How to determine RAID 6 config from HP400

[PAD-dataservice]

Hi
Need some help to determine the configuration for a RAID 6 from the config files extracted from the end of the drives (originally from a HP400 controller).
Could anyone shead some light on how to do this?

The files are attached (the extention txt is just to get the forum to accept the upload - they're raw extracts)

extract_1.txt

(22.71 KiB) Downloaded 805 times

extract_2.txt

(224 Bytes) Downloaded 766 times

extract_3.txt

(1.94 KiB) Downloaded 716 times

/Michael

[lcoughey]

The three files contain different data and are different sizes...not much help.

[ddrecovery]

What software are you using to try to assemble the array?
Have you made images of the drives?

[PAD-dataservice]

@Luke The files are from the end of the drive (each drive has it) and contains the serialnumbers (though listed in reverse !?) for the drives in the raid - and should contain quite specific info on the raid configuration from the HP400 controller...

@ddrecovery Yes we have full images of 5 out of the 6 SAS-drives. One of the 5 drives might have failed temporarily, when the raid tried to reconstruct from loosing the defect drive and breaking it for real. But there's no logfile for the HP400 we can get our hands on. Nor is there any info from client except it was raid 6 - and drive order is not provided with certainty. But it looks like there is a parity delay.

We often use R-studio to set up Raids - but this one we're using UFS explorer pro (because the raid contains a VMFS partition/VMDK files and UFSexpl pro can handle both)

[DR-Kiev]

According to your files order of parity blocks based on S/N comming from the end of list which is correct to HP layout. There is have been used totaly 7 drives , 6 from the begining, than 1 was added additionally . Probbably made attempt to rebuild new drive.

[PAD-dataservice]

@dr-kiev Do the files contain information on which kind of reed-solomon that were used, parity delay, index type and so on?
I've send you another PM - with a request for paying help... I need to have this done before leaving on holyday

[DR-Kiev]

@dr-kiev Do the files contain information on which kind of reed-solomon that were used, parity delay, index type and so on?
I've send you another PM - with a request for paying help... I need to have this done before leaving on holyday

HP Raid6 usually use Along with XOR HP "ADG" type second parity block in line not Reed-Solomon but very similar.
P.S. Replyed to PM

[ACE]

Hello everyone,

It looks like this RAID has non-standard Reed-Solomon polynomials with unusual coefficients. Data Extractor RAID Edition allows to auto-identify these coefficients and select the type of RS polynomials.
Please feel free to contact Dmitry Chub ts@acelab.ru to get support on it.

[PAD-dataservice]

A follow up on this case:

It ended up being ACE-lap support and DataExtractor RAID edition that finally solved this case:) It was complicated by the fact that the Raid lost two members (and not one as original presumed), so to recover data both XOR and RS had to be used.
With the words of Dimitry fra ACE: "It's easy to use XOR and recover one member fom XOR. With RS it's more difficult because of on HP servers have non-standard coefficients and RS table. (Only Data Extractor RAID Edition supports these types of reed solomon polynom). Also HP raid arrays have delay which makes raid recovering complex case."

I'm looking forward to play a bit with DataExtractor RAID edition and figure out just how he did it

[drHDD]

If only Data Extractor RAID Edition supports these types of reed solomon polynom how all other pro do similar cases?
It's good if ace has this polynom, but it's not true that only them have this polynom.

[hddguy]

If only Data Extractor RAID Edition supports these types of reed solomon polynom how all other pro do similar cases?
It's good if ace has this polynom, but it's not true that only them have this polynom.

You can write custom block layout for a RAID in R-studio - this is one way.

[drHDD]

It's look like you do not understand TS problem. Problem is not in block layout. Problem is in hp tricky reed solomon polynom. R-studio doesn't support it yet. If you can't work with polynom - part of data corrupted if you lost 2 drives in raid 6.