Call me kooky

This was frightfully easy to get working.  I had some qualms about setting up the raid array because it’s usualy a mess to get things working.  Not this time!

Enter dmraid….

First to see if dmraid finds the array/controller at all;

root@phosphers-desktop:~# dmraid -s
*** Active Superset
name   : nvidia_dfebefhj
size   : 1250284800
stride : 128
type   : raid10
status : ok
subsets: 2
devs   : 4
spares : 0

Yup, even notices the raid10!  Now to check if the devices are showing up correctly;

root@phosphers-desktop:~# dmraid -r
/dev/sdb: nvidia, “nvidia_dfebefhj-0″, stripe, ok, 625142446 sectors, data@ 0
/dev/sdc: nvidia, “nvidia_dfebefhj-0″, stripe, ok, 625142446 sectors, data@ 0
/dev/sdd: nvidia, “nvidia_dfebefhj-1″, stripe, ok, 625142446 sectors, data@ 0
/dev/sde: nvidia, “nvidia_dfebefhj-1″, stripe, ok, 625142446 sectors, data@ 0

Another yep.  As a side note, once dmraid was installed and activated the first time I never had to touch it again.  Nice and easy.

Now to the confusing part.  Since the 2.6 series of kernel does some fancy new stuff the partition information ends up in /dev/mapper.

root@phosphers-desktop:~# ls /dev/mapper
control  nvidia_dfebefhj  nvidia_dfebefhj-0  nvidia_dfebefhj1  nvidia_dfebefhj-1  nvidia_dfebefhj2  nvidia_dfebefhj3

Whole bunch of stuff.  The numbering is what threw me for a loop.  Notice the -0 and -1?  Those are the stripe sets, and what I tried to mount first to no avail.  The non dash numbers are the actual partitions, and once I understood that it was easy to get things mounted.

Normal sata2 disk;

/dev/sda:
Timing cached reads:   7958 MB in  2.00 seconds = 3982.14 MB/sec
Timing buffered disk reads:  208 MB in  3.01 seconds =  69.17 MB/sec

RAID disk/s;

/dev/mapper/nvidia_dfebefhj:
Timing cached reads:   9068 MB in  2.00 seconds = 4538.92 MB/sec
Timing buffered disk reads:  332 MB in  3.01 seconds = 110.32 MB/sec

Just a bit faster, which is fine by me.