Pekka Savola
2007-Apr-22 18:40 UTC
[CentOS] Centos5: RAID1 on root/boot? [+lilo mkinitrd issues]
Hi,
I did a command-line upgrade of a RHL73 server to Centos 5. It was a
bit rocky road, but in the end it was successful.
There's one thing that bugged me. I'm using software RAID1 consisting
of /dev/hd{a,c}. No LVM or anything fancy, a number of /dev/mdX
partitions, including the root (+/boot).
I'd have preferred to continue using lilo as it works more easily with
RAID1 root/boot setups. The deployment guide for Centos5 claims to
support RAID1 boot partitions (RHEL4 didn't, and provided a pointer a
site in UK which no longer works for more info), it's not clear how
exactly a redundant boot block configuraton is achieved with that --
so that if you remove the primary drive, it'll still boot.
Anyone have pointers for this?
[I've gone through a number of sites with various hacks how to install
grub on both disks, but I'm not quite sure which ones are valid, and
whether those are needed anymore]
.. read on if you're interested about the lilo/mkinitrd problem ..
However, lilo doesn't really work, the root cause as far as I've been
able to go is that '/dev/root' isn't recognized and as such not
mounted. With grub this works.
The init script in the initrd includes the following (automatically
generated):
===8<=echo Creating root device.
mkrootdev -t ext3 -o defaults,ro /dev/md0
echo Mounting root filesystem.
mount /sysroot
===8<=
And looking at the vast amount of RH bugs about this, it seems this
root autodetection has caused significant grief, e.g.,
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=197701.
However, no bugs on this (or similar) have been filed under RHEL5.
Any thoughts? Anyone run with similar setups?
*)
=======8<===md: autorun DONE
..
Creating root device
Mounting root filesystem
mount: could not find filesystem '/dev/root'
Setting up other filesystems
Setting up new root fs
setuproot: moving /dev failed: no such file or directory
no fstab.sys, mounting internal defaults
setuproot: error mounting /proc: no such file or dir
setuproot: error mounting /sys: no such file or dir
...
and finally, a kernel panic.
=======8<=====
--
Pekka Savola "You each name yourselves king, yet the
Netcore Oy kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings
William L. Maltby
2007-Apr-22 19:12 UTC
[CentOS] Centos5: RAID1 on root/boot? [+lilo mkinitrd issues]
On Sun, 2007-04-22 at 21:40 +0300, Pekka Savola wrote:> Hi, > > I did a command-line upgrade of a RHL73 server to Centos 5. It was a > bit rocky road, but in the end it was successful. > > There's one thing that bugged me. I'm using software RAID1 consisting > of /dev/hd{a,c}. No LVM or anything fancy, a number of /dev/mdX > partitions, including the root (+/boot). > > I'd have preferred to continue using lilo as it works more easily with > RAID1 root/boot setups. The deployment guide for Centos5 claims to > support RAID1 boot partitions (RHEL4 didn't, and provided a pointer a > site in UK which no longer works for more info), it's not clear how > exactly a redundant boot block configuraton is achieved with that -- > so that if you remove the primary drive, it'll still boot. > > Anyone have pointers for this? > > [I've gone through a number of sites with various hacks how to install > grub on both disks, but I'm not quite sure which ones are valid, and > whether those are needed anymore]This has been touched on several times in the CentOS mailings. Did you also search CentOS? If not, ISTR that there are answers there. For grub, of course. <snip> HTH -- Bill
Maybe Matching Threads
- mkinitrd doesn't know how to handle encrypted CCISS devices - RHEL5.4 and Xen4.1.2
- CentOS 4.4 LILO Raid (also e2fsprogs in general)
- unable to recover software raid1 install
- keytab-lilo: update to support kbd 2.0.3 format
- keytab-lilo: update to support kbd 2.0.3 format