[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Request for help - my computer 'bought the farm'



On Sun, May 30, 1999 at 02:36:08PM -0400, Roland McGrath was heard to say:
> >   I've been trying to compile Guile on the Hurd and after a few changes (see
> > my earlier posting about poking inside FILE *s) it compiles.  However, after
> > touching build-stamp the computer dies with messages about: [unfortunately this
> > is only an approximate transcription, I don't have any way to save the actual
> > message]
> > 
> > (default-pager): refusing paging request because of previous errors
> > ... [repeated lots of times] ...
> > *** [Error 2] Computer bought the farm
> > make: [Error 2] Computer bought the farm
> 
> Whenever you see any "(default pager): " messages, stop and investigate
> right away.  That error "... because of previous paging errors" should come
> only after some previous errors from the default pager.  (Most likely from
> running out of swap).

  Unfortunately I wasn't present when this happened, my usual routine with
compilation is to start it and then go do something else :-)

> >   It looks like this has something to do with the pager but I can't find the
> > previous errors it refers to (since the Mach console doesn't have
> > backscrolling..maybe I could try piping the compile through less..)  
> 
> Unfortunately, those messages from the default pager go directly to the
> console device, not through the terminal driver or anything else you can
> catch. 

  Ouch.  Yeah, I noticed that when I tried to catch the output of make (a lot
of things seem to avoid stdout and stderr, even normal compilation output..)

> > It may help to mention that my computer doesn't have swap space under the
> > Hurd due to the fact that I can't get it to recognize my Linux swap
> > partition.
> 
> Indeed, having no swap is no doubt the problem.  Please report the details
> of your problem with swap.  I believe the current Hurd handles Linux 2.0
> and 2.2 swap signature pages correctly, though I have myself only tested a
> 2.0 page.

  The problems were that it didn't work right off when I boostrapped the Hurd
last week and I decided not to worry about it immediately unless it caused
problems.  It turned out that I had to edit one line in /boot/servers.boot :)
  Unfortunately that didn't make the problems go away.. (see my other message)

  I'll do a reboot to try again and see if I can collect more data for you
(eg, it looks to me now like the pager is coming up correctly but I could be
 wrong)

  Daniel

-- 
All generalizations are dangerous.


Reply to: