Drobo

WD20EADS with firmware: 00R6B0

Hi,

I bought 2 weeks ago a Drobo at the same time with a new HDD Western Didital WD20EADS 00R6B0
Drobo tells me that the HDD failed (after 2 weeks of good service) and to replace it.

I tried the “Failed HDD Wester Digital” on a USB box and it works fine !
I managed to format it , copy files etc.
I put it back in the Drobo and tell me that the drive has a problem !

Could you give me a help here, the drive is fine but Drobo does not want it.

Configuration Mac OS X
Drobo -> 300 GB Seagate
500 GB WD
500 GB WD
2000 GB WD

I’m sure I remember reading the Drobo uses its own system of checking a drive and/or can sense when its in trouble and may be going to die. Maybe the it has discovered something wrong with the drive?

I downloaded the disk utility under windows for the Western HDD. There is nothing wrong with the drive. I spend the day to detect bad cluster and there is none.

Help needed please!

I’m not an expert, but I would presume all the various disk utilities out there work on a similar principle. If the drobo really does have its own way of looking at a hard drive - then maybe everything else would give it a pass and say its ok?
I think Data Robotics tech support opens monday morning so maybe one of their representatives will reply to your post in the week.

UP !

Please, if you got time to see the problem.

If you open a support case, we can get a diagnostic file from that drobo to see what and why that drive is being flagged.

Just a reminder, the Drobo does look for so many errors in a certain time frame. Once that drive makes that criteria drobo marks that drive as failing/failed. It is more sensitive than your computer.

The tools that drive manufacturers use to determine if a drive is bad, should be run several times. I have seen drives pass the first and second run but after that the diagnostic tools from WD, Seagate, Samsung, etc, do finally show the drive as having issues.

IOW: once drobo says “bad drive”, it will always be a bad drive.

I open a case for this incident. Like I’ve said, the drive works well…
I bought maybe 15 days ago and if I return the HDD they will found nothing bad with the drive I supose.

If you are lucky the Drobo have just saved your data. The Drobo have a more sofisticated error analysis than the USB box.
Hopefully the support-case will help you diagnose the drive and assure that you will get it exchanged from your supplier.

That the drive is only two weeks old isn’t any sort of guarantee of function.
Check out the masters of harddrive statistics, google: http://labs.google.com/papers/disk_failures.pdf
You have a peak of drives that fail in the begining, then it levels out until the drives are two years old when they start failing at almost 10% per year.

Btw, it would be really nice if the dashboard showed some more information about the reason for marking a drive faulty.

I second that! I’ve had Drobo mark at least three drives as failed that have gone on to productive lives as external HDs without a problem. Most recently, I had to replace a <6 month old 1TB Seagate… Drobo had been spontaneously doing relayouts (lasting from 5-20 minutes) for months without me “catching it in the act” (I only found out via the email alerts). It was only after I opened a support call that they were able to tell me which drive was causing the problem from reading a trace.

It would be very nice if Dashboard could give me a little history of the drive in each slot.

Remember too that drives have spare sectors - the diagnostic utilities (even from the manufacturer) won’t mark the drive bad until there are no more spare sectors available to use.

Since Jennifer mentioned that Drobo also flags based on error frequency, it could be a communication problem - for example, flakey contacts. Imagine a sub-par modem connection with error correction. The data will get through, but only after numerous retries.

If that’s the case, it would be quite a shame. A perfectly good drive, once marked bad by the Drobo, is thenceforth unusable in that Drobo. There ought to be a way within Drobo Dashboard to manage the Drobo’s internal database of drive conditions.

If the data is stored on the drive (which the ability to migrate drive sets from one Drobo to another seems to suggest), then using the manufacturer’s utility to zero-out the drive should delete its data.

However, Drobo might be storing a “bad drive list” internally somewhere too, in which case we’d need some DRI-provided function to reset it.

I’ve had a similar problem with a non-Drobo RAID array when a port went bad. The controller had issues and said the drive was bad…
But I’ve also had the opposite - a drive that was marked bad by my RAID controller, seemed to be OK elsewhere, but quickly went bad outside the array too. shrug

Brandon

I’ll get to test that theory shortly as I happen to have a drive in this exact condition (failed by the Drobo, but seems fine otherwise) – I’ve run it through OS X’s “overwrite drive with zeros” format, which passed, but I haven’t tried the manufacturer’s low level format. That will be next…

The drobo does store a list of bad drives. It looks at the serial number and stores that information. Some customers have had success in formatting the drive in their computer and it will work in the drobo again but not all the time, the drobo will see it as a bad drive right off the bat.

I have seen customers do a PIN reset of their drobo with that one drive in and that seems to wipe the information clean.

I honestly wouldn’t recommend those things. There is a sound reason why the drobo is marking those drives as bad. Personally I would rather swap out drives then have to worry if the drive is going to fail at any moment. If you already know a drive is having issues, why not get rid of it? What happens if a different drive fails later on down the road and during the relayout process that first drive that was having “known” issues fails too? That’s a dual disk failure and a drobo can’t protect you against that.

I agree with Jennifer… the best practice would be to “retire” the presumably-bad drive from Drobo use. You might use it for other uses if you want…

Now if you get a new drive and it too is marked bad by Drobo, then either you have incredibly bad luck, or Drobo is confused somehow.

Definitely be good to let DRI look at the logs and determine exactly why the drive is being marked bad.

Jennifer,

i think the issue is more that retailers and drive manufacturers wont replace a drive under warranty, when it appears to work perfectly well in a normal pc/usb enclosure and when it also passes SMART with flying colours… simply because drobo has marked it “bad”.

there should be a “print out the technical reasons why this drive is bad so we can put a post it on top of the drive stating why we are returning it” option

Well the reason was a timeout of the drive.
I did a reset with the drive in (without the others), and now it is working fine … for the moment.
I now have 3 x 2Terra Western Didital WD20EADS 00R6B0 and a 500 go WD

Hi,
I have the same problem with an WD10EACS 1T hdd. I wont get it replaced because the disk works just fine outside of the drobo. It has also passed the analytic tools multiple times.

I reformatted the disk and put it back in the drobo but it stil wont work.

What can I do? How does this reset process work? Is there any support document for that? Why is there no UI way to reset the status of an hdd?

I can`t use the disk in any other way. It was just bought to be used inside the drobo.

Please help.

Thanks a lot!

What I did :
I took off the disks. Place the one that was marked failed by Drobo and did the reset procedure (unplug Drobo, press the reset bottom behind, and while still pressing the button put on the plug back, continue pressing the reset button until the power indicator blinks green.

The HD should then no more be red on the Drobo. Once that done you plug it off again and put all your HD at the same place there use to be and switch back on.

It worked for me :slight_smile: