WD replaced my TLER-enabled drive with one that won't enable TLER

I run a RAID-5 setup with my Thecus N4100Pro NAS. I have 4 Western Digital 1TB Caviar Black drives (WD1001FALS). I bought all 4 pretty soon after the line came out, and I was able to enable TLER on all 4 drives. The RAID array has been great for about 2 years, but about 2 weeks ago, one of the drives died on me. I did an RMA through WD and got a replacement drive only to realize that I couldn't enable TLER on it. I had a long conversation via their support website ranting to WD about how I need a TLER-enabled WD1001FALS, but they just blew me off.

Anyway, I figured I'd try it, but as I expected the RAID rebuild wouldn't even get to 4% complete before the drive would drop out. Then, the replacement drive actually started throwing SMART errors, too, so I had to replace it. The 2nd replacement arrived today, and as I expected, it also won't support TLER.

My plan is to not even try to go through the rebuild process. I want to sell the replacement drive since it's brand new and pick up a drive that I know will work. Now, I know that the RE line is designed specifically for RAID and supports TLER out of the box. But I also saw the price tag.

So my question is this: what drive(s) do you guys recommend I replace my WD1001FALS with? I'm also not opposed to picking up 4 1.5TB or 2TB drives and growing the RAID.

Thanks!

those drives will sell well even though they are about two years old. Personally I would go for new drives as there is no chance that you will get a tler drive from wd.

Since you are running a raid array you need 7200rpm and right now Hitachi is the only company that makes cheap 7200rpm drives in the size of 1.5tb and up so your options are very limited. They recently released a new drive series recently which looks good and is not 4k so you wont run into any 4k issues. Personally I would go for a few 2tb 7k3000 drives.

It doesn't support TLER? It is forced on by default at the factory.

If you are trying to use a tool to change it, it will fail. The reason it is failing to rebuild is either the drive is faulty or TLER is causing the issue. TLER is very bad in RAID arrays when set too high, which that new drive definitely does.

thideras said:

It doesn't support TLER? It is forced on by default at the factory.

If you are trying to use a tool to change it, it will fail. The reason it is failing to rebuild is either the drive is faulty or TLER is causing the issue. TLER is very bad in RAID arrays when set too high, which that new drive definitely does.

Click to expand...

I think you're a little confused. So TLER is necessary for hard drives in a complex RAID array; without it, they may fall out of the array. The primary difference between Western Digital's Caviar Black line and their RE line is that the RE line supports TLER out of the box. The Caviar Black line has it disabled out of the box. WD's first batch of 1TB Caviar Black drives supported the WDTLER tool which would enable TLER (set to 7 seconds) on the drives. However, around late 2009, all WD Caviar Black drives no longer support the mod (so it is impossible to enable TLER on them).

I have a 4-disc RAID-5 array. 3 of the discs are my original drives that I was able to mod using WDTLER, but the 4th died, and the replacement no longer supports the mod. The replacement kept falling out of the array, and I don't want to deal with that. I'd rather get drives that are known to work in a complex RAID array.

I'm leaning towards one of the Hitachi drives (probably the 2TB 7K2000).

Sort of.

TLER = Time Limited Error Recovery.

Time limited is the important part. Their home drives are set extremely high (90 seconds) and their RAID drives are set lower (7 seconds). If a drive stops responding for more than 15 seconds, there is a high chance that it will be dropped from the array. Either TLER needs to be set to a lower value or disabled. The reason that your drive is dropping out, just as I said in my previous post, is either the drive is actually faulty or it is taking too long for the recovery. Since you can't disable or set TLER values on the newer drives, you shouldn't use them in RAID arrays.

Believe me, I've done a lot of research on this subject.

http://www.overclockers.com/forums/showthread.php?t=652911

Oh, I realize fully what it means. The drive isn't necessarily faulty, though, if it takes longer than 7 seconds. It could just be that the sector trying to be written to is corrupt. A hard drive can have a decent amount of sectors be bad without the drive as a whole being affected (this is why SMART has an upper limit on bad sectors before saying the drive is faulty). If just one of these is bad, though, it can fall out of a RAID array which shouldn't be allowed to happen.

