md's ability to make arrays from partitions offers users great
flexibility, while simultaneously providing the novice enough rope to
tie his own noose...

You could have done worse.  I've read of cases of people making a RAID10
by taking two disks, laying two partitions on each, creating an md/RAID0
across the two partitions on one disk, then making a RAID1 across the
stripes.  50+ posts later, the members of that forum were finally able
to convince the guy why it wasn't working the way he envisioned...

On 12/19/2013 7:57 AM, Wilson Jonathan wrote:

On Thu, 2013-12-19 at 03:17 -0600, Stan Hoeppner wrote:

quoted

On 12/18/2013 3:12 PM, Wilson Jonathan wrote:

quoted

I realise if it is possible its not going to be the best solution, but a
temp stop gap... 

The primary purpose of RAID is to protect your filesystem and files
against failure of a disk in an array.  The purpose of RAID6 is to
protect against a second disk failure or read error while you're
rebuilding the first failed drive in an array.  By creating a RAID6 with
one missing, you've essentially created a RAID5 array, as you can only
lose one drive.

Why didn't you simply create a RAID5, then reshape it to RAID6 when you
acquire your "4th" disk?  If you'd done that you wouldn't be worrying
about "plugging this hole".

I think I over thought things/was trying to be to clever :-/

quoted

The tone of your request suggests the time frame to add the 4th drive is
now indeterminate, and you have a realization that it may possibly never
happen.  This being the case, I'd suggest you backup all your files,
blow away the 'kissing' RAID6, and make a fresh RAID5 of your 3 drives.
 Or a RAID10.  Then format and restore.  Your performance will increase,
and you won't have to worry about this missing drive in your crippled
RAID6 array.

The intention was to copy the existing raid6-6drive data (the end of the
new 3tb drives now holds/replaces the original 1tb drives) then copy the
data across to the new raid6 (held in the first 2tb of space on the
3-3tb disks) then destroy the 6 disk raid and finally re-use the disks
until a new drive can be purchased (was looking at 1-2 months time)
however thinking about it again and in light of your comments I think
I'll leave everything well alone and leave the 6 disk original running
(with 3 new drives)... when I've installed the missing 4th disk at that
point I will copy over the data, then tear down the 6 disk array, that
way I will have my 2 drive redundancy should any disk suffer an early
death. It will also mean that the existing 3 new disks will have had a
1-2 month workout, again allowing for 2 drive deaths in the original
array.


--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html