Drobo

Drobo has decided my drive is bad (its not)

Spotted a red drive today, so removed it, plugged it into the mac and after reformatting it, its totally fine.

Put it back in the drobo, but the drobo is not having any of it.

What am I supposed to do?

If the drive light was blinking red, replace the drive.

http://support.datarobotics.com/app/answers/detail/a_id/207/related/1/kw/red%20light/r_id/100004

It is understandable that you would be confused or suspicious if your Drobo product flags a bad drive when the drive works fine in your computer.

If a drive produces a lot of errors in a short amount of time (it’s going bad soon), Drobo storage devices display a red blinking light to indicate a failed drive. Your computer may not catch the bad drive as early in its failure process as your Drobo product can. This difference in perception is especially exaggerated if you format the disk before testing it in the computer and/or it has just a small amount of data. Eventually, the computer will also flag the drive as bad.

Thats fine, but it kind of puts me in a position.

Its a seagate drive and in warranty. But its not failed so they won’t take it back. I don’t believe for one minute that a Drobo is any more clever than lets say Apples Disk Utility, or Seagates own ‘SeaTools’ which both report the drive as fine.

So my guess is the drobo is confused or its buggered the drive. Afterall if I have reformatted it and its blank how is drobo then flagging lots of errors (I reinserted the drive)

I don’t pretend to be an expert or even knowledgeable, what I do know is that the seagate software says the drive is good so no claims, and I am willing to bet data robitics ain’t gonna help either.

Open a case, give me the case number and I will look at the diagnostic file.

OK, Is it worth me doing that in as much as I have put another drive in now and its doing its thing which will take 20 hours? Will that ruin the log file?[hr]
OK this has become alarming. I decided I would delete a folder of films that I have a back up of as the drive I put in as a replacement is small leaving me critically low. I deleted the 50 gig folder and stopped it half way through the delete as drobo dashboard is telling me its freed up 200 gigs so far.

I fancy this is heading to a s**t fest.

ONe other thing how to I open a support case?

Either call our number or create one from the support login page.

Wait until after it’s done the relay out process before generating a diagnostic file.

Well it started out at 26 hours to relay now this morning it says 72 hours.

IS there a way to stop it doing its thing, so I can get my data off, I don’t trust this drobo at all.[hr]
i created a support account but see no where to raise a case, presumably because my drobo is over a year old. After a year does DR wipe its hands of responsibility for its hardward and my data?

Your data is still accessible while it’s doing relayout. It’s just not fault tolerant.

@ garethi - its kind of funny that you dont trust drobo - when in fact its taking better care of your data than your computer is?! its flagged a drive as bad, when your computer cant yet detect the fault/impending failure (or more likely can but just doesnt think its a big deal - yet…)?

we all know (I hope) the two states that drobo and the manfuacture’s diags (pass & fail / working and rejected by drobo ) produce are both based on a wide variety of variables fed back by SMART and (probably) in drobos case ongoing monitoring of drive behaviour.

Now with several variables available to look at, and some of those having quite wide ranges you would not expect drobo and the manufacturer to produce corresponding pass/fails 100% of the time anyway… but if you look at their motivations -drobo is (hopefully) applying the harshest criteria to maximise data security, and the manufacturer pushing in the other direction, trying to minimise the number of returns, I’m not in the least bit surprised that they frequently give different results!

On top of that, drobo could well be watching the rate of change of some of the smart variables, so even though they haven’t yet reached the threshold value for the manufacturers diags, drobo could be well aware the values are accelerating south! Don’t forget drobo has 24/7 access to drives under operational conditions, and will have been watching them for weeks if not months, and expecting a high standard.

Manufacturers tests will have access to the drives for (probably) a few hours, and are, frankly, trying to see what they can get away with without having to RMA the drive.

If you are serious about data security (if you aren’t - why have a drobo ?) Then which would you trust?

I agree drobo should be more verbose and tell you WHY its failing the drive, but send DRI your logs and they will!

