naim@nucsrl.UUCP (Naim Abdullah) (10/16/86)
I need some advice/information about a csnet/arpanet/smail/sendmail/
pathalias combination and I hope somebody out there can help. Please,
I am looking for illumination here, so no nasty replies about how obvious
the solution is.
First, some background: We use a vax 780 running 4.3bsd for netnews/uucp
mail. Currently we are using the standard berkeley provided, sendmail.cf
file for a site that only has uucp connections. We are not using any
smart mailers/pathalias etc.
The computer center, at this university, will soon get an arpa net connection
on a uvaxII running ultrix. Our vax 780 will be on an ethernet with this
uvax so presumably we can get arpanet access by using rsh and having the
uvax forward us our arpa net mail etc. (correct me if I am wrong..).
I was thinking of installing smail/pathalias on the vax 780 so that we
can get domain based uucp. This shouldn't interfere with our arpa
connection. All fine and dandy, so far.
Now the complication: We also have a vax 785 running vms that has a csnet
connection. It is using the pmdf phone net software provided by ok-state.
It has been proposed that we switch the csnet connection from the vms
machine (the 785) to vax-unix (the 780). This way, the vax 780 will be
our main connection with the outside world (arpanet/csnet/uucp). The 780
can forward mail, to the vms machine from the outside world (we have
a vms uucp that can talk to the berkeley uucp).
My questions:
1) Does this necessarily mean that we will have to install mmdf on the
780 and (as a consequence) throw away all the sendmail/smail stuff ? In
other words, is mmdf a prerequisite for a vax-unix site with a csnet
connection ? Being triply redundant, is there a unix site out there that
has a csnet connection, but is *NOT* running mmdf ?
2) If we are forced to install mmdf to support csnet for the 780, would
it still be possible to do domain based uucp ? Can pathalias be integrated
into the mmdf mess ?
Sorry for the rather rambling nature of the posting, but I did not want
to leave any ambiguities.
Thanks in advance (as they say).
Naim Abdullah
Dept. of EECS,
Northwestern University,
{ihnp4, chinet}!nucsrl!naim