On Fri, 9 Oct 1998, Geof Goodrum wrote:
> I was just going to ask the same question. I've been using a RAID0 and
> two RAID5 arrays created with raidtools 0.41 running under RedHat
> 5.0/2.0.31, but I started getting seg faults with 'mdadd -ar' (libc
> related?) after an upgrade to RedHat 5.1. I patched a vanilla 2.0.35
> kernel with raid0145-19981005-C-2.0.35, installed raidtools-0.90, created
> raidtab, did the 'mkraid --upgrade' on the arrays (RAID0 didn't like it -
> "magic-less superblock", but RAID5 seemed to upgrade without complaint),
> but like you, I get "invalid argument" for the arrays whenever I do a
> raidstart. The only time I can get /dev/md0 (RAID0) to activate is when I
> run 'mkraid -force'; the RAID5's fail.
the old RAID0 array is a 'superblock-less' array. This means you have to
mkraid it after every bootup. (mkraid starts it automatically). After you
stop it, you have to mkraid again. This is because the configuration is
only in the /etc/raidtab file and you have to re-create the array again
and again.
if you have lots of data on your RAID0 array and cannot do a backup, then
this is how it goes. But if you can recreate the RAID0 array after backing
up all data in it, you can put this into your raidtab:
persistent-superblock 1
and the (newly created, with destroying all data content) RAID0 array will
be autostartable just like RAID5.
-- mingo
