>>>> I am running raid 1 on a centos 4.4. One of the harddisk (sda1) >>>> failed. How can i carry on running the server using only sda2? >>> >>> Generate a grub floppy and use that to load the grub menu from >>> the sdb (probably now sda) disk. >>> >>> If you are really talking about sda1 and sda2, those are >>> partitions on the same disk. >> >> Is there a detail step by step howto? The raid 1 has no LVM. just >> md0, md1 and md2. md0 is /boot, md1 is swap and md2 is the >> storage. I had replace sba with a new disk. I tried to boot up >> and it says kernel panic. How am i going to reconstruct the raid >> and sync sdb to sda? > > It might be easier to swap the old sdb into the sda position so > you'll boot from it, but you should also be able to boot the > install cd with
If swapped and booted, and got a kernel panic error.
> 'linux rescue' at the boot prompt, let it detect and mount your > system (which will be the 'broken' raid devices with their single > members),
If i use linux rescue, The 3 mds I created are gone. /cat /proc/mdstat says Personalitlies: [raid0] [raid1] [raid5] [raid6], no longer Personalities : [raid1]
Perhaps your raid wasn't really working the way you thought before. From the rescue boot, does fdisk show the 3 partitions on the old disk with type 'fd'? Can you mount the old /boot and / partitions somewhere by hand? You should be able to do this with the /dev/sda1 and /dev/sda3 device names if the md devices aren't detected at boot.
cat /proc/partitions still shows me the 3 partitions.
Does fdisk say that they are type 'fd'(raid autodetect)?
I actually copied /boot to the "replaced disk" and it is able to boot up, but without any filesystem, so i guess the boot is still intact. So do i need to mount /boot and /?
If you can get the original partitions to be detected as their md devices you should fdisk matching partitions on the replacement disk, then 'mdadm --add ...' to add them and they will automatically sync up.
mdadm --create --verbose /dev/md0 --level=1 --raid-devices=2 /dev/sda1 /dev/sdb1 mdadm --create --verbose /dev/md1 --level=1 --raid-devices=2 /dev/sda2 /dev/sdb2 mdadm --create --verbose /dev/md2 --level=1 --raid-devices=2 /dev/sda3 /dev/sdb3
If you already had raid devices on one of the disks you should not have had to --create them again. The original ones should have been detected and you should have been able to --add new matching partitions.
I created them as md(s) are not longer there.
After that i reboot and got the kernel panic again.
md: considering sdb1 md: adding sdb1 md: created md0 md: bind<sda1> md: running: <sdb1><sda1> raid1: raid set md0 active with 2 out of 2 mirrors md: ... autorun DONE md: autodetcting RAID arrays md:mautorun ... Creating root device Mounting root filesystem switching to new root switchroot: mount failed: 22 umount /unitrd/dev failed: 2 Kernel panic
When you --create a new raid it will start to sync the mirrors. It may have done this the wrong direction, overwriting your old contents. Can you still do a rescue mode boot, mount /dev/sda3 (or sdb3 if the old drive is in the 2nd position) and see the contents?
I am unable to mount sda3.
# mount /dev/sda3 /mnt/part3 mount: Mounting /dev/sda3 on /mnt failed: Invalid argument
sdb is not longer detectable.