*BSD News Article 5434


Return to BSD News archive

Xref: sserve comp.unix.pc-clone.32bit:9 comp.unix.bsd:5482
Path: sserve!manuel!munnari.oz.au!uunet!spool.mu.edu!agate!soda.berkeley.edu!wjolitz
From: wjolitz@soda.berkeley.edu (William F. Jolitz)
Newsgroups: comp.unix.pc-clone.32bit,comp.unix.bsd
Subject: Re: Attempted summary (Was Re: UNIX on a PC clone)
Date: 22 Sep 1992 23:16:01 GMT
Organization: U.C. Berkeley, CS Undergraduate Association
Lines: 25
Message-ID: <19o9fhINNplu@agate.berkeley.edu>
References: <19ntp1INN2qo@almaak.usc.edu> <1992Sep22.220331.21122@terminator.cc.umich.edu>
NNTP-Posting-Host: soda.berkeley.edu

In article <1992Sep22.220331.21122@terminator.cc.umich.edu> pauls@umich.edu writes:
>
>Furthermore the system has crashed so far only for understandable reasons.   
>That is, only while compiling, and then as far as I can tell because the  
>installation's default swap quantity is too low.  That's an important  
>thing to note for beginning installers -- DON'T INSTALL IT UNTIL YOU HAVE  
>FIGURED OUT HOW TO MAKE SWAP BIGGER OR YOU WILL MAKE LOTS MORE WORK FOR  
>YOURSELF LATER, AND/OR JUST END UP REINSTALLING.  Just something to keep  
>in mind.

Just a note -- we've been running the standard system (5MB swap space)
recompiling the system, X windows, running debuggers, the BSD command set,
and so forth and have not had to raise the swap space at all.

We've done this on 386 and 486 systems with all manner of controller and
drive types and all amounts of memory (2MB, 3MB, 8MB, 16MB, 20MB).

While upping swap space may appear to solve the problem, it is questionable
whether it is actually dealing with the REAL problem, or is just masking
it in some way.

I suggest that we start looking more closely at these situations, to 
deterministically localize this problem.

Lynne.