Originally Posted by pkraus109
The subnet is 255.255.0.0 and the one gateway is 10.0.0.1
Hmmm, it shouldn't be an issue then. Is it just two pieces of equipment that are talking to each other slowly? Or every
piece of equipment on that subnet?
The way it should be working (with some made up IP addresses):
10.0.1.5 wants to talk to 10.0.4.4
10.0.1.5 looks at its own subnet mask and sees that it is 255.255.0.0
10.0.4.4 fits into its subnet of 10.0.x.x
So, it sends out an arp who-is request ("who-is 10.0.4.4") broadcast
10.0.4.4 hears the who-is broadcast and responds to 10.0.1.5 ("arp response: 10.0.4.4 is at 00:11:22:33:44:55 mac address")
10.0.1.5 puts the correct mac address as the destination as it creates the packet and sends it out.
The level 2 switches send it to the destination machine (they don't care less about IP addresses)
10.0.1.5 remembers the MAC <-> IP address association so it doesn't have to ask for future packets.
I was thinking that maybe one of your subnet masks was set wrong so that it had to send every packet to the gateway. But if you've got everything set up right, I don't see why there would be a problem.
Does it seem to only be every segment having the problem?
Another thought: are your switches "smart" level 3 switches that are looking at the IP addresses of the packets and have default gateways programmed into them as well?