Strange ARP problem
Summary:
I'm setting up a 100baseTX network between 2 Debian'ish computers with a
switch between them. The problem is that neither computer replies to an
ARP request from the other (or maybe neither gets the other's ARP
request). If I hand code the appropriate MAC address into /etc/ethers and
then run "arp -f"on both boxes, the connection works.
The details:
I'm in the process of migrating from a 10base2 network to a 100basetx
network. I need to keep everyone connected, so both networks must be up
during the switch.
One computer (Gateway) is a Debian Woody kernel 2.4.4.
eth0 is static 192.168.1.10/255.255.255.0 serving the existing 10base2 network.
eth1 is connected to the cable modem and is configured using DHCP.
eth2 is static 192.168.100.10/255.255.255.0 serving the new 100basetx
network, broadcast is 192.168.100.255.
both eth0 and eth1 are ISA cards (dlink combo using ne drivers)
the eth2 card is a Netgear FA311 using the driver that comes with 2.4.4
The other computer (Sandbox) is a Progeny 1.0 kernel 2.2.19
eth0 is static 192.168.100.9/255.255.255.0, 192.168.100.10 is the gateway,
broadcast is 192.168.100.255
eth0 is a Netgear FA311 using the natsemi driver.
The switch is a Linksys 8-port Workgroup Switch.
When I run iptraf on Gateway, it shows the ARP requests going out but no
ARP replies or ARP requests coming in.
Could the switch be filtering out the ARP requests?
I'm running ipchains on Gateway. Does ipchains filter ARP requests too?
Any other ideas?
Ed
Reply to: