Re: postgresql + apache under heavy load

On Wednesday 21 January 2004 22:20, Alex Madon wrote:
> Well the top issued was corresponding to a concurrency of 20 (so a level
> where swap is not sollicited).
> With a concurrency of 60, swap is very high
>
> Here is a "movie" of how swap is maanged (a snapshot every 5 sec)
[used sits around 110MB, spikes to 240MB then gradually drops back to 153MB
just under a minute later]

> and the parallel with Mem
[buffer memory plumets from 20MB down to 6MB, then stabilises around 6.5MB]

OK - I'm guessing the two spikes in the figures are as the ab requests hit,
spawning http daemons and postgres backends.

> One can see that at the maximum feeling of swap (74700k free swap), the
> full picture is:
>
>
> 22:51:54 up 3:58, 6 users, load average: 47.38, 18.53, 7.79
> 131 processes: 130 sleeping, 1 running, 0 zombie, 0 stopped
> CPU states: 5.3% user 3.0% system 0.0% nice 0.0% iowait 91.6% idle

Although Tom pointed to the high CPU in a previous post, here you're not
suffering on CPU. Run a "vmstat 5" while this is happening - that'll show us
the details.

--
Richard Huxton
Archonet Ltd