Hi I am still tracking down issues on our cable network here. (See previous old posts - its getting better!)
We seem to be getting an issue with rogue ARP data, for example LAN addresses getting replies from the WAN side, as logged, for example: kernel: arp: unknown hardware address format (0x0000) kernel: arp: unknown hardware address format (0xdd1f) kernel: arp: 192.168.0.7 is on em1 but got reply from 00:00:cd:1c:14:1a on em0 kernel: arp: 192.168.0.7 is on em1 but got reply from 00:09:bf:55:71:b0 on em0 I cannot identify these ARPs on the LAN so they seem to be WAN side. It may be possible that some sort of erroneous ARP traffic is the problem here that causes the Motorola Cable modem to occasionally reboot and also to occasionally lose the connection with pfSense, which has been re-established with ifconfig code as has been detailed on the list before. With reference to http://www.geekzone.co.nz/forums.asp?ForumId=44&TopicId=19840 what sort of ARP caching does pfSense do? This is on 1.2 RC2 embedded, have there been any changes since that might be relevant to this issue here since? (I have not upgraded as an upgrade I expect needs to be customised with the ifconfig rescue code we did!) Can anyone make anything of what I have described here? Kind regards David Hingston