On Sun, Mar 23, 2003 at 12:38:34AM -0800, philo vivero wrote:
> On Sat, 2003-03-22 at 21:50, Marty Shannon, RHCE wrote:
> > > > > I added a second IP address / interface (eth0:1)
> > > > >    selfcheck request timed out.  Host down?
> > > > > If I shut down eth0:1 amcheck works.
> > > > If you're on a Redhat-based system, edit
> > > > /etc/sysconfig/network-scripts/ifcfg-eth0:1 (or whatever has the
> > > > definition for interface eth0:1) and delete the GATEWAY= entry from it.
> > > [...]
> > > > The network script that brings up that interface sets a new default
> > > > gateway that then causes routing to not function.
> > > Thanks for the suggestion - I tried it, but it didn't work.  Is there 
> > > something else I can try?
> > Did you actually look at the output of "netstat -r" and/or "route"?
> 
> I'm going to have to agree with Marty here. Do "route -n" and make sure
> your default gateway belongs to eth0 and not to eth0:1.
> 
> It really, really sounds like this is your problem.

When I've had systems with two gateway's (intentional) they were selected
in a round-robin fashion ... until a valid route to a destination was
detected.  Then that interface was used for that destination.

To establish specific interface -> specific hosts routes you can create
"static routes".  I had to set up boot time routines to do this.

It may be necessary to flush your routing tables to see any effects.

-- 
Jon H. LaBadie                  jon AT jgcomp DOT com
 JG Computing
 4455 Province Line Road        (609) 252-0159
 Princeton, NJ  08540-4322      (609) 683-7220 (fax)