Velociraptors: RAID problems - new firmware

A

aitoribarra

Weaksauce
Joined
Sep 16, 2009
Messages
92
Hi,

If you use WD Velociraptors in RAID and run them 24/7, and you get occasional failures where they drop out of RAID, yet rebuild fine with no loss of data, then you may be interested in this thread:

http://forums.storagereview.net/index.php?showtopic=27303&st=50

So... there's a bug in the TLER code in the firmware that causes errors after 49 days of continous operation. I was faily sure that my problems with Velociraptors were down to the RAID controllers I've used them with (Areca 1680ix and Dell PERC 6i) - but it turns out that the drive firmware is to blame. Obviously the defnition of "Enterprise drive" varies, but I'd have thought that more than 50 days of continous operation would have been one of the criteria...

But! New 04.04v02 firmware fixes the bug!

However... the 04.04v02 firmware is not available on the WD website, and 1st level support claimed that if it wasn't on the website, it didn't exist... I've escalated to 2nd level support and they claim that they're unaware of the issue etc etc. Hopefully a bit more pushing and they will release it to me, but I would rather that they posted it publicly.

After struggling for months with Velociraptors, I was ready to give up and pay the extra for more SAS drives - but this new firmware may make me stay their execution.

cheers,

Aitor
 
It's my understanding that this has been an issue for a lot of people (not just with VRaptors). You can (and should) disable TLER when using RAID controllers with non-enterprise drives. There is a utility that will let you do so, so there shouldn't be any reason to have to jump through hoops for a firmware update.
 
aitoribarra said:
Hi,

If you use WD Velociraptors in RAID and run them 24/7, and you get occasional failures where they drop out of RAID, yet rebuild fine with no loss of data, then you may be interested in this thread:

http://forums.storagereview.net/index.php?showtopic=27303&st=50

So... there's a bug in the TLER code in the firmware that causes errors after 49 days of continous operation. I was faily sure that my problems with Velociraptors were down to the RAID controllers I've used them with (Areca 1680ix and Dell PERC 6i) - but it turns out that the drive firmware is to blame. Obviously the defnition of "Enterprise drive" varies, but I'd have thought that more than 50 days of continous operation would have been one of the criteria...

But! New 04.04v02 firmware fixes the bug!

However... the 04.04v02 firmware is not available on the WD website, and 1st level support claimed that if it wasn't on the website, it didn't exist... I've escalated to 2nd level support and they claim that they're unaware of the issue etc etc. Hopefully a bit more pushing and they will release it to me, but I would rather that they posted it publicly.

After struggling for months with Velociraptors, I was ready to give up and pay the extra for more SAS drives - but this new firmware may make me stay their execution.

cheers,

Aitor
Click to expand...

Excellent information.
 
Unfortunately, for whatever reason, WD support aren't giving me the firmware update. They say there are no firmware updates available for the velociraptor. It's almost as if they don't want anyone to know that there's a problem - which is crazy if they've got a fix. However, they have sent me a utility which patches the existing firmware that is supposed to fix the 49 day problem. I would rather have the new firmware though, as it has other fixes & improvements too. I haven't tried this patch yet but will do just as soon as I can find a window of opportunity to take down my RAIDs...
 
OK, now I've got the firmware update. Looks like 2nd level support escalated when I sent them the PDF. Hopefully now they will actually make the new firmware publically available, so that everyone else who experiences the same problem doesn't have to go through the same hoops.
 
I would have tried this but my second Velociraptor died on me. It won't even spin up anymore.
 
Well, at least you can RMA it. If they send you a new drive, or even a refurb, it would hopefully come with the new firmware...
 
I have also had numerous problems with these Velociraptors wd3000GLFS drives. I use to get these raid access errors on my 680i system that were only fixed temporarily by turning on TLER on both drives and running that utility. After 6 months the problem came back and I assumed it was the 680i fault. So I plugged them into my new x58 EVGA board and now only one of the drives successfully posts at start up. The other one only sometimes will register at post and most of the time does not despite swapping power cable and port cables. It will even slow the post process if plugged in. So I must have a bad drive? Or is this a firmware problem as was mentioned before?

In any case I am so pissed at WD I don't think I could ever support their company again...
 
You probably just have a bad drive. I know I've got one. (Actually I've got a growing stack of failed Western Digital hard drives.)
 
