Path: utzoo!utgpu!water!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!bloom-beacon!bu-cs!purdue!decwrl!ucbvax!pasteur!ames!amdahl!pyramid!octopus!sjsumcs!hom From: hom@sjsumcs.UUCP (Hom Bahmanyar) Newsgroups: comp.protocols.tcp-ip Subject: A host sending an ARP request to itself? Keywords: ARP ethernet Message-ID: <117@sjsumcs.UUCP> Date: 7 Jul 88 21:34:34 GMT Organization: San Jose State Univ., Math & Computer Science Dept., San Jose, CA Lines: 32 We have an ethernet in our Math & Computer Science Department that has a Sequent Balance, a Convergent Technologies Miniframe, and more than 50 PC compatibles on it. The Sequent and the PC's work perfectly on the network (for example with PC-NFS); however, the Convergent machine has some problems with the network: Commands such as telnet and ftp to or from the Convergent machine always fail but other commands such as rwho and ruptime from the Sequent show the Convergent host as being on the network. Also the Convergent host never responds to pings from the Sequent or the PC's. After we checked all the network setup files and couldn't find anything wrong, I wrote a simple ethernet monitor program for one of the PC's that receives all the packets directly from the PC's 3c501 ethernet controller and displays them on the screen. To my surprise, I saw that the Convergent host was sending an ARP request about every 5 seconds to itself to resolve its own IP address. That is, in the ARP packet there were: sender's physical addr: Convergent host's ethernet addr sender's IP addr: Convergent host's IP addr target's physical addr: 0 0 0 0 0 0 target's IP addr: Convergent host's IP addr But why would the Convergent host want to resolve its own address when it already knows its ethernet address? Any help you can give me on this will be greatly appreciated. ---- Hom Bahmanyar (408) 924-5107 San Jose State University, Math & Computer Science Department San Jose, CA 95192-0103 ...!{sun,ames,pyramid,hplabs}!sjsumcs!hom