Seen yesterday at RRI:
RRI uses a 1.5 TB external RAID array as their primary storage device. The device used to be mounted at boot time with an entry in /etc/fstab. The problem is that if the device is powered off during the boot, Linux will halt the boot and dump the user into a Ctrl+D environment to 'fix' the missing drive problem. This is undesirable in this case because end-users don't know what to do. Switching this to using an automounter also had an undesireable effect, however: The filesystem is an ext3 type fs and keeps track of mount count between fscks. The max number allowed had been hit, and the automounter appears to stall in this case.
The work-around is simple: run fsck on the drive, but is there a better way?