>From: Stroller <MacMonster at myrealbox.com> > >On Apr 9, 2005, at 7:43 pm, SeaSoft Systems wrote: > >>> From: SeaSoft Systems <seasoft at west.net> >>> >>> I have really no clue here. Could anyone with some network expertise >>> help me understand this situation or point me to a resource or more >>> appropriate forum? I have Googled my brains out but have not found >>> anything helpful. >> snip >> So very sorry... >> >> I should have Googled *after* I found the system log entries. There is >> in fact quite a lot to be found under "mach_kernel: arp: " > >Since you've piqued my interest (and probably that of others, too) do >you think you could let us know what has caused this, when you've >finally got to the bottom of it? > Most certainly; it's a bit long and tortured, but for the terminally curious... Turns out a year ago I set up a Brother USB printer that also had a network interface. During the printer network setup, I assigned it an unused IP on my LAN (192.168.167.21). After testing, I never used the IP number for printing (I only set up the IP print feature to access the printer via an NT box also on the network). Since I never used the IP print interface, I forgot I had assigned the IP number to the Brother. (DHCP, anyone? :) I inadvertently set the *same* IP for a web serving iMac when I got my Linksys firewall/router a few months back. The conflict never created a problem, even when the printer was on. Because I was only using the USB connection to the printer, no IP-addressed printer traffic intended for the printer ever hit the network. Why did the duplicate IP not cause other problems? I'm more than a little fuzzy on this; I am just not sure what the router was doing since whenever the Brother was powered on, there were two (192.168.167.21) IPs online. I have a couple guesses: 1. The router had the errant IP assigned to the iMac server MAC from its initial configuration and simply didn't send any of the server traffic to the printer. The printer, never seeing any IP traffic for itself, never broadcast its IP? 2. Perhaps both the Brother and iMac got the WAN-traffic? And the Brother, not seeing recognizable printing commands, simply ignored them silently? Sorry I can't be more helpful here; I'm just way out of my league, technically... The day of the "incident", however, I actually tried to print using the IP print setup on the OSX machine. It didn't work (timed out or something), and I just printed via USB as usual. It wasn't until days later that I noticed the error messages (quenched automatically by OkeyDokey on the iMac server) that triggered my investigation. So, the OSX box evidently "re-assigned" the duplicated IP from the iMac server MAC to the printer MAC in an attempt to satisfy my IP-based print commands. I have no idea how this affected the router's handling of WAN I/O to the iMac server. It seems that the router kept sending the WAN/LAN packets to the right machine (iMac server) and that the only real conflict was on the iMac server which then saw the duplicate IP conflict and protested. Note that whenever I have inadvertently replicated LAN IPs on my own, this has always caused the affected OS8.6 machine to crash; that didn't happen this time. It was a *very* educational process. One thing I learned that has altered my OSX behavior and may be of use to other neophytes: Keep a dedicated terminal window open that displays real-time additions to the system.log. This is really, really informative about what's happening on your machine. If you lack a hardware firewall, for example, it will display all the cracking attempts from Russia/China/India/etc. Combined with a WHOIS-type IP lookup, this is extremely fascinating. The command I used was $ tail -f /var/log/system.log The -f flag keeps tail adding log records to the bottom of the window as they show up. Hardware firewalls, of course, make this game much less educational :) This "Tail Window" experiment made me take the OSX box out of the DMZ of my hardware firewall/router; just too much stuff going on from the Internet. Its also quite useful in following & understanding all the "OSX vulnerability" brouhaha currently in progress. Whew, Richard