In the last episode (Aug 09), Jules Bean said:
> On Mon, 9 Aug 1999, steve procter wrote:
> > The following output is from running "top" after several hours of
> > uptime/browsing our webpages. As you can see, mysqld is eating a
> > lot of memory (it has hit the limit of 500MB of memory and so is
> > now swapping out).
> You're misunderstanding the output.
> Mysql is taking a mere 40M per process - of which a large amount will
> be shared between all the processes. The 500M used is almost
> entirely OS-level caching - as is shown by the figure 430612K cached.
> So only about 70M is really used by all your applications together.
Another thing to look at is the "mysqladmin extended-status" output; If
key_reads is high, or keeps going up after a few hours of uptime, you
might want to raise your key_buffer value above its current 32MB
setting (which is probably where the 35MB RSS is coming from in the top