Inexplicable 2.9BSD crashes and curiosities

[jeffrey w percival]

I sure hope someone can help me.  We have just recently started
running 2.9BSD on our PDP 11/70 (RM03, Fujitsu Eagle, 2 TE-16 drives,
1 DH-11, 1 Emulex DH, Decwriter).  Our system is crashing at least
twice a day, and we don't understand why.  Some random thoughts:
1. Is there any problem with running executables that were compiled
   under 2.8?
2. What about executables that came off the distribution tape?  I ask
   this because I've noticed some bothersome peculiarities about
   source vs. object code in the distribution...
   A. The released C compiler was not passing the "-i" flag to the loader,
	  and we were unable to produce separate I/D modules.  Not even the
	  C source was coded to do this!  It needed to be hacked a bit.
   B. In trying to remake the f77 compiler, yacc bombed out on one of
	  its operations.  I re-compiled yacc, then "make f77" again, and
	  yacc performed OK!  Nothing was changed, only re-compiled.  That
	  wasn't enough, though, because the f77 make failed later on with
	  a message from the loader about "no text" or something like that.
	  Can anybody explain the many discrepancies between distributed
	  source and distributes executables? Has anybody re-made f77 on a '70?
	  How?
   Is there some generic problem with running executables that we have
   not re-compiled ourselves?  What machine was 2.9 developed on?
3. We named the root file system as the pipe device, and we only have
   a few hundred free blocks in that system.  Could that be the problem?
4. The crashes seem to happen when the CPU gets busy, e.g. when we
   are re-making kernels.  There is usually much file system damage.

I would appreciate any comments...  Sorry if this note is a little
disorganized, but I guess it reflects our mood these days.

-- 
	Jeff Percival ...!uwvax!uwmacc!jwp