Dan_D said:
You probably just have a bad drive. I know I've got one. (Actually I've got a growing stack of failed Western Digital hard drives.)
Click to expand...

Me too! I just shot an email to the customer relations manager at Western Digital about this. I don't think it ever worked right from the first day I owned this drive. I just tested my two drives with an usb adapter and only the 'good' drive was recognized by one another machine. Definitely a bad drive!

I just hope they'll send me a couple of newer release drives with newer firmware as I am getting sick of this as well....

Thanks.

:mad:
 
Last edited:
Can you post this firmware somewhere? I have a couple of these that are now in a server and would like to avoid this issue. Also will the firmware update work correctly with the drives in a RAID config or do they need to be removed from the RAID before updating?
 
I applied the new firmware (which WD support gave me after much hassle) on Saturday...

Firmware update seemed to go ok (you need to do this connected motherboard sata after booting to DOS).

Started rebuilding RAID... failed after a few minutes. It tried again, failed again, at this point RAID controller got really confused and tried to rebuild from the failed drive -> raid volume corrupted.

I had to pull all my velociraptors and restore from backup to other drives (luckily I had enough free space!).

So, no, don't trust the firmware update with RAID. RMA the drive instead and hope they send you a good one.
 
I just spoke with a representative from WD and seems like I had bought the wrong kind of Velociraptor. Mine are GLFS and they strongly advised me not to run them in raid and that raid was not supported and that I needed the HLFS type. They also told me that running them in raid voids the warranty!

But he said just rma it and get a new drive and don't run it in raid!

Feeling depressed I bought the wrong drives....
 
davidm71 said:
I just spoke with a representative from WD and seems like I had bought the wrong kind of Velociraptor. Mine are GLFS and they strongly advised me not to run them in raid and that raid was not supported and that I needed the HLFS type. They also told me that running them in raid voids the warranty!

But he said just rma it and get a new drive and don't run it in raid!

Feeling depressed I bought the wrong drives....
Click to expand...

I think he was B-Sing you. I'm beginning to think that's standard MO for WD reps... Anyway, as far as I know, there is no difference between BLFS, HLFS and GLFS models apart from the heatsink (or lack of). Sure, removing the heatsink from the HLFS or GLFS counts as a mod and probably voids the warranty - but running in them RAID? Why would they have the TLER function then? Or the RAFF feature? Anyway, don't worry, even if you'd bought GLFS (no heatsink) or HLFS (heatsink but with connectors in right place for a backplane) you could still be having RAID problems...
 
My GLFS have heatsinks, but the rep led me to believe that running them in raid could shorten their life in few words more or less. Is this true?
 
davidm71 said:
My GLFS have heatsinks, but the rep led me to believe that running them in raid could shorten their life in few words more or less. Is this true?
Click to expand...

You're right, I got BLFS and GLFS models mixed up... Sorry!

Shortening their life? I don't see how. RAID doesn't necessarily mean more work for the drives. I guess if you've got two drives, then running them as two indpendent disks can mean less work for them than in RAID, because every access will involve both drives, but then, a two drive system with the disks in RAID should be no more work than a single disk system. The real issue is how hard you hammer the drives.

But these are "enterprise" drives, with a 1.4 Million hours between failure rating (I assume at 100% duty cycle). They are meant to be hammered.
 
Well the WD rep told me that my GLFS drives are not 'enterprise' but desktop grade drives. So I guess I'm going to run them individually from now on. And when I asked him why I can't run them in raid he referred me to a document that described TLER and the lack of it with my drives. But then I told him I had a utility that turned it on and he said that was for their own company internal testing and not meant for the consumer. He wouldn't say that running them in raid is dangerous but gave that impression more or less and urged me not to do it. Probably because of a timing issue in the time error recovery in running them in raid.

The funny thing is when I first got these drives I had them in raid 0 and was getting all these raid access errors. I almost returned them and got a $70 check from WD to help with the restocking fee to make me happy. Then I found the TLER utility and the errors disappeared for like 6 months. So atleast I got them on the cheap thanks to WD consumer relations departement. Now I have to rma one of them as its dead as a door nail. Just wish I could run them in raid...

Thanks.
 
WD agreed to replace the drives with the HLFS versions instead, but whats kind of odd is that there are six tracking numbers. Weird...
 
