[no subject]

CC: (BUG LISPM) at MIT-MC
    Date: Sun ,13 Jan 80 00:16:00 EDT
    From: Agre at MIT-AI (Philip E. Agre)
    To: (BUG LISPM) at MIT-AI
    
    In system 25.0, with microcode 612, on LISP Machine One:
    I am trying to load in a very large file, and at some random time into the
    reading of the file, I get CHAOS net lossage of the following flavor:
    
    Attempt to get a packet from #<CHAOS Connection 7003750>, which is in
    CHAOS:HOST-DOWN-STATE, not a valid state.  While in CHAOS:GET-NEXT-PKT <- ...
    
    Neither AI nor my Lisp Machine was down, so I can't figure out the sense
    of this.  It's happened 5 times in a row with different amounts of the
    file read in.

You probably have something in your file that computes for more than
90 seconds without allowing interrupts.  Thus things time out and each
end thinks the other end is down.  Try turning on sequence breaks
(si:sb-on ':clock)