Hi,
In the raidtab file you posted, both of the member deivces were listed as
being "raid-disk 1". You want the first one to be raid-disk no 0 and the
second one to be raid-disk no 1.
Try using this raidtab:
-----
raiddev /dev/md0
raid-level 0
persistent-superblock 0
chunk-size 32
nr-raid-disks 2
nr-spare-disks 0
device /dev/sdb1
raid-disk 0
device /dev/sdc1
raid-disk 1
-----
Now, I've not used raid-0 (just raid-1) but the example raidtab file
provided in the raidtools-0.90 shows a raid-0 configuration for /dev/md1,
and it has a raid-disk no 0 and no 1 just like all the other raid
personalities.
By the way, the really important error message would be from "mkraid".
Everything else _will_ fail if mkraid has not done its magic. Unfortunately
mkraid does not emit nice error messages, so if you dive into the source to
find something, dive into mkraid to really understand any error-like
messages it is throwing at you.
- David Harris
Principal Engineer, DRH Internet Services
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]] On Behalf Of Chance Reschke
Sent: Friday, December 18, 1998 12:54 PM
To: [EMAIL PROTECTED]
Subject: Problems using RAID-0
Hi,
I'm having trouble setting up a RAID0 array using the
raid0145-19981215-2.0.36 kernel patch and the raidtools-19981214-0.90
tools.
I patched up and rebuilt the kernel, ran lilo and reboted. I created an
/etc/raidtab based on the example, and rebooted again. The kernel reports
the following in /var/log/messages:
Dec 18 09:41:29 rats kernel: md driver 0.90.0 MAX_MD_DEVS=256, MAX_REAL=12
Dec 18 09:41:29 rats kernel: raid0 personality registered
Dec 18 09:41:29 rats kernel: raid5: measuring checksumming speed
Dec 18 09:41:29 rats kernel: raid5: using high-speed MMX checksum routine
Dec 18 09:41:29 rats kernel: pII_mmx : 1024.128 MB/sec
Dec 18 09:41:29 rats kernel: p5_mmx : 974.217 MB/sec
Dec 18 09:41:29 rats kernel: 8regs : 717.423 MB/sec
Dec 18 09:41:29 rats kernel: 32regs : 431.673 MB/sec
Dec 18 09:41:29 rats kernel: using fastest function: pII_mmx (1024.128
MB/sec)
...
...
Dec 18 09:41:29 rats kernel: autodetecting RAID arrays
Dec 18 09:41:29 rats kernel: autorun ...
Dec 18 09:41:29 rats kernel: ... autorun DONE.
I then try running "raidstart -a", but it returns the error "/dev/md0:
Invalid argument", and prints the following to /var/log/messages:
Dec 18 09:41:58 rats kernel: (read) sdb1's sb offset: 8883840 [events:
00000000]
Dec 18 09:41:58 rats kernel: md: invalid raid superblock magic on sdb1
Dec 18 09:41:58 rats kernel: md: sdb1 has invalid sb, not importing!
Dec 18 09:41:58 rats kernel: could not import sdb1!
Dec 18 09:41:58 rats kernel: autostart sdb1 failed!
Dec 18 09:41:58 rats kernel: huh12?
My /etc/raidtab looks like:
raiddev /dev/md0
raid-level 0
persistent-superblock 0
chunk-size 32
nr-raid-disks 2
nr-spare-disks 0
device /dev/sdb1
raid-disk 1
device /dev/sdc1
raid-disk 1
I've tried it with devices /dev/sdb and sdc as well. I've tried using
fdisk to put a single partition on each disk before running raidstart.
I've tried doing this with persistent-superblock set to "1". I've tried
using mkraid. I've tried rebooting between each step. No dice.
I've run md arrays in the past with no problem, but the new fangled
approach has me stumped. I'd really appreciate any help rather than
having to spend hours or days pouring over the source.
Thanks,
Chance
