Drobo

Does the Drobo regenerate data as the disk pack is accessed during a rebuild?

I’m in the process of upgrading the disks on my 4-bay Drobo v2 as I’m low on free space.

I had 4 x WD10EADS-00L5B1 installed and bought two WD20EARS-00MVWB0 drives to swap in.

I’ve found that the rebuild process is taking a very long time (yep, I realise the drives I’ve bought won’t help matters but bare with me).

When I disconnect the drobo from my computer, the drobo seems to be idle but not quite idle enough to spin down the drives. The drives chatter but not like a very intensive process is accessing them as I assume a rebuild would.

When I connect the drobo to the computer (i.e. boot up in the mornings) the rebuild hours have barely changed, maybe an hour or two less than when I left it the night before.

So having read a few forum posts here I thought ‘better safe than sorry’ and proceeded to copy everything off to a DroboFS in case I got into a rebuild loop or a power cut occured or I got bored enough to do a full reset of the drobo.
But, as the copy is proceeding, I’ve noticed the approximate hours remaining has come down significantly (i.e 84 hours down to 60 down to 42 as it currently stands).

So my question is, if I touch all the data in my disk pack i.e. read every byte of every file, will this force my drobo into rebuilding that data there and then or am I still going to have to wait a seriously long time for those green and orange LEDs to stop flashing at me?

Some additional information could help to understand. You have 4x 1TB drives and you have swapped out 1 drive from 1TB>2TB, or you swapped out 2 of the 1TB>2TB at the same time?

Rebuild process: it works whether DROBO is connected to your computer or disconnected. I would disregard Dashboard estimates, in this situation. Because IME, the Dashboard takes more than a few minutes to accurately assess the rebuild time. So, it might be accurate right now. But if you shutdown the computer in the evening and then turn up in the morning, Dashboard rebuild time assessment starts from scratch.

Relayout of the diskpack, when swapping from 1TB>2TB could take up to 3 days(per disk), so bear with it. I haven’t attempted to upgrade two 1TB>2TB at the same time.

also have you got any virus programs running with autoscans etc?
eg does it virus scan the drobo - if so, that could also be slowing down the rebuild since its trying to access and scan files too

No, I’ve switched all the virus scanning software off as well as adding the drobo to the exceptions list.

Yoyoma, I’ve only swapped the first disk out so far and it looks like it will take 4 days in total if the current estimate is to be believed. I don’t really fancy having to wait another 4 days when I swap the 2nd disk so I think I’ll continue to copy everything off and start from scratch when it’s done.

I’m going to assuming the answer to my initial question is no, the the drobo doesn’t rebuild data as it is accessed and that the drobo rebuild will finish in its own time regardless of whether I read every byte or not.

I misunderstood your question, didn’t read it properly. You are doing a disk relayout/disk upgrade and attempting to copy off all the data to make a back up at the same time? That will take a l o n g t i m e. While the Drobo still lets you access your data during a rebuild, you should do so sparingly, or not at all.
In your case, I would do what you are planning to do. Stop touching the Drobo, let it finish. Then make a back up copy of data. Then proceed to upgrade your Drobo with second 1TB>2TB.

Thanks Yoyoma, I’ll definately remember that for future disk upgrades but I’ve just returned to my machine after leaving it copying all night and the hours remaining is now at 14. It looks like the drobo is well on track for 4 days even though I’ve been reading data off it.

I’ll let you know if it actually finishes in 14 hours though. :slight_smile:

Well that put an end to that little adventure… The 2TB disk is now marked as failed. Looks like the 2nd disk I bought (from the same batch) is dead right from the word go too (it won’t mount when attached to any PCs or enclosures).

That’s probably why the Drobo has been so slow when left on its own without the computer hassling it.

I’ve put the old 1TB disk back in which has triggered another rebuild but it looks and sounds much more energetic and on some of the larger archive folders, all the MD5 sums are coming back fine so thanks to data-robotics engineers, I may well have dodged a bullet there. :slight_smile:

Well that put an end to that little adventure… The 2TB disk is now marked as failed. Looks like the 2nd disk I bought (from the same batch) is dead right from the word go too (it won’t mount when attached to any PCs or enclosures).

That’s probably why the Drobo has been so slow when left on its own without the computer hassling it.

I’ve put the old 1TB disk back in which has triggered another rebuild but it looks and sounds much more energetic and on some of the larger archive folders, all the MD5 sums are coming back fine so thanks to data-robotics engineers, I may well have dodged a bullet there. :slight_smile:

That sucks. A lot of wasted time with a failed disk. But, it’s a great feeling to see it doing its thing, rebuilding on the original 1TB disk, isn’t it?

It certainly is. I’ve had a few glitches here and there with this Drobo (I’ve had it for over a year and a half) and inspite of everything I haven’t lost a single byte of data so even though my first disk upgrade was a bit of a nightmare, I still trust it.

I think the only thing I’d ask DR for is better diagnostics. When I placed the first 2TB disk into the Drobo (the one that was dead from the start) the Drobo just excepted the drive, said it was rebuilding and then went to sleep.
The power supply I run the Drobo through has a current meter and when it’s busy, the current can reach 2.5amps but when it sleeps the current sits at somewhere around 500mA which is exactly what it did but nothing on the dashboard indicated any failure.

After a couple of hours I decided to pull the first 2TB drive out and replace it with the other 2TB drive which triggered a rebuild during which the Drobo didn’t go to sleep but wasn’t fully active (current was at about 1.7A indicating it was awake but not accessing the drives). Hence my initial post of this thread.

Looking at the log file there’s alot of data that looks like it’s scrambled (not quite encrypted but not quite raw binary either) followed by some TMDiag and DDDiag text. The TMDiag says the drive stopped transfering data at 11:52 but my guess is there’s alot more information in the unreadable block at the start of the file.