It all started with the blinking red light...

I've got an Iomega IX4 StorCenter 4TB(1TB x 4) setup as RAID 0. It appears one of the drives is failing, however it still seems operational. I'm able to access the web admin pages and check the device logs and was able to identify the failing disk.

What I've done so far is take a newer 1TB disk and duplicated the original disk to the new one. After replacing the drive, the disk errors in the logs go away(no hardware errors), however, it seems the RAID array is somehow corrupt or messed up.

A) Is there any way to recover the data on these disks?
B) even better, can the RAID setup be repaired or recovered?

Any help or insight would be appreciated.

Best practice is to clone each drive & build a virtual raid ie in rstudio & then recovery the data from that.

Loki

How did you duplicate the drive to a new one ? what program ?

@loki - thanks, I'll give a go, however my challenge would be the additional 4TB needed for that.

@networks - I used a hardware device that supports direct duplication, no software required

@networks:

Good question

@tcelias:

What specific error information is shown for the "suspect" drive, in the logs that you have?

Were you aware that RAID 0 has no redundancy? (the discussion on why it is called RAID 0 crops up quite often in other forums). Having a multiple disk RAID 0 volume (as you have) is even worse regarding MTTDL (mean time to data loss), than using 1 disk. In that configuration, you will have a problem if any of those (in your case, 4) disks reports any kind of error. In enterprise RAID systems, RAID 0 would only be suggested for data that the customer is willing to lose (e.g. temporary files which can be easily recreated), and where the tradeoff of high speed for low reliability is one they are willing to make. That should tell you something Therefore you may want to reconsider the use of RAID 0 when this situation is finished, if (as is the case now) you have data that you don't want to lose, stored on this box.

Edited to add: I just saw your answer above, to the question from networks. How do you know that the duplication was successful? What is the specific duplicator that you used? Other people here may be familiar with it. The reason for the questions is that, based on your comments so far, I have some concern about whether the duplication was actually complete or not.

I want to know if the duplicator is a sector level duplicator

@Vulcan - sorry, should have prefaced the original post with the understanding that the data is not super critical and I've contemplated just buying a new 1TB drive replacement and resetting the whole thing. That being said, I was aware of the risks in originally choosing RAID 0 when I configured it and accept the fact that losing everything may occur. But I don't give up easy, so that is why I'm reaching out.

I have an Aluratek (AHDDUB100) HDD Duplicator and duplicated the original hard drive onto a new drive. After the duplication was complete, I took a peek of the result in UFS explorer to see and verify that the partitions were there, just as the original. I am not an HDD guru, so I do not know what else to check beyond that which is why I'm reaching out here.

The drive in question is the 3rd one in the array and looking at the HD error log:
*see attahced file*

@tcelias:

Thanks for the info - wish I hadn't spent time explaining why RAID 0 was a bad idea, if you already knew and don't really care about the data, but anyway...

In short - although there is an interesting anomaly in the SMART data, the disk has been reporting read errors.

Having briefly read about your disk duplicator, it is designed for copying perfect disks only and seems to have no ability to tell you whether it encountered unreadable sectors or not. Also, what it does for an unreadable sector on the source is not defined anywhere that I've read e.g. does it leave that LBA untouched on the target; fill that LBA with zeros; fill that LBA with some other value; or something else? In fact one unreadable LBA may cause a larger block of data (e.g. 128k or more, for example) to not be copied from source to target (I've seen that behaviour on another duplicator). In other words, IMHO it isn't suitable for your purpose, which is trying to copy a failing disk.

So if you're taking the risk of DIY, and if you accept that the disk is deteriorating and could become even worse, then one common approach as loki said, would be for you to clone all disks but especially this failing one, using something (software or hardware) which can be coaxed into skipping just each unreadable LBA if needed (and then going back and doing many retries to see if any sectors are readable after more retries), as well as logging which sectors were unreadable. Then attempt to reconstruct your RAID using the clones. (Don't overwrite your existing, likely imperfect, copy of the failing disk made by your duplicator yet - if the failing disk dies during your next cloning attempt onto yet another disk, that copy might be the best you have ).

Personally I use ddrescue on *nix for that purpose, but that's due to having experience in driving it and altering its parameters (and *nix drivers), to interactively get the behaviour I want. IMHO it's not easy to get the very best out of it without plenty of practice & experience, but it does a reasonable job in some cases, even without tweaking. Other people recommend different software (DMDE, HD Clone etc.) which I haven't tried. Search the forum for posts about clone or cloning for previous discussions about suitable software.

A DR company would likely use a duplicator specifically designed for cloning failing disks. You may want to consider that option, but if you only send them one disk, then you can't be sure of success when you try to reconstruct the RAID yourself.

Good luck!

Thanks for the info Vulcan! No worries as looking back at my original post, I failed to include all the details (kids running about and so on...)

I was considering the option Loki mentioned but was also comparing the cost of buying new drives to have the capacity to setup the virtual raid against cost of a professional.

I may even be in denial (lol) and end up scraping the whole thing and reformatting everything, I just hate to have not gone through all reasonable options.

I'll try the software route on duplicating and see if that's any better.

Thanks again for the insight.