Hi-
It would be great if somebody can give me a "recipe" answer for this, but
I'm equally interested in informed conjecture and stories from similar
experiences.
We're getting ready to spec our hardware for the imminent release of an
application which uses JDBC to funnel data from our database to the web via
Tomcat.
We expect to set up a database server running PostgreSQL, an app server
running Tomcat, and a web server running Apache.
In our development environment we've observed that whenever we open a
connection to the database from the app server, a process owned by postgres
gets started on the database server. It appears that the process is the same
whether our remote client is a Tomcat servlet or psql running from the
command line.
Given this background, I have a specific question & a couple of general
questions:
How much memory should we be expecting to use for each of these processes
that get started on the server? What factors determine the load that each
one puts on system resources & what information should I gather to estimate
the impact on resources?
More generally, can someone point to an explanation of the division of tasks
between postmaster, per-client backend processes running on the database
server, and client processes running on the application side?
And finally, if anyone else is running PostgreSQL->JDBC->Tomcat->Apache, I'm
curious about your hardware configuration & experiences with it. Any
mistakes that you can help me avoid?
I've looked at Tony Grant's article on his version of this setup, & I'm
trying to digest Bruce Momjians's big "PostgreSQL Internals" article, but so
far, I haven't found these particular questions addressed.
Thanks everyone!
-Nick
--------------------------------------------------------------------------
Nick Fankhauser nickf@ontko.com Phone 1.765.935.4283 Fax 1.765.962.9788
Ray Ontko & Co. Software Consulting Services http://www.ontko.com/