By my reading, the machine is definitely swapping, and not writing to a log
file (unless its writing obscene amounts of data to the log, which
presumably the default settings won't do).
postmaster -i -D /home/mg/pgsql -B 100
produces almost identical results in terms of performance and disk
activity. top shows that each child only has 2.2MB shared instead of 12MB.
What other program/means do I have to tell if the machine is swapping? Can
I get a reading as to the amount of data that is actually in physical
memory within a process?
-Xavier
At 08:24 PM 4/23/01 -0400, Tom Lane wrote:
>xbdelacour@yahoo.com writes:
> > I'm no Unix expert, but this would seem to indicate that shmget is
> > successfully allocating 400385024/1024/1024=381MB of shared memory. I
> don't
> > know enough about how the postgres parent/child/shmem scheme works to know
> > why this is working yet the children only register 12MB of shared memory
> > under top.
>
>On most of the systems I've worked on, top does not seem to count shmem
>blocks that a process is attached to in the process' memory usage. So
>that doesn't prove much one way or the other.
>
>I am wondering if your version of 'top' fails to count swapped-out shmem
>segments against swap space, or something like that. That'd be a tad
>weird, but it seems very improbable that your machine is not swapping;
>I just do not believe top's claim that no swapping is happening.
>
>Anyway, the most direct experiment would be to reduce your -B request to
>100MB or so and see how things change...
>
> regards, tom lane
_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com