Freebsd change default Internet channel route

Question

I have two Internet channel and Gateway on freebsd. When I switch channel with the command route change default chan2, the command netstat -nr shows changed default route. But traceroute shows that the packets go through the old route chan1.

Example:

$netstat -nr 
Routing tables Internet: Destination Gateway  Flags    Refs   Use  Netif Expire
                         default     xxx.xxx.183.54 US 0 8432    em3

$sudo route change default xxx.xxx.144.125 
change net default: gateway> xxx.xxx.144.125

$netstat -nr
Routing tables Internet: Destination Gateway Flags Refs Use  Netif Expire
                         default     xxx.xxx.144.125   US  2  16450  em3

BUT

$ traceroute 8.8.8.8
traceroute to 8.8.8.8 (8.8.8.8), 64 hops max, 52 byte packets
 1  xxx.xxx.183.53 (xxx.xxx.183.53)  0.527 ms  0.415 ms  0.483ms

All works if I run the following combination:

$sudo route del default

$sleep 10

$sudo route add default xxx.xxx.144.125

Answer 1

If you use route change command, you need to restart network service to apply the changes, e.g.:

$ sudo /etc/rc.d/netif restart

Answer 2

Is it possible you have a cached route to 8.8.8.8? The full output of netstat -nr would show that. If so you would need to remove it before testing your change - either that route specifically or you could route flush and then add rather than change the default gateway (but bare in mind this will interrupt non-local traffic whilst between the two commands if you go that way).

Answer 3

I don't know why I don't recognize the output from your netstat -rn command, but I'm used to seeing a NETMASK there.

Not sure why your netmasks aren't showing up, but since they're not, you might have completely incorrect netmasks and the certainly could be your problem.

Answer 4

From the output of the ping command, it seems that you are targeting the local IP address, rather than the remote gateway.

So if xxx.xxx.183.54/29 is assigned to your FreeBSD node (as seen under ifconfig) you need to use the next hop IP address which looks like it could be xxx.xxx.183.53/29. The same may be true in that the xxx.xxx.144.125 address is the address assigned to a local interface.

As an example, I have a machine with the following:

ifconfig:
  em0: flags=8863<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
  inet xxx.xxx.123.99 netmask 0xffffff00 broadcast xxx.xxx.xxx.255
  
  em1: flags=8863<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
  inet xxx.xxx.234.99 netmask 0xffffff00 broadcast xxx.xxx.xxx.255

netstat -rn
  Destination        Gateway            Flags     Netif Expire
  default            xxx.xxx.123.254    UGS       em0

In my example above , all traffic is currently going out of the em0 interfacec towards the ISPs #1 Router IP (xxx.xxx.123.254) as the next hop.

If I wanted to route all traffic out of em1, rather than em0 I would use: $sudo route change default xxx.xxx.234.1, if xxx.xxx.234.1 was my ISPs #2 router IP address on that LAN segment.

I would also suggest checking to see if dhclient is running with either ISP, as it maybe installing routes unbeknown to you in the background.