[KLUG Members] linear raid troubles

[Adam Williams]

>I have 2 40 gig hard drive that were configured as a linear raid array
>to have one large 80 gig device.  This is in a system where I had some
>hardware troubles and I ended up losing the entire operating system
>drive and unfortunately, the entire /etc/ directory.  So there is lots
>of important information on this device that I am afraid may be lost. 
>So I have attempted to recreate the device with the following raidtab.
>raiddev /dev/md0
>  raid-level            linear
>  nr-raid-disks         2
>  nr-spare-disks        0
>  chunk-size            4
>  persistent-superblock 1
>  device                /dev/hdg1
>  raid-disk             0
>  device                /dev/hdh1
>  raid-disk             1
>It is not starting up.
>If I execute 
>raidstart /dev/md0
>my command prompt comes back to me and /proc/mdstat looks like
>[root@edgesrv samba]# cat /proc/mdstat
>Personalities : [linear]
>read_ahead not set
>unused devices: <none>

Look at dmesg.  raidstart should write something to the kernel ring
buffer.

Make sure you have your chunk size the same, or else......

>if I execute 
>raidstop -a
>I get 
>[root@edgesrv samba]# raidstop -a
>nothing to do!
>usage: raidstop [--all] [--configfile] [--help] [--version] [-achv]
></dev/md?>*
>So, I may have tried to start the raid device with the wrong parameters
>before.  I can't remember exactly what I changed, could have been
>raid-level 0.  Have I lost all of the information that used to be on
>this 80 gig raid device.  Please tell me all is not lost, humor me for a
>while.

Well, the data is still there. Accessing it is the problem! :)

Recreating an identical raidtab and rerunning raidstart WILL work to
bring the drives back as no relevant meta-information is stored in the
raid volumes themselves.