markg@CAEN.ENGIN.UMICH.EDU (Mark Giuffrida) (04/25/89)
> Can anyone explain the rationale behind the -h option? Is there something >I'm missing? Are we just uniquely lucky in having routed work for us? Because some sites have a small limited number of networks that do not change very often. Why take up a process slot with Routed in this situation? Instead, you would run Routed in "chatter mode" on gateways nodes and use the '-h' for nongateway nodes. Note that this scenario buys you the grace of not maintaining static route tables. Mark Giuffrida University of Michigan
appel@ocf.Berkeley.EDU (Shannon D. Appel) (11/01/90)
We just installed 10.2, and we are having problems with the routed. Basically, on our gateway, we start up with /etc/routed -f -h and it goes fine for a minute or two and then crashes. The actual gateway machine can still talk to everything, but the rest of the machines going through it can not. Occasionally, connections will come back up on the other machines, but the only way to surely restore connections is to kill the routed and restart it. Anyone dealt with this problem out there? Shannon Appel
austin@METO.UMD.EDU (Austin L. Conaty) (11/01/90)
In reference to your routed problem, The gateway node should be running /etc/routed -f without the -h. This may clear up your problem. austin@meto.umd.edu
mcguire@math.uiowa.edu (Charlie McGuire) (11/01/90)
In article <APPEL.90Oct31105630@typhoon.ocf.Berkeley.EDU>, appel@ocf.Berkeley.EDU (Shannon D. Appel) writes: |> We just installed 10.2, and we are having problems with the routed. |> Basically, on our gateway, we start up with |> |> /etc/routed -f -h and it goes fine for a minute or two and then crashes.... The routed with SR10.2 is bad. There is a patch with a new routed available. Contact the support center and request a patch tape. *************************************************************************** * Charlie McGuire | INTERNET: mcguire@math.uiowa.edu * * Systems Programmer | mcguire@cs.uiowa.edu * * Computer Science Dept. | * * The University of Iowa | PHONE: (319) 335-2730 * ***************************************************************************
bergum@CIM-VAX.HONEYWELL.COM ("David Bergum") (11/02/90)
Direct mail bounced, so... Date: 31 Oct 90 18:56:30 GMT From: appel%agate%agate.uucp@ucbvax.Berkeley.EDU (Shannon D. Appel) Organization: ucb Subject: routed Message-Id: <APPEL.90Oct31105630@typhoon.ocf.Berkeley.EDU> Sender: apollo-request@umix.cc.umich.edu To: apollo@umix.cc.umich.edu We just installed 10.2, and we are having problems with the routed. Basically, on our gateway, we start up with /etc/routed -f -h and it goes fine for a minute or two and then crashes. That is what the -h is supposed to do. It tells routed to run four a couple routing updates until the routing table is stable and then exit. You don't want to use -h for a gateway; it is intended for a non-routing host. The actual gateway machine can still talk to everything, but the rest of the machines going through it can not. Occasionally, connections will come back up on the other machines, but the only way to surely restore connections is to kill the routed and restart it. Anyone dealt with this problem out there? Shannon Appel
appel@ocf.Berkeley.EDU (Shannon D. Appel) (11/02/90)
I got a lot of responses back on my last routed query saying that it was not working because I had /etc/routed -f -h instead of /etc/routed -f. Actually, my -h was a temporary hack, in an attempt to do anything:) to get the thing to work. Right now, I have it back with the correct /etc/routed -f and it still dies regularly. (Note: when I say die, the routed process is still up, but it does nothing in the way of routing. If it is restarted at this point, it works fine.) One person out there said that the routed for 10.2 was just bad, and that I should get the tape with the latest patch on it. In the meantime, does anyone have a temporary fix so that I can keep my non-gateway machines on the net while we are waiting for the patch? Shannon Appel
hj412fr@duc220.uni-duisburg.de (Martin Anantharaman) (11/02/90)
I got a lot of responses back on my last routed query saying that it was not working because I had /etc/routed -f -h instead of /etc/routed -f. Actually, my -h was a temporary hack, in an attempt to do anything:) to get the thing to work. Right now, I have it back with the correct /etc/routed -f and it still dies regularly. (Note: when I say die, the routed process is still up, but it does nothing in the way of routing. If it is restarted at this point, it works fine.) One person out there said that the routed for 10.2 was just bad, and that I should get the tape with the latest patch on it. In the meantime, does anyone have a temporary fix so that I can keep my non-gateway machines on the net while we are waiting for the patch? Shannon Appel Now, I did not really read the postings thta led up to this, BUT as I am successfully running a ATR gatewayed to a subdomain of the Internet, here are a couple of things that helped me: 1) Run routed interactively with -t to see what your routeds are up to. 2) Pay attention to your subnet-masks: When I did 1) I found that the routeds were basically ignoring the gateway routed, because they did not think it was on the same net. 3) Maybe use the -g flag on the gateway routed so it keeps offering its routes. Martin Anantharaman FB7, FG7 (Mechanik) Work: +49 (203) 379-3336 Universitaet -GH- Duisburg Home: +49 (203) 37 65 89 Lotharstr. 1 FAX: +49 (203) 379-3052 4100 Duisburg 1 E-Mail: hj412fr@duc220.uni-duisburg.de West Germany
krowitz@RICHTER.MIT.EDU (David Krowitz) (11/02/90)
Use the "route add default <gateway> 1" command to tell the non-gateway machines to forward any packet not intended for another node on the local network to the gateway. You can also use a similar command on the gateway itself to point to the next gateway on your route to the outside world. Our Apollo token-ring is connected to our local building ethernet, which in turn is connected to the MIT campus optical ring, which in turn is connected to everywhere else in the world. We don't use "routed" on any of our machines, and it works just fine. As long as your Apollo network only has one connection to the outside world, /etc/route works just fine. -- David Krowitz krowitz@richter.mit.edu (18.83.0.109) krowitz%richter.mit.edu@eddie.mit.edu krowitz%richter.mit.edu@mitvma.bitnet (in order of decreasing preference) '
appel@ocf.Berkeley.EDU (Shannon D. Appel) (11/03/90)
Got it working with the new patch of routed. Thanks for all the help. Shannon Appel