Vodka status

[Aaron Turner <aturner@linuxkb.org>]

Well I've been here for almost 2 hours now and I'm getting tired.  Good
news/bad news:

Good news:
1)  Rewrote the /etc/rc.d/init/named script.  It now no longer gives a
bogus error message that prevents it from working.  Not sure what was
"broken", but it works fine now.

2)  No corruption on the disks afterall.  Ran fsck numerous times and
never found a thing wrong.

Bad news: 

1)  The system will *not* go all the way down.  It just hangs after
stopping kerneld and init reporting all the processes at this runlevel are
killed.  IE, it hangs just *before* unmounting all the filesystems.  At
least it doesn't corrupt anything.  I'm not sure what the problem might
be, I'm going to ask around on the kernel-list and linux-raid lists.

This means that after an init 0 or init 6, someone has to physically hit
the reset button on the box.  It will reboot, fsck the drives, and re-sync
the RAID partitions.  At least this isn't NT, so we shouldn't have to
reboot often. :-)  Oh, it takes about 10 minutes of the drives being idle
to sync all the RAID partitions.  If you're running anything that hits the
drives hard, the time to re-sync will increase since the sync is a 'nice'
process.

The good news is that NaviSite is adding a feature that allows you to
access a web server and remotely powercycle a box.  It is intended for NT
(obviosly) but it should be a doable interum solution.  I'll ask around to
get more info.

Later people.