[edit]
After research, I'm gonna pick up some of the Hitachi drives. Looks like they don't need any sort of TLER-like modification to work in a RAID array. Thanks guys!

Fangs404 said:

The drive isn't necessarily faulty, though, if it takes longer than 7 seconds.

Click to expand...

I know, but I was simply saying that the drive will go offline as long as it needs to fix the error or when it hits the TLER value. In this state, it will not respond to any command and some RAID controllers may drop the hard drive and mark it as failed. In your case, the SMART errors indicate there is something else wrong with the drive. I've seen reports of TLER running without issues in a larger array and I've heard of non-stop issues with three drives. It really seems to be hit-or-miss, assuming your drive is functional. I'm not sure what RAID controllers were used in either case, so I can't really compare.

Either way, if I was in your scenario, I would ditch those drives for ones that didn't have TLER, TLER equivalent or simply get ones that were made for RAID arrays (7s TLER). Any drive they send you is going to be a newer one with TLER values locked and it might cause issues. Knowingly keeping a problematic drive in an array defeats the purpose of using RAID in the first place.

thideras said:

I know, but I was simply saying that the drive will go offline as long as it needs to fix the error or when it hits the TLER value. In this state, it will not respond to any command and some RAID controllers may drop the hard drive and mark it as failed. In your case, the SMART errors indicate there is something else wrong with the drive. I've seen reports of TLER running without issues in a larger array and I've heard of non-stop issues with three drives. It really seems to be hit-or-miss, assuming your drive is functional. I'm not sure what RAID controllers were used in either case, so I can't really compare.

Either way, if I was in your scenario, I would ditch those drives for ones that didn't have TLER, TLER equivalent or simply get ones that were made for RAID arrays (7s TLER). Any drive they send you is going to be a newer one with TLER values locked and it might cause issues. Knowingly keeping a problematic drive in an array defeats the purpose of using RAID in the first place.

Click to expand...

Yeah, that's exactly what I decided to do. I don't trust a WD drive in there that's not TLERed. I wound up buying 4 http://www.newegg.com/Product/Product.aspx?Item=N82E16822145473 because they're on sale for $99 a pop with promo code EMCKHKG39. I can get back a decent chunk of that when I sell my current drives.

Thanks for the help guys. What a pain in the ***....

Just an update - I was able to rebuild and grow my RAID array just fine with those 2TB Hitachi 7K3000 drives. They seem to be excellent drives, and the fact that they work in a complex RAID array without any mods is great.

On the off chance you continued your research -- do you know if Samsung F4 are fine for RAID10?

BossBorot said:

Since you are running a raid array you need 7200rpm...

Click to expand...

Please, why is this? I am considering a RAID with Barracuda Greens which run at 5900RPM.

SHODAN said:

Please, why is this? I am considering a RAID with Barracuda Greens which run at 5900RPM.

Click to expand...

It isn't a requirement that you run 7200 RPM drives, but it will make a difference. If you have two threads reading from the hard drive (say a VM and a network share), the drives are going to be seeking substantially. Upping the RPM of the drive lowers the seek time and will make the array noticeably faster. If you are going to have a low load on the drives and don't care about performance, you can go with lower RPM drives.

The other issue is that the lower speed drives tend to have features (TLER, etc) that break RAID arrays. This is a hit or miss issue and it seems to depend on the drive and controller combination. I've seen reports of drives continuously dropping out of an array and some that ran for years with drives that "don't work in RAID". I wouldn't risk my data without thoroughly testing the drives in that setup.

Are there certain specifications I can look for to know if the drive is likely to work well?

The only way to know for sure if it works is to use the same combination as someone else (which still isn't guaranteed) or using actual "RAID" drives. The latter is going to be substantially more expensive, but it will work.