--------------------
GOAL
--------------------
RedHat 7.0
Kernel 2.4.0-test10 (also have tried 2.2.17 + ide + raid patches)
'/' mounted on RAID-5
--------------------
PROBLEM
--------------------
* When booting with newly created '/dev/md0' as '/', it is reported
corrupt and needs a manual fsck
* The manual fsck reports millions (almost) of:
inode 12156 is in use, but has dtime set. Fix?
inode 12156 has imagic flag set, clear?
...
And various other sundry errors.
* After fsck'ing it (holding the enter key down for an hour), and
rebooting, kernel stops at "Warning: unable to open initial console"
I'll describe my process below. I'm assuming I'm doing something
wrong between the "copy your old root over to the raid" and the "boot
with raid as your new root" steps. I've read that Software-RAID howto
and thought I following it word-for-word, but I must be missing
something.
--------------------
PROCESS (how I got the problem)
--------------------
- Install RedHat 7.0 to hda2.
- Compile & boot new kernel (I am currently trying with 2.4.0-test10,
but have also tried 2.2.17+ide+raid patches).
- fdisk /dev/hda /dev/hdc /dev/hde /dev/hdf (add 17000M "fd" linux
raid partition)
- vi /etc/raidtab...
raiddev /dev/md0
raid-level 5
nr-raid-disks 4
nr-spare-disks 0
persistent-superblock 1
parity-algorithm left-symmetric
chunk-size 32
device /dev/hda6
raid-disk 0
device /dev/hdc2
raid-disk 1
device /dev/hde1
raid-disk 2
device /dev/hdg1
raid-disk 3
- mkraid /dev/md0
- mke2fs -b 4096 -R stride=8 /dev/md0
- init 1
- cp -xa / /mnt/md0
(During the cp I get a bunch of errors like the following:)
"EXT2-fs error (device md(9,0)): ext2_new_block: Allocating block in
system zone - block = 416"
"EXT2-fs error (device md(9,0)): ext2_free_blocks: freeing blocks in
system zones - Block = 419, count = 5"
- cp finishes. So I edit the /mnt/md0/etc/fstab and set
"/dev/md0 / ext2 defaults 1 1"
- then I edit /mnt/md0/etc/lilo.conf to set:
image=/boot/vmlinuz-2.4.0-test10
label=2.4.0-t10md0
read-only
root=/dev/md0
- '/sbin/lilo -v' runs fine
- Reboot.
- For some reason the kernel mounts the /dev/md0 and thinks that it is
very corrupted (even though I had a clean shutdown). A check ensues,
in which hundreds of thousands (not kidding) of the following errors
are fixed (by me holding down enter key):
inode 12156 is in use, but has dtime set. Fix?
inode 12156 has imagic flag set, clear?
...
- This goes on forever until it finally says it's done, and I hit
'exit'.
- Upon reboot, I just get 'Warning: unable to open an initial
console'.
--------------------
DETAILS:
--------------------
* Four Seagate 20gb barracudas (ultra-66)
* Two are on the Intel i815 Mobo (onboard LAN / vid)
* Two are on the Promise Ultra66 controller (2.00b18 firmware) <--
that's why I use the ide patches when using 2.2.17
* raidtools-.90-13.rpm (plus .90 in the kernel)
Best regards,
Dan Browning
Network/DB Admin
Cyclone Computer Systems
_______________________________________________
Redhat-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list
