Just thought I'd share the results so far:
As I mentioned earlier, I went ahead and upgraded the kernel from 2.2.16
to 2.4.4. The server has been running for nearly 36 hours since then, and
has reached an RSS of 100M, well over what it ever reached before. And
things are still running very smo
Hi,
>From a system point of view, there are 3 main reasons for a slowdown:
- CPU : Your machine is slow because the CPUs are at 100% and can't
do anything more.
- I/O : Your processes are waiting for data from the disk.
- RAM : You don't have enough RAM so your machine is swapping and al
On Mon, 7 May 2001, Rene Tegel wrote:
> On Mon, 7 May 2001 07:40:26 -0400 (EDT)
> <[EMAIL PROTECTED]> wrote:
>
> >
> > Thanks for the quick reply.
> >
> > I set max_connections so low because even at peaks I never need more, and
> > when the slowdowns happen, it seems to have an easier time r
On Mon, 7 May 2001 07:40:26 -0400 (EDT)
<[EMAIL PROTECTED]> wrote:
>
> Thanks for the quick reply.
>
> I set max_connections so low because even at peaks I never need more, and
> when the slowdowns happen, it seems to have an easier time recovering if
> there are 15 slow connections hanging tha
That's the weird part; it doesn't seem to be swapping at all. When trying
different combinations in my.cnf, I had key_buffer as low as 64M without
any effect.
Jon
On Mon, 7 May 2001, Joseph Bueno wrote:
> Hi,
>
> Are you sure that you need 200Mb of key_buffer cache ?
> Since your machine is o
Jon Valvatne wrote:
>
> Hello,
>
> I have a hopefully simple question here:
>
> My web site is dynamically serving 300k page views a day from a MySQL
> database, running on a dual P3/700 with 512 megs of ram. Considering the
> complexity of my queries and the data amount involved, I know I shou
Thanks for the quick reply.
I set max_connections so low because even at peaks I never need more, and
when the slowdowns happen, it seems to have an easier time recovering if
there are 15 slow connections hanging than if there are hundreds.
I'll try the thread concurrency thing, thanks. I was u
On Mon, 7 May 2001 06:27:08 -0400 (EDT)
Jon Valvatne <[EMAIL PROTECTED]> wrote:
> As you'd expect, the server gets bogged down rather quickly at this
> point, serving new requests very slowly if at all. Restarting MySQL helps
> right away, buying me another 24 hours of stable uptime.
If this i
Hello,
I have a hopefully simple question here:
My web site is dynamically serving 300k page views a day from a MySQL
database, running on a dual P3/700 with 512 megs of ram. Considering the
complexity of my queries and the data amount involved, I know I should
expect the server to be overwor