On Friday 04 Jun 2004 11:06 pm, Brian Meadows wrote: > On Wed, 2 Jun 2004 23:58:06 +0100, Derek wrote: > >On Wednesday 02 Jun 2004 22:37, brian wrote: > >snip > > > >> So most of those I understand, but anyone know what the numeric > >> entries represent? > >> > >> Second question - 9.1 ran just fine on this PC (600 MHz PIII, 512 > >> MB of memory) but once I'd installed 10.0 I noticed a lot of disk > >> thrashing going on. I ran up KDE system guard (I"m using the > >> version of KDE which came with 10.0, and that's the only desktop > >> I've installed) to find that I'd only got a couple of megs of > >> memory free, which explains the thrashing, > > > >Linux uses all unused memory as a disc cache. It is perfectly normal > > for memory usage to be 100% After all unused memory is 'wasted' > > memory. > > Hmm. And a hard disk which is being *constantly* accessed is a > hard disk that is likely to have a short lifespan - assuming > you're not running server-class drives, which I'm not on my Linux > box. I wouldn't have noticed the memory usage had it not been for > the disk thrashing. > ... > Thanks for the info, but I'm still not convinced. If this > constant disk access really is normal for a Linux system, I'm > going to buy shares in some hard drive manufacturers!
You're right in that it shouldn't be constantly thrashing the disk, but Derek is correct that most of memory is always used for disk caching, so it isn't obvious. Try: $ top and give it the commands: fuOu Check the nFLT column (page fault count.) With an uptime of 1 3/4 days I have X at 19k, and several desktop (xfce) processes at 1-3k. Anything increasing constantly might be a problem. I'm using 9.2 with 320M of memory. -- Richard Urwin
____________________________________________________ Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Join the Club : http://www.mandrakeclub.com ____________________________________________________
