We are just finishing off a 24 x 750Gb Sata RAID 6 in a HP system.

Needless to say, its taken a long time, but its working now.

Anyone worked on a larger RAID?

Simon

http://www.1ct.com
http://www.disklabs.com
http://www.disklabsforensics.com

_________________
Simon

http://www.disklabs.com

International Professional Data Recovery Association

Not quite, but recently had a 2 x 12 Disk (80GB Deathstars) RAID5's configured as an NT Spanned Volume. Recovered 1 of the arrays fine, but the other had 3 x see-thru drives so that F***ed the whole thing up. A catastrophic waste of time in the end!

I hope you charged handsomely!

_________________
Data Recovery Ireland

Did a 28 drive compaq (2x RAID 5 then configured as RAID 0 in LDM), but they were only 72GB. A 14 x300gb drive FCal RAID 6.. But dont really see many Raid-6 systems.

At this point the best is mine?

RAID 4+0 from 12 x 4 x 300GB (48 disk).

The admin triend to replace two disk on one 12 x 300GB RAID4, and one of the working disk has disapeared (sata connection), and the next reboot, the Linux kernel re-ordered the drives, but the admin used the up arrow, and enter to try again the backup....

14TB XFS goes bye, but some part i can recovered with the older, and already dropped drive...

Regards,
Janos

What would be even more interesting, would be a description of the type of failure that was solved.

Cases where metadata were completely wiped out from the disks or even incorrect data was written (e.g. when operator attempted forced re-assembly with wrong sequence)? Or the combination of stale drives, or hotspare with old data etc.

It can be quite a puzzle.

@Simon.... (very interesting, but sarcastically I view these posts on this board and your frequent youtube.com videos verging on blog leveraged marketing rather than contributing to the greater good!!!)

Its not a matter of the volume of the disks, but the nature of the failures across multiple disks. Rebuilding RAID sets of groups of 12 disks, 14 disks, 16 disks (i.e. 24,48,60,72,84,96... 16,32,64,128) has been around in RAID firmware and volume management software for a while, so in reality it is not new.

Clearly different RAID levels leave more remnants of data and metadata hidden on different disks. We view the complexity of a recovery as the likelihood of finding ALL the correct sectors of data in the right order.

RAID-6 was designed as an advancement of RAID-5 to design for larger and larger disks...to ASSIST with recovery. The additional parity disk; enables the RAID set to suffer duplicity of disk failures - i.e. two concurrent disk failures will not suffer in data loss.

Recovery from RAID-0 is the challenge and highest risk. Partial file recoveries often result...requiring additional human intervention to obtain data from "deleted files" in previous file versions. Clearly a health warning needs to be put on any RAID-0 system, though the funniest thing I ever saw was a Sony laptop with RAID-0 disks!

RAID-5 and RAID-6 both effectively use Reed-Solomon encoding - the same alogrithms used for very early solid state disks to protect DRAM memory so is extremely proven and well founded in mathematics. (RAID-6 can also engage traditional parity as the second protection mechanism).

Sorry to put a downer on your marketing efforts Simon, but we must look at the facts and understand the start point.

You are not suggesting that a corporate member of the IPDRA would be up to abusing a forum to get links are you?

Its no wonder no one in the trade wants to join his and his group.... I did a search on Simon Disk Labs and found this beauty ... Look at the post on October 17, 2005, 06:38 AM

http://www.resource-zone.com/forum/show ... hp?t=41031

Looks like he has tried to get his competitors delisted from dmoz claiming they have abused their posting guidelines.. At least my website isnt on his hit list, and all the time pleading that he wanted to help dmoz - now he's trying to say he wants to help the trade !!!

I think I will be staying clear of this one.

One of those is my site!!

Why did you do that simon? Are you trying to get all your competitors kicked off DMOZ ?

<itch>

_________________
If you can keep your head when all about you are losing theirs, you probably don't fully understand the situation. ... Mr Kipling

https://www.mjm.co.uk/

You asked for it....

Originally a RAID 6. This failed. It was then rebuilt in RAID 5, then RAID 6, Then RAID0, RAID 5 and RAID 6. None of this was originally told to us, the client denied everything!

5 x rebuilds with 2 dead hard drives. Hard work!!

_________________
Simon

http://www.disklabs.com

International Professional Data Recovery Association

Only the ones who abuse the guidelines!

You will note that my main site isnt listed....

_________________
Simon

http://www.disklabs.com

International Professional Data Recovery Association

Kangaroo - your point being...

The idea of the IPDRA is pure. We deal with companies who dont have decent skills therefore stop clients having the opportunity of having their data recovered due to the damage they cause. The purpose of the IPDRA is to get members that clients can have confidence in.

Someone has to do something. Its very easy to criticise, much harder to do something about it. I have never charged for membership - so its not a financial benefit. I am trying to do some good.

Never mind....

_________________
Simon

http://www.disklabs.com

International Professional Data Recovery Association

You will note that mine still are listed. So I am not abusing the guidelines, otherwise they would have listened to you.

<itch>

_________________
If you can keep your head when all about you are losing theirs, you probably don't fully understand the situation. ... Mr Kipling

https://www.mjm.co.uk/

I did a RAID 5 with 14 1TB Seagates once and the customer pleaded poverty to me. The RAID card he was using cost a small fortune.

Got a good recovery in the end though

_________________
The last I heard of God he was worshiping me

Thats what its all about

I once did a two disk Raid 1. Disk 1 was completely failed with a pcb problem. I got all of the data. Who's the man!

You are now an established GURU!!

Well I had a Lacie netowork unit once with 4 500Gb MAxtors as a RAID 0.

Two disks had platter damage but I got 350Gb of data recovered off it. Now who da man!

_________________
The last I heard of God he was worshiping me

i have one new candidate again.

One RAID5 set from 3 disk. (adaptec scsi raid card)
The server started to beep, the owner re-started.
(have used again with 2 disk)
The server starts to beep again, the restart have no success. (of course)
Than the owner removed the bad disk, inserted the new one, but can't add the new to the old set.
Next step was, deleting the array.
And re-created the RAID1 from the 2 good disk.
And the sync was cancelled at 45%.

Great.

now i need to recover....