Hi,
I would like to set up access to Internet via two providers. When they both
works OK I use
ip route add default scope global nexthop via x.x.x.x dev eth0 weight 1
nexthop via y.y.y.y.y dev eth1 weight 1
Next, I use script that regularly pings their upper providers for see if some
of providers are down. If one is down I want to guide all communications via
other interface. If second one is down -
ip route del default
ip route add default via x.x.x.x
It works OK but at that moment I can''t ping any nonlocal address thru
eth1 so
I can''t check when provider come up.
ping z.z.z.z -I eth1 - gives nothing
Tcpdump shows that eth1 broadcast arp request for nonlocal address ?!?
Tables for both interfaces are setup correctly and it seems that ping checks
only main table.
At same time when haven''t default route via provider and that provider
is up,
all communication initialized by peers on Internet going on nondefault
interface, even pings, works OK.
When the default routes manually sets back to go thru both ones it comes back
to normal .
One more strange thing is that my script works OK for over one year until I
have to change netmask and default gw for one provider.
TIA, Darko
