Path: utzoo!attcan!uunet!snorkelwacker!usc!zaphod.mps.ohio-state.edu!caen!hellgate.utah.edu!hamblin.math.byu.edu!news From: smithw@hamblin.math.byu.edu (William V. Smith) Newsgroups: comp.sys.next Subject: Re: A few minor prob left... Message-ID: Date: 18 Oct 90 14:24:53 GMT References: <9102@milton.u.washington.edu> <1990Oct15.050058.27579@ccng.waterloo.edu> <53614@brunix.UUCP> <1990Oct17.231958.10800@ccng.waterloo.edu> Sender: news@hamblin.math.byu.edu (Usenet News) Organization: Brigham Young University Lines: 20 In-Reply-To: tgoldtho@ccng.waterloo.edu's message of 17 Oct 90 23:19:58 GMT Thomas A. Goldthorpe < tgoldtho@ccng.waterloo.edu> writes: >that dynamic characteristic. If it runs out of space, swapon allows for other >swap files in other partitions to be used anyway. On a single hard disk >system, of course the only swapfile will be the one created. On a multiple >hard disk system, there are advantages to having the primary swapfile on >another disk along with /tmp. There is a bug (in 1.0 at least) which can cause problems if you have multiple swapfiles. When one fills, the vm system doesn't seem to always know that it should not write part of a page to one and part to another. It then seems to lose track of things and . . . PANIC> This happened to me a couple of times before I figured out what to do about build.660 putting in references to two partitions. Then I rebuilt my disk and now no problems with vm. This is an annoying bit though. It would be nice to be able to make just a few changes and then add another disk which you could swap to (in addition to the boot device say) or mount some remote disk partition down the hall as an overflow if necessary, etc. Anyone know if this is fixed in 2.0? -Bill-