[FC6] Using a software RAID0 created by SuSE 10.0
lanas
lanas at securenet.net
Tue Mar 20 14:00:25 UTC 2007
Folks,
I'm now using Fedora Core 6 and I'd dearly like to use an existing
RAID0 that was created when the system was running SuSE 10.0. In fact,
SuSE is still installed so I can compare both systems regarding how
they use the RAID0.
Now it looks like there are many things which are similar. The
output of dmraid with several switches such as '-r', '-l' are the same
in FC6. There's also a md0 device in FC6. There are even in FC6 dome
mapper devices like:
crw------- 1 root root 10, 63 mar 19 16:18 control
brw-rw---- 1 root disk 253, 0 mar 19 16:18 nvidia_ffcdjdfc
brw-rw---- 1 root disk 253, 1 mar 19 16:18 nvidia_ffcdjdfcp1
So far so good. I think.
I looked at the FC6 rc.sysinit bootscript a bit to see how it setups
a RAID and after making some script tests around the resolve_dm_name()
and get_numeric_dev() functions I found out that the following command
was executed in the case of this system:
/sbin/dmraid -ay -i -p nvidia_ffcdjdfc
Moreover the output of that FC6 command is looking very positive:
"RAID set "nvidia_ffcdjdfc" already active"
But, whereas I use the following under SuSE to mount the
single-partition RAID0, it fails on FC6 (I've created the /raid mount
point in FC6):
mount -t ext3 /dev/md0 /raid/
I've also tried some variations such as:
mount -t ext3 /dev/mapper/nvidia_ffcdjdfcp1 /raid/
mount /dev/mapper/nvidia_ffcdjdfcp1 /raid/
But they all fail with:
"mount: wrong fs type, bad option, bad superblock [...]"
It really looks like both systems are perceiving the RAID0 in a very
similar way. What could be missing in the procedure I use in FC6 that
prevents from actually mounting the RAID0 ? I don't know much about
RAIDs and I'm careful to not try anything that would damage the actual
stored RAID0 data.
Any comments/suggestions welcomed - Thanks !
Al
More information about the fedora-list
mailing list