And so far I haven’t heard of anyone who has been refused an RMA (and i’ve returned several drives which drobo failed and the manufacturers tools said were fine - without any problems), which suggests to me the manufacturer looked at the drive, and thought “fair enough, it probably is on the way out”

Yea thats cool.

My trust is failing because a folder which was reported as weighing 50gigs once half deleted drobo has freed up over 300 gigs.

Explain that one. Thats where I am worried.

This is the sort of thing that happened before when I removed my drobo from the droboshare. The droboshare is dangerous, sorry but it is. DR say you can happily swap between droboshare and direct USB. Well not in my experience. all sorts of crap starts happening and its doing it again. Instead of the 70+ hours its telling me it needs to relay 500 gigs of data it would be quicker for me to copy the data back onto the drive drobo says is bad (but that all other software says is good) and reformat the drives.

Drobo is suposed to make me feel safe about my data, sorry but I don’t.

Not going to comment on DroboShare as I never ran it with HFS+…

But in terms of Drobo marking drives bad, it’s important to understand that Drobo’s decision seems to be based on trending while the drive manufacturer’s diagnostic utilities (all the ones I’ve used so far - Quantum (now) Maxtor (now) Seagate, Western Digital, IBM and Hitachi) measure full, flat-out failure.

Excluding true “false positives” due to conflicts between the drive and the Drobo logic, a red-marked drive in Drobo is a “warning” or a true “fault” - I believe there’s a thread in the feature request forum requesting a “warning” state for “drive is very likely to fail soon” versus “drive has failed.”

Or another way to look at is Drobo calls the fire department whether the smoke detector goes off, or someone sees flames. Paranoid, you may thinking, but a little paranoia is often helpful. :slight_smile:

perhaps flashing yellow rather than flashing red, since they dont use just simply “flashing yellow” for anything at the moment?

OK, I have ordered a new drive and also sent the other one off via RMA and see what happens there.

I have no idea why its going to take so long but basically its been messing with data for twenty hours now with around 40 hours to go for a 500gig drive.

My new drive will be in tomorrow, If I yank the drive I put in yesterday whilst its doing its thing will that be OK. Basically I want to stick the new 1.5tb drive in tomorrow and basically ignore the damn thing while it takes four days (yes count them) to move around 800 gigs of data.

You are not fault-tolerant during relayout. Therefore, you should NOT remove any drives!!

Relayout doesn’t affect just the drive that was replaced, it affects data on all the drives, so the time it takes is a function of your total storage capacity.

You cannot remove a drive if the drive is in the relay out process, if it’s blinking amber/green.

You will lose data and probably cause a Too Many Hard Drives Removed error message.

OK, But seriously I have calculated it as being about twice as fast to remove all the data to other drives and copy it all back to a reformatted drobo, 70+ hours is not acceptable really, I hope Pro users don’t suffer the same issue if its used in a business environment.

Ok, then you can do that.

Normally you’d replace the failed drive with the one you intend to permanently replace it.

In a business environment, it’s common practice to have a spare drive to use as a replacement.

There wouldn’t be downtime, though there is no fault tolerance while the relayout is happening (unless you use Dual Disk Redundancy).

Moving your data off to a non-fault tolerant storage, then reformatting the Drobo and copying the data back incurs a good amount of added risk.

In your current state of relayout, the only risk factors are that another drive in the Drobo will fail.

If you copy the data off the Drobo, to non-fault-tolerant storage, that’s fine.
But if you then reformat the Drobo, now you are relying on that non-fault-tolerant storage as it is your only copy of the data.
If there’s an error in the copy to or from, that’s data lost.

It’s a balance between being patient and waiting or increasing the risk to restore the fault-tolerance faster.

OK, I will calm down now. I have a new cavier green in my hand and will wait for the process to complete.

I will hopefully get a new 1.5tb back from the RMA and keep this as spare for the next time.

Non-fault-tolerance puts me on edge too. :slight_smile: