I've got raid1 running on all partitions but /boot with 2.2.10 and
raidtools-19990724. Now I'm thinking about recovery. I experimented with one
partition. Here it is in raidtab:
# /home
raiddev /dev/md7
raid-level 1
nr-raid-disks 2
nr-spare-disks 0
chunk-size 4
persistent-superblock 1
device /dev/hda7
raid-disk 0
device /dev/hdc7
raid-disk 1
It was originally built with hda7 as a failed-disk 1. I tried switching it
to 0 in raidtab to see what would happen. Apparently nothing. cat
/proc/mdstat still shows:
Personalities : [raid1]
read_ahead 1024 sectors
md2 : active raid1 hdc2[0] hda2[1] 264000 blocks [2/2] [UU]
md5 : active raid1 hdc5[0] hda5[1] 526080 blocks [2/2] [UU]
md6 : active raid1 hdc6[0] hda6[1] 66432 blocks [2/2] [UU]
md7 : active raid1 hdc7[0] hda7[1] 66432 blocks [2/2] [UU]
md8 : active raid1 hdc8[0] hda8[1] 34176 blocks [2/2] [UU]
unused devices: <none>
even after raidstop, raidstart, even reboot. It appears then, that raidtab
is only used by mkraid. Correct?
The reason for this is that it seems the failed-disk directive would be nice
for bringing the machine back up with a new disk after a failure. However,
the docs say that failed-disk cannot be first. What happens if hdc fails?
If I bring up the PC with a new hdc, I expect RAID would come up in degraded
mode of some kind, I could do a raidhotremove, partition hdc, and then
raidhotadd. Is this right?