eho@clarity.Princeton.EDU (Eric Ho) (09/20/89)
Has anyone encountered situations where some of your D-machines (1108 or 1186) live on same segment as a number of Suns (say they all live on the same floor/building) and then the segment is in turn connected to other buildings (main campus) via a pair of Cisco boxes and that whenever you reboot your Cisco box or whenever you test the external ethernet loop on your Suns, you'll get barfs on ALL the D-machines all at the SAME time with something like "Illegal Address Arg: -1" ? In order to further use other system services like the Clearinghouse Service, I'll have to get out of the break loop first. All my Clearinghouse Services run on two 8000 fileservers and both of them are on the same segment but they're not affected in any way ? This is particularly bad if your 1108/1186 is in idle state -- i.e. you've to know the password of the last user using that D-machine in order to login so that you can get out of the break loop (giving a different login name is no good because it'll just cause the D-machine to go out to the net for authenication and it can't go out unless it is out of the break loop). I think that something is seriously wedged in the Lisp microcode or even in Pilot that handles the ethernet. I don't know what is the situation in Lyric since all the D-machines here are running Koto. Any pointers will be much appreciated. -- Eric Ho Princeton University eho@confidence.princeton.edu