On 2/01/2014 04:10, Krzysztof Adamski wrote:
On Wed, 2014-01-01 at 23:33 +1300, Pieter De Wit wrote:
Based of what you are saying you probably would not believe that RAID5
can be made with 2 drives, DATA + 1 PARITY. But Linux RAID supports that
just fine (thanks Neil).
Yeah - you right there - I can't see why you would have a "RAID1 set
with parity", nor can I see how the calculations would work. Then again,
I am not a MD developer and I am not going to look at the source code. I
would like to know what the command line would look like to create this,
for educational purposes.
Hot spares are a joke, you have to wait for a drive to fail before you
use them, that can be to late. A 3 drive RAID1 make more sense. On one
system I have a 8 drive RAID1 (the /boot partition) the / (root)
partition is a 8 drive RAID6.
IMHO, you have answered your question of what RAID level you need to
run, RAID1. As for your original question of the data migration, I would
move the data over on a filesystem level. I used to have this neat trick
where by I created soft symlink between the "new" and "old" locations,
from memory, it only works between different block devices. You then
instruct linux to copy the data to "itself". Since it's a symlink it
copies the data to the new device. It was something along those lines.
You need to mount the source as read only, iirc. Contact me off list if
you want more details.
Cheers,
Pieter
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
