The supplied zm script for startup and shutdown was being invoked a bit late on a stock CentOS 3.4 box. Basically MySQL is being shutdown before Zoneminder and this was causing Zoneminder to choke when it tried to access the closed database, leading to problems umounting disks prior to power off.
Solution:
Make sure Zoneminder is being shutdown before MySQL. In my CentOS 3.4 box a stop priority of 30 is fine:
Code: Select all
#!/bin/sh
# description: Control ZoneMinder as a Service
# chkconfig: 2345 99 30