We have had numerous problems with TCP start over the years (especially when
there is an abnormal system end, or some user doesn't sign off before the
interactive subsystem is ended, et cetera.
As such, we have our own application that terminates each TCP service in
order, and then starts each service, checking the result of each command and
retrying when necessary.
On systems where signon choices are 1) use TELNET; or 2) drive 150 miles to
signon to the console and then start TCP, we don't rely on the simplistic
IBM-supplied approach. Weekend upgrades, off-hours work, people not showing
up at the hosting site... all of these things have confirmed the need for a
solid approach. Shame on IBM for supplying such a simplistic approach on
such a wonderful platform! If AIX (or any of the other *IX systems) were so
lax about TCP startup, their systems would never have gotten off the ground!
"The direct use of force is such a poor solution to any problem, it is
generally employed only by small children and large nations."
-- David Friedman