If the /etc/ssphostname file gets corrupted on a STARFIRE domain, then netcon
will be unable to connect to it, getting stuck in a loop. When attempting a
netcon session you get the following looping errors from it:

  cvcd host daemon unavailable
  cvcd host daemon unavailable
  Jun 20 14:47:19 Starfire1 /sbin/cvcd-(HOSTNAME:Starfire1):
     connection request from illegal host
  cvcd host daemon unavailable
  cvcd host daemon unavailable
  cvcd host daemon unavailable
  .
  .
  .

NOTE: It is no use trying to boot single-user, as this is affected too!

There are 3 ways to fix this:

1)	Telnet into the domain as root, then correct /etc/ssphostname.
	(only works if you've previously commented CONSOLE=/dev/console out
	of /etc/default/login AND the system is up enough to run in.telnetd)

2)	Telnet into the domain as a user, su to root, then correct
	/etc/ssphostname.
	(only works if the system is up enough to run in.telnetd)

3)	Panic the domain, boot net or boot cdrom, fsck & mount the root
	filesystem, then correct /etc/ssphostname.

NOTE: When running training courses where user accounts are never set up:

To aviod wasting time booting from CDROM, always ensure that the domains have
CONSOLE=/dev/console *commented* *out* of /etc/default/login.