[plug] trying to revive a raid5 system

Brad Campbell brad at fnarfbargle.com
Sat Jan 31 04:31:46 UTC 2015

On 31/01/15 11:15, Jon L. Miller wrote:
> I've asked this before to a certain point and now I would like to expand
> this in a slightly different direction.
> I had a Promise SmartStor unit that had 4 hdd setup as raid 5.
> This unit died the good death (after 6 years of no issue operation)
> recently.
> Worked with Promise and they keep stating that there is no raid on the
> system that they can find.
> On another document on the web someone mentioned that you can take the
> drives and put them in a linux distro with LVM and it'll find the R5 on the
> drive and I should be able to copy the data which is all I want to do atm.
> What I would like to know is if anyone has done any hacking of a NS4600 or a
> NAS of this type, I could use some input.

The odds of someone here both having one of those machines and having 
done serious investigation into it are probably a hiding to nothing. The 
fact you got no response last time you asked would tend to back that up.

Pull the drives and hook 'em up to a PC. Use blkid or similar to examine 
the drive/partition looking for signatures and go from there. Sometimes 
when moving from an embedded machine to a standard PC you have issues 
with endianness but from memory mdadm and friends have command line 
switches to help with that.

They may use LVM or MD or a combination of the two. Either way (or both) 
you should be able to assemble the drives into an array in some way 
shape or form. It's been a _long_ time since I've seen any small NAS 
that wasn't running Linux, so the odds are pretty good.

If you find md superblocks, post mdadm --examine results for _all_ of 
them. If it's LVM I'll have to defer to someone else. I personally never 
found a use case where the additional complexity justified the outcome. 
I'm sure it has a place however.

Dolphins are so intelligent that within a few weeks they can
train Americans to stand at the edge of the pool and throw them

More information about the plug mailing list