Are you saying the HLFS models don't have this problem? I thought it applied to the entire family.
 
cZm said:
Anybody know how to update the firmware on my VR's??
Click to expand...

Create a MS-DOS startup disk.
Boot into DOS.
Run the firmware update utility.

Volkum said:
Smells like BS to me. They are all the same drive.
Click to expand...
The only difference is that the GLFS was discontinued (I believe) before this firmware was developed, and WD, being lazy, didn't do the necessary QA tests with the new firmware to release it with the GLFS supported. If the GLFS had TLER explicitly disabled via firmware while the BLFS & HLFS didn't, that may be part of the reason why. This is of course beside the fact the BLFS, HLFS, and GLFS are all identical hardware.
 
Last edited:
Thanks for the help

Couple of things thou. Where do I get the utility form and how to I run it??
 
The firmware update utility is on the previous page, and in my previous post I already explained how to run it.
Create a MS-DOS startup disk.
Boot into DOS.
Run the firmware update utility.
Click to expand...

After reading it again, if you still don't understand how to run it, you shouldn't even be considering trying to update the firmware on your VR by yourself. You would be better off finding someone to help you, or not doing it at all.
 
Western Digital sent me HLFS drives to replace my GLFS drives and not only that they made a mistake and sent me double the drives I needed! I emailed them and pointed out the shipping error and they said keep 'em! So setting up a raid 5 array as we speak!
 
Does applying this firmware actually affect the data on the drive? Should I theoretically be able to just plug one drive in at a time to the onboard sata controller, boot into dos, update the firmware, and then put them back on the other raid controller and have them continue to work and have the same data?
 
Cyberbeing said:
Here is the firmware:
http://www.mediafire.com/?imzyjmumrjm
Click to expand...

13063_wdc.jpg


:(
 
Odd, maybe that particular firmware update utility is only good for the 300GB VRs?

Anybody have success on their 150GB or 74GB VRs? If not, I guess you will have to get the firmware directly from Western Digital.
 
davidm71 said:
Western Digital sent me HLFS drives to replace my GLFS drives and not only that they made a mistake and sent me double the drives I needed! I emailed them and pointed out the shipping error and they said keep 'em! So setting up a raid 5 array as we speak!
Click to expand...
Hey davidm71, I also have 3 GLFS and one of them is acting up. How did you get a replacement from WD? Did you say anything in particular that made them send HLFS? I am planning on calling them and wondering what and who to speak to to get the exchange. Thanks!
 
hello,
I wanted to ask if I can update two velociraptor GLFS in raid0 with this new firmware.
thanks
 
It doesn't work on GLFS drives...I have two in RAID 0 and got the same error that is shown on the screen shot in this thread...
 
Hallo everybody, I'm new... I just run firmware update for my 2 VR Raid 0, first one had 0404v01, but second one had 0404v03!
Anyone have this new version, please?
 
imyourzero said:
It's my understanding that this has been an issue for a lot of people (not just with VRaptors). You can (and should) disable TLER when using RAID controllers with non-enterprise drives. There is a utility that will let you do so, so there shouldn't be any reason to have to jump through hoops for a firmware update.
Click to expand...

You want to enable TLER when the drives are used in a RAID array. TLER is disabled by default on non enterprise drives.
 
UPDATE!

did the update on my WD1500HLFS- 01G6U0 (04.04V01) to firmware version: 04.04V02
no problems. my drive did not have any recent problems but just thought i do the update

other notes i checked with hp.com and they have their vraptor firmware versions up to version 04.04V05 Rev.A And this is to fix the same issues but for their specific models:

WD800HLFS-60G6U0 or U1
WD1600HLFS-60G6U0 or U1

i first tried the hp update but it did not work so i ended up going with the 04.04V02 update.

and drives in raid must be taken out of raid so the firmware updates can take place, and then the drives can be put back into raid array without affecting raid setup/data what so ever.

btw-glad to have this thread!:)
 
I found that firmware: readme says that it will not work in raid configs: did you tried removing it?
 
Last edited:
That HP firmware won't work unless you own a HP OEM Velociraptor, since it checks for 60G6U0 within the model name. WD branded drives are all 01G6U0 I believe.
 
Smart... I have Dell OEMs, anyone has a direct link or knows how to modify HP's exe to change serial check?
 
You must log in or register to reply here.
Back
Top