Thanks to everyone who replied. This mailing
list is great. Anyway, my inetd problem seems to
be a 2 part problem. The first part was that telnets
weren't allowed. This was happening before I did a
kill -1 on the inetd process. This problem is not
consistent, so its hard to figure out what was going on.
Some reasons given for the problem was a bad pty
or out of swap space. My question is, how do you get
a bad pty? And how do you tell if you have a bad pty?
Someone else said this is a known pty bug (ID 1014706).
I've got a call into Sun now about that. The second part
of the problem was the kill -1 to the inetd process. This
is a known bug and there is a workaround (ID 1024879).
The below fixed the second problem, but the first problem
is still sort of up in the air. The good news is that the
second problem ( kill -1 to the inetd process) has been
fixed in 4.1.
The below came from Debbie Eckel.
Quoted from the Customer Distributed BugsList (March 1990):
"Bug number: 1024879
Release: 4.0.3, 4.0.1
Synopsis: Inetd/telnetd restart problem
Description: If inetd is started from the shell, in.telnetd will
hang after an initial telnet connect has been accepted. This
inetd process must be restarted by init from /etc/rc (the normal
default case) for in.telnetd to work correctly."
Alternatively, rename /usr/etc/in.telnetd to in.telnetd.real and
substitue the following script for in.telnetd:
#!/bin/sh
#
# This hack is a workaround for bug 1024879. in.telnetd.real is
# the in.telnetd shipped with the system.
#
trap '' 21 22
exec /usr/etc/in.telnetd.real "$@"
Remember to chmod +x in.telnetd!"
And this from Frank Fiamingo
This is a known problem in SunOS4.0.x. Apparently if inetd is NOT
started by init it can't pass along certain environment variables
that SunOS's telnet is expecting. What always amazed us when we
saw this problem is that we could telnet from an HP into the sun,
but not from a Sun.
Thanks everyone,
Craig
J. Craig Manning
System Administrator for CSC
on site at NASA Ames Research Center, Moffet Field.
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:05:57 CDT