I have a three-computer internal LAN: two Debian (potato) boxes and a
Windows NT machine. For now, we won't worry about the NT computer.
Both Linuxes have identical /etc/hosts files:
127.0.0.1 localhost
192.168.0.1 septictank
192.168.0.2 sewage
septictank is my diald server/gateway/firewall machine for the rest of
the LAN (and sewage is my workstation).
Both machines are using OpenSSH-3.0.2p1 that I self-compiled and
installed (with the same configure options).
Here's the strange part:
Ssh'ing from septictank to sewage can be done using either of the
following commands:
ssh sewage
ssh 192.168.0.2
and the behavior is as expected (i.e. diald is not triggered).
However, if I try to ssh from sewage to septictank, I have to use the
following command:
ssh 192.168.0.1
else diald is triggered (in other words, "ssh 192.168.0.1" from sewage
will login to septictank immediately. Using "ssh septictank" will
cause diald to bring up a ppp connection, and ssh will wait until ppp
is up and running).
Two files I thought might be of interest are /etc/nsswitch.conf and
/etc/host.conf. On both systems, however, these files are the same
(Debian's default).
Finally, one more interesting point that may or may not be related: I
have not configured the Windows NT box to recognize the symbolic
hostnames "sewage" and "septictank". So when I connect to the Linux
boxes from the NT boxes I always use their numerical IP addresses.
Anyway, I use a program called "putty" for ssh on NT. When I ssh from
NT to 192.168.0.1 (septictank) the connection is established
immediately. However, if I ssh from NT to 192.168.0.2 (sewage), the
connection will go through, but takes forever---the modem might be
dialing, but I'm not sure (forgot to check that before sending this
email).
Anyway, thanks for any thoughs or suggestions!
Matt
--
Matt Garman, [EMAIL PROTECTED]
``I ain't never seen no whiskey, the blues made my sloppy drunk!''
-- Sleepy John Estes, ``Leaving Trunk''
