I didn't search the forums because I'm an avid reader and don't recall anything like this. :D

Thanks.
 

FlatLine84

[H]ard|Gawd
Joined
Apr 7, 2005
Messages
1,521
I know there are a crapload of those drives on the market. I don't know if there is a bad batch. I'm currently running 2 in the dreaded "RAID-0" array and they both preform fine, better than my single 120gb seagate.
 
Joined
Oct 24, 2001
Messages
856
The other 9 systems in this order are working fine and the replacement WD800JD in the first system is working fine. These hard drives were all purchased from our supplier at the same time, so I'm not terribly sure about a bad batch.
 
Joined
Oct 24, 2001
Messages
856
FlatLine84 said:
I guess I would just try swapping it with one of the other WDJ800s.
We know this solves the problem but I need to understand the initial problem so that we don't make the same mistake in the future...
 

FlatLine84

[H]ard|Gawd
Joined
Apr 7, 2005
Messages
1,521
If it's a bad drive, it's a bad drive, there really isn't a mistake made, drive can come already bad. If one of the other WD's of the same model work, then I would say it's just that drive in particular. Just RMA it directly to WD, they're really good about it.
 
Joined
Oct 24, 2001
Messages
856
A failure rate of 3 drives out of 12 is very atypical though possible. Ultimately, the drives will be replaced if found to be defective or if no solution is found. I'm here to exhaust all other alternatives.
 
Joined
Oct 24, 2001
Messages
856
Here's an update if anyone is paying attention to this.

We have started to uncover more and more of these drives in this order doing I/O at about 3 MB/sec. The original drive we replaced last week is also experiencing the same problems. This leads me to believe that this is a degrading problem because the replacement drive worked fine in my labs. We are able to confirm that replacing with a Seagate drive fixes this issue.

Here's the real kicker: the drive reads in Linux at 56 MB/sec. This was done by doing a basic "dd" command and monitoring the I/O traffic using "vmstat."

So, my conclusion at this point is that there is some compatibility problem between Windows and the SATA controller (and hence the drive) or a compatibility problem between Windows and the drive (sans the controller). We are using the latest drivers and BIOS -- everything is up-to-date.

Any ideas?
 
Top