steve@dartvax.UUCP (Steve Campbell) (10/01/84)
We recently brought up our Ethernet connecting our 780 and 750 (both running 4.2bsd). Both VAXen have DEUNAs; we're using Lou Salkind's driver. It has all worked without a snag except for... Yesterday I found that one machine - host dartvax - was refusing rlogin requests. Inspection showed the following evidence: 1. The rlogind daemon on the dartvax was gone (hence the refusals), and could not be restarted. Ie when root said /etc/rlogind, no daemon was spawned though no error messages were produced. 2. A netstat command showed the following: Active connections Proto Recv-Q Send-Q Local Address Foreign Address (state) tcp 0 0 dartvax.login dartvax.1023 FIN_WAIT_2 tcp 0 1 dartvax.1023 dartvax.login LAST_ACK 3. A netstat -i command showed the following: Name Mtu Network Address Ipkts Ierrs Opkts Oerrs Collis de0 1500 K_Ether dartvax 4392 0 2524 0 0 lo0 1536 127.0.0 localhost 81635 0 81635 0 0 ...and the unusually high activity on the local loopback interface continues at a rate of about 1 packet per second. I have found nothing short of rebooting that clears this situation. (It has happened twice in a month of Ethernet operation, once on each machine.) My questions are: (1) what caused this situation and how can it be prevented and/or cured? [See clue below.] (2) how could I have solved this probelm myself? Ie, where is the documentation (seriously, folks)? Clue: For reasons of local historical compatibility, we have a "bye" program that disconnects users from the hosts. For this to disconnect an rlogin'd user from his/her remote host - what most people would do either with a ^D or a "logout" - the bye program does a kill (getppid(), SIGHUP); It works; it seems to cleanly disconnect the rlogin connection. Could it also cause the lockup? Stephen Campbell | Dartmouth College | dartvax!steve (603) 646-2643 | Hanover, NH 03755 | steve@Dartmouth.csnet