Subject: Re: Possible swapping bug (design 'feature?')
To: None <>
From: der Mouse <mouse@Collatz.McRCIM.McGill.EDU>
List: netbsd-bugs
Date: 07/06/1994 07:53:39
> % uname -a
> NetBSD starwars 0.9a MUFFIN#9 i386

> I'm attempting to run a program called a MUSH [...]

> [children forked for database saves continue to hold VM after
> exiting, according to swapinfo -k]

> Since this problem hasn't occurred under Mach, linux, or SunOS,

I run a MUCK (for those who don't know mucks, it's the same as a mush
for the purposes of this message).  I run it on a NeXT.  I saw behavior
very similar to what you describe - every time it forks to dump the
database, the kernel loses VM.  I eventually had to set dumps to be
done inline, in the main process, to make it workable.

I mention this because you mentioned Mach, and the NeXT runs a Mach
variant.  I too had no such problems with SunOS; I have not tried the
thing on NetBSD.

					der Mouse