My route is
10.243.49.0/28 dev eth1 proto kernel scope link src 10.243.49.4
192.168.0.0/24 dev eth0 proto kernel scope link src 192.168.0.254
10.160.64.0/21 dev eth2 proto kernel scope link src 10.160.65.233
169.254.0.0/16 dev eth2 scope link
nexthop via 10.243.49.1 dev eth1 weight 1
nexthop via 10.160.64.1 dev eth2 weight 1
this is my route for 2 uplink
my squid is listening on this 2 routes ,This is OK for 2 uplinks up time.
But not OK for failed 1.I got net but slow and error got No route to host
Can I skip failed route with squid ?
Please help me
I believe this problem is about kernel route.
nexthop via 10.243.49.1 dev eth1 weight 1 <--- the first link
nexthop via 10.160.64.1 dev eth2 weight 1 <--- then second link
you setup router round robin. so if one link down, your kernel still send package to the down one. so you have to use script to propect your link down or up and rewrite route table to kernel.