Hello
I had a drive failure on the second drive in a mirrored pair. After
rebooting the system ( and correcting the jumper settings). The drive came
back up fine. The boot and swap partitons came up fine, but the root
partition noticed the drives were out of sync and only used the first
drive. It then says, " no spare drives for reconstruction".
I have since then also run badblocks -w on this partition to check it, no
errors reported. Then ran dd if=/dev/hda3 and of=/dev/hdc3 to repopulate
it( of course I did this with hda3 mounted and being used, so its probably
not the most consistent filesystem).
How do I force a reconstruction of /dev/hdc3 ? Preferably in multi user
mode. Sorry the Software-Raid-howto seems to refer to older versions or
utilities not used now that raid is in the kernel.
a) What I want to do( but ckdisk is no longer used )
ckraid /etc/raid1.conf --fix --force-source /dev/hda3
Of course I don't have ckraid or raid1.conf
b) Or in single user mode,
dd if=/dev/hda3 and of=/dev/hdc3
mkraid raid1.conf -f --only-superblock
mkraid.c does not have --only-superblock ( so I assume it's not there)
c)
raidstop /dev/md2
mkraid /dev/md2 --force-resync --really-force
raidstart /dev/md2
Once again mkraid.c does not appear to have -force-resync.
d) raidhotadd /dev/md2 /dev/hdc3
Don't have raidhot add.
Setup, Raid installed straight from install
RedHat 6.1 -> 2.2.12-20smp
Two IDE disks on separate IDE channels
/md0 hda1 hdc1 # /boot
/md1 hda2 hdc2 # Swap
/md2 hda3 hdc3 # / , hdc3 now invalid running off of hda3
only.
When I discovered the drive down, the system was complaining about hdc not
responding. BadBlocks reveal the drive is okay( for now). I suspect the
problem was caused because I removed a defective tape drive a week earlier
and did not reset the master/slave jumper settings on hdc.
Thanks, Peter
[EMAIL PROTECTED]
