Tim Rightnour writes:
> I attempted to configure a RAID device by following the manpage for raidctl..
>  
> Whenever I attempt to configure.. the kernel spits out something like the bel
> ow:

You did the first config using:

 raidctl -C raid0.conf raid0

right?  (The '-C' is important for the first configuration of a RAID set...)

> The sd10 thing is just wierd.. I can get raidframe to attempt the disk again 
> if
> I dd /dev/zero over the first few KB of the disk.. Every disk spits out an
> error similar to the below.
> 
> Am I just doing something horribly stupid?  These disks work great with ccd.

This doesn't have anything to do with the disks :)  It's simply RAIDframe
looking for component labels (which contain all sorts of information 
that RAIDframe needs to make your life easier :) ) and not finding any
valid labels.

> This is on an older copy of 1.4.2_ALPHA, on an alpha machine.. my raid.conf i
> s
> included below..
> 
> RAIDFRAME: protectedSectors is 64
> Hosed component: /dev/sd10a
> raid0: Ignoring /dev/sd10a
> raid0: Component /dev/sd13a being configured at row: 0 col: 1
>          Row: 0 Column: 0 Num Rows: 0 Num Columns: 0
>          Version: 0 Serial Number: 0 Mod Counter: 0
>          Clean: 0 Status: 0
> Column out of alignment for: /dev/sd13a
> Number of rows do not match for: /dev/sd13a
> Number of columns do not match for: /dev/sd13a
> /dev/sd13a is not clean!

There should have been a bit more output here... specifically, for /dev/sd14a
(RAIDframe should have been complaining about that one too...)

The above is normal for the first configuration.  The Examples section in 
'man raidctl' could perhaps be a bit more clear about that.

> START array
> # numRow numCol numSpare
> 1 3 0
> 
> START disks
> /dev/sd10a
> /dev/sd13a
> /dev/sd14a
> 
> START layout
> # sectPerSU SUsPerParityUnit SUsPerReconUnit RAID_level_5
> 108 1 1 5

For 3 disks, using 64 for sectPerSU will probably give the best performance.

Later...

Greg Oster