zemon@felix.UUCP (04/09/84)
We are having a lot of trouble with the 4.2 Ethernet stuff. If any of you have had troubles like ours, please contact me. Our situation: We are running 4.2 bsd on two VAXen (a 750 and a 780). The VAXen are connected with an Ethernet. We also have a number of M68000 workstations on the same Ethernet cable. The VAXen can communicate both with Berkeley's networking software and with a proprietary protocol. Both protocols occur on the same wire. The M68000 workstations communicate only with our proprietary protocol. Our problem: The Berkeley networking stuff is not reliable. Connections made with, for instance, rlogin often close spontaneously. MDQS, also running through sockets, often loses connections. The ruptime command, which displays the up/down status of the machines, thinks that one of the VAXen is down. Often, ruptime reports that the machine you are currently logged in on is the one which is down! This strange condition can only be cleared by resetting the interface with /etc/ifconfig on the machine which seems to be down. Our proprietary protocol communications are always reliable, even when the Berkeley stuff stops working. These problems make all the "wonderful" networking capabilities such as rlogin, rdump, etc. virtually useless. Do we, perhaps, have an early and buggy distribution tape? Ours is #24, made 8-23-84. If you are located in southern California and have a tape made significantly later and would be willing to let me borrow it, please let me know. Art Zemon FileNet Corp. Costa Mesa, CA (714)966-2344 ...!{decvax, ucbvax}!trwrb!felix!zemon ...!bmcg!felix!zemon