Question for anyone running WD 640 blues in RAID 0

S

Sabrewulf165

2[H]4U
Joined
Jun 23, 2004
Messages
2,974
1) Any chance you could post an hdtune read bench screenshot, or at least your min/max/avg?

2) Do the 640 blues still have that problem with TLER that older non-RE WD drives in RAID always had?
 
HDTune.jpg


(3) WD6400AAKS in RAID 0 on Asus P5Q-E (ICH10R)

I'll be honest, I never really paid any attention to the TLER issue. That being said, I haven't changed the TLER setting of the drives since I got them and I haven't had any problems - going on 3 months now.
 
What TLER issue are you referring to?

TLER is only meant for drives that are going to be installed in a fault tolerant RAID array such as RAID 5 or 6. If you use TLER in RAID 0, you significantly increase your chance of data corruption. Read the Wiki here.

The only thing that TLER does is limit the on board error recovery to 7 seconds. After that, the RAID array will handle the bad sector and data recovery. If there is no data redundancy, like RAID 0, that will result in data corruption since there is no way to rebuild the bad sector in those instances.

Here is why I know a bit about this issue. I recently bought a Thecus N5200 NAS box. I bought 3 WD GP 1TB desktop drives to go in it. During the initial format for the RAID 5, 1 of the drives came up with an error. I ran the drive diagnostics on all 3 and they all came up clean. I started the RAID build again, and the same drive failed again. I did not know about TLER at this point. I RMAd the bad drive, and while waiting for the replacement, I learned about TLER. When the replacement drive arrived, I enabled TLER on all 3, since they were going into a RAID 5 array, and they have worked fine for about 6 months now.

TLER is a very misunderstood technology.

Don
 
Splendid, thanks for the caps guys, they look great :)

@ Don : a couple years ago when I was running WD drives in RAID on an Intel "soft" controller, there was a known issue where drives would randomly drop out of the array after a reboot or powercycle which supposedly had something to do with consumer-level WD drives and TLER (or lack thereof, I forget which). From what I can remember, the solutions to this problem were (1) run some kind of tool that disabled TLER on the drives (or maybe ENabled it, again I can't remember) (2) use non-WD drives, or (3) use WD "RE" or enterprise-class drives.

Sorry my memory is a bit hazy on the whole thing, but the bottom line is the last time I tried to run consumer-class WD drives in RAID 0, the stripeset would crap itself on almost every reboot. Just wondered if that was still an issue or not.
 
Yea, it sounds like you may have just had a flaky drive. Turning on TLER in RAID 0 is not the answer for that problem though.

Read the first paragraph of the Wiki article. It is not a WD problem. All hard drives will drop out of an array if they take too long to recover from bad sectors. The other manufacturers have the same fix, they just call it something else.

Don
 
You must log in or register to reply here.
Back
Top