Newsgroups: comp.sys.apollo
Path: utzoo!utgpu!cunews!rmallett
From: rmallett@ccs.carleton.ca (Rick Mallett)
Subject: Re: Rotting routing tables update.
Message-ID: <1991Mar19.132326.22799@ccs.carleton.ca>
Summary: bug in SR10.2 routed causes headaches
Keywords: routed tcp/ip ftp telnet
Organization: Carleton University, Ottawa, Canada
Date: Tue, 19 Mar 1991 13:23:26 GMT

Many thanks to all the folks who responded to my plea for help with 
routing tables that disappeared shortly after bootup. Thanks especially 
to Charles Fuller who recalled that the SR10.2 route daemon had been 
patched at some point after initial release of 10.2. Sure enough the 
release notes for psk7 contained the following info:

> 4.3.2  routed Bug Fixed in TCP/IP
>
> The routed command was timing out active physical interfaces.  We've
> modified routed to prevent it from timing out, and therefore marking
> "down", interfaces that are configured "up" with the ifconfig command.

which pretty well summed up my situation. Of course the reason that I had
been ignoring psk7 was that you can't apply it on top of psk5 and it seemed
to mainly consist of support for the 400 series machines which I doubt I'll
ever want to purchase. Still it was an important lesson to read the release
notes thoroughly. I ended up solving the problem by grabbing routed off an
SR10.3 machine and it seems to be working fine. 

In my message I had pointed out that I was using Apollo netid to isolate
my network from some other machines on the same cable. Some people pointed
that this is not considered proper by the vendor and that network cells are
the proper way to achieve the same result. I do agree but in this case the
separation is only required for a few weeks while I do the conversion to
SR10.2 and the point is that I don't want to convert the old registry and
its hundreds of inactive accounts. Anyway I want to report that it seems to
be working fine in this context (i.e. SR10.2 with SR9.7 and a few 10.2
nodes on net 0), in case anyone else needs to do the same.