I have an apparently 100% working server with a pair of MD devices set up as
mirrors.

(The system is running Debian Linux with the 2.4.25-1-686 kernel. The mdadm
package version is 1.7.0-1; mdam --version says v.1.7.0.)

This system has been working properly for months. A few days ago I upgraded
the CPU and motherboard. I was careful to make sure that the drives lined up
in the same place (old hde becomes new hde, etc.), so the hardware upgrade was
trivial (zero software config work was required to get it back up and
running). It's been problem-free for about 3 days and I tested it thoroughly
to make sure all the major devices (network, drives, etc.) were discovered and
working, and that the stuff I use it for still works. No problems.

Then, I did a bunch of software updates last night via aptitude and rebooted
to make sure everything was OK. It seemed fine.

Today, though, I got a strange root email:
/etc/cron.daily/raidtools2:
lsraid: Unable to open device "/dev/ide/host2/bus0/target0/lun0/part1": No
such file or directory
lsraid: Unable to open device "/dev/ide/host2/bus1/target0/lun0/part1": No
such file or directory
lsraid: Unable to open device "/dev/ide/host2/bus0/target0/lun0/part3": No
such file or directory
lsraid: Unable to open device "/dev/ide/host2/bus1/target0/lun0/part3": No
such file or directory

Well, that's accurate; "ls -l /dev/ide" says:
ls: /dev/ide: No such file or directory

I looked at the raidtools2 script, and it's just running "lsraid -A -f -a
/dev/md0" and "lsraid -A -f -a /dev/md1". When I run those I get the same
error messages that were in the email. Hmm.

"more /etc/raidtab" says:
raiddev /dev/md0
raid-level 1
nr-raid-disks 2
nr-spare-disks 0
chunk-size 4
persistent-superblock 1
device /dev/hde1
raid-disk 0
device /dev/hdg1
raid-disk 1

raiddev /dev/md1
raid-level 1
nr-raid-disks 2
nr-spare-disks 0
chunk-size 4
persistent-superblock 1
device /dev/hde3
raid-disk 0
device /dev/hdg3
raid-disk 1

Which all makes sense and agrees with my notes from when I set this up
originally.

However, more /proc/mdstat says:
Personalities : [raid1]
read_ahead 1024 sectors
md0 : active raid1 ide/host2/bus0/target0/lun0/part1[0]
ide/host2/bus1/target0/lun0/part1[1]
56128 blocks [2/2] [UU]

md1 : active raid1 ide/host2/bus0/target0/lun0/part3[0]
ide/host2/bus1/target0/lun0/part3[1]
78975424 blocks [2/2] [UU]

unused devices: <none>

My question is, where does this stuff in /proc/mdstat come from and how do I
fix this? Am I missing some device files that ought to be there (/dev/ide/...)
or is there some old config data somewhere (MD superblock) that needs to be
updated?

Since the system appears to work fine right now I don't want to just futz
around trying to fix it, but I also don't want to leave it in a
not-quite-happy state lest I discover that one day a drive fails and say "oh
no, the mirror has been broken for a year and I don't actually have a mirror
at all."


Thanks in advance...