Sun 4/330 going 'dead'

[john powell]

We have two Sun4/330's.  One runs the generic 4.0.3 kernal, the other is
configured for dial in/out modem support.  Generally we are running X11R4
on both machines.  Occassionally when we are running a commercial package
from Intelligenitics for DNA sequence assembly, the system dies.  It goes
completely dead ... even 'L1-A' doesn't have an effect.  We have waited
for over a half hour to see if it was coming back (it didn't).  Telnet &
rlogin to the 'dead' machine do nothing.  We only have about 20M of disk
available so a even a core dump should not take this long.  This has only
occured on a machine running the IG program.  It seems to occur more
frequently on the machine with the modems when someone is also running
c-kermit. 

My question is has anyone else ever seen these symptoms?  What are the
conditions where a machine doesn't respond to L1-A?  How do you trace this
when the only recoverary seems to be a power down restart?

                                jip at alw.nih.gov
                                John Powell