Re: 2.1.92(16MB) swap death

Michael L. Galbraith (mikeg@weiden.de)
Thu, 2 Apr 1998 20:56:00 +0200 (MET DST)


On Thu, 2 Apr 1998, Alan Cox wrote:

> > 2.1.92 is about as bad as pre-2.1.92-2 on my 16MB system. I
> > commented out the "raising capabilites" message in fs/exec.c and
> > tried:
> > (1) compiling init/main.c -- no problem
> > (2) starting X and compiling from an xterm -- no problem
> > (3) starting top in another xterm and compiling --
> > cached went down from 5548K to about 1024K
> > free stayed about 1024K
> > after a minute, screen froze
> > 20 minutes later I gave up and hit reset
> >
> > Is there any point in my going through this exercise?
>
> This roughly matches my 8Mb box. On 32Mb 2.1.92 feels a bit better but has
> the odd disk frenzy, on the 8Mb box its as bad as 2.1.90/91 etc. I tried
> booting it on my 4Mb Zenith 386SX20 for a joke.
>

80 mb doesn't help much.

On my machine, the buffercache is running at 50% of normal (write) bw.
Read is still hovering at 40mb/sec.. write is down from typical 32 to
15 mb/sec. (my hardware is old)

Find / on a 6 gig fs leaves 35mb ram untouched.

-Mike

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu