LARTC - Jul 2004 - dead gateway detection/failover without load ballancing?

Greetings,

I wanted to ask the list a question about dead gateway detection.  I''ve
followed the wonderful writeup by Christoph Simon on how to use Julians patch
for dgd on two loadballanced connections, but I don''t want them
loadballanced.  How can I setup two default gateways but always have one be the
preference unless it goes down.  Once the dgd code sees that have it switch
until it become available again.  I''ve tried just adding two default
routes to the same route table and it seems to switch between them randomly
whether I use the "nexthop" code or put them both on their own lines. 
I''ve also tried using the weight code but obviously that just gives
more preference to one connection but some connections will still go the other
route.  I''ve also tried using a higher metric on the second gateway,
but that doesn''t work at all.  The second route never gets taken even
if the first goes down.  According to Julian''s writeup this makes
sense, he states that they must have the same metric for the dgd to work.

My second question is about proto static.  It is my understanding that proto
static is suppose to keep the route in there even if the interface goes down.  I
run dhcp on my external interfaces and when the interface goes down and the dhcp
tries to renew it deletes all my proto static lines related to that interface. 
Is this a route code or a dhcp code problem?  Is dhcpcd manually removing those
lines even though they are not in the main route table?

thanks for any help you can provide,

joshua
_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/