The problem is "probably" that the host is changing the route to go to a host x. The host keeps a cache (the routing cache) where it stores through what interface, ip blah blah the host reached another host. This is kept alive for a while, and then flushed to make another routing decision to the same destination host. Probably that's why ssh could get "hung" anytime. On 12/8/05, Dave Weis <djweis@xxxxxxxxxxxxxxxxxx> wrote: > > I have a machine with two routes to the internet doing load balancing on > the connections and NAT for internal machines. For inbound ssh connections > and outbound connections to anything, it will occasionally "lose" packets. > If I Control-C out of ssh and try again, it usually works, but not always. > After a few tries it does connect, but frequently stalls and hangs > permanently. I did add rules to try and keep connections stuck where they > started: > > ip rule add from 200.200.200.1 table eth1 > ip rule add from 192.168.0.2 table eth2 > > In /etc/iproute2/rt_tables: > 201 eth1 > 202 eth2 > > ip rule list does show the rules: > 0: from all lookup local > 32764: from 192.168.0.2 lookup eth2 > 32765: from 200.200.200.1 lookup eth1 > 32766: from all lookup main > 32767: from all lookup default > > Is there some other command I can use to see if they really are sticking > where they should? > > Thanks > dave > > > -- > Dave Weis > djweis@xxxxxxxxxxxxxxxxxx > http://www.internetsolver.com/ > > _______________________________________________ > LARTC mailing list > LARTC@xxxxxxxxxxxxxxx > http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc >
