Re: Success !!

jon,

i'm sorry, but i can't tell you anything about perl. i only did some
tests with the postgresql jdbc interface.

> The regression tests hit just one failure, which was for
> numerology. Should I be worried about this ?
in most cases no, but simply run src/test/regress/checkresults.
this gives you a diff of expected vs. your regression test results for
the failed tests. src/test/regress/README explains some reasons for
differing results.

> In the /contrib/linux area there is a script
> called postgres.init, which looks promising, however, the comments
> say this again is a RedHat script (so after my previous experience
> am slightly nervous to set this one up). Also in that script it will
> bomb out if networking is not running. I'm currently building my
> server standalone. Also, it uses something called local 5 (don't
> know what thats about).
for local[0-7] see man 3 syslog
the syslog local5 stuff works w/o problems. but i'm not shure about the
network thing. i think you should have a loopback device configured at
least. (you will need it for apache, too. right?)

> So, if you have something that you use
> on your Linux machine to autoload/autoclose the postmaster
> at boot time, that would be just groovy.
i use a modified script from an older postgres contrib dir (or the
faq?). here it is:
--- /sbin/init.d/postgresql.init -----------------------------
#!
/bin/sh
#
# postgres.init Start postgres back end
system.
#
# Author:       Thomas Lockhart
<Thomas.Lockhart@jpl.nasa.gov>
#               based on news startup by David
Myers
#
# Modified by Daniele
Arduini:
#       - support for bash
shell
#       - /usr/local/pgsql
installation
#
#
Assumptions:
# - the postgres user is named
"postgres"
# - the postgres user is running
bash
# - /etc/syslog.conf contains the line "local5.*
-/var/log/postgres"

[ -f /usr/local/pgsql/bin/postmaster ] || exit
0

# See how we were
called.
case "$1"
in

start)
    echo -n "Starting postgresql service:
"
    # force full login to get path names and environment
variables
    # postgres runs bash so use proper syntax in
redirection
    # change this line if the postgres superuser account is not
"postgres"
    # change this line if another shell syntax is
necessary
    su postgres -c 'nohup /usr/local/pgsql/bin/postmaster -i -B 256
\
       -D/usr/local/pgsql/data 2>&1 | logger -p local5.notice
&'
    sleep
2
    pid=`pidof
postmaster`
    echo -n "postmaster
[$pid]"
    touch
/var/lock/subsys/postmaster

echo

;;

stop)
    echo -n "Stopping postgresql service:
"
    killproc
/usr/local/pgsql/bin/postmaster
    sleep
2
    rm -f
/var/lock/subsys/postmaster

echo

;;

*)
    echo "Usage: postgres.init
{start|stop}"
    exit
1
esac

exit
0
-----------------------------------------------------------------------

to run this script with the appropriate parameter, you have to create
two links under rcX.d. example:
  cd /sbin/init.d/rc2.d
  ln -s../postgresql.init K05postgresql.init
  ln -s ../postgresql.init S86postgresql.init

this starts postgresql in runlevel 2. see /sbin/init.d/README,
man 1 init and the suse manual. you can use the script /sbin/rctab to
create and manage these links for the different runlevels.


> [snip]
> I'm building this web server using SuSE Linux 5.1, Perl,
> Apache 1.2.4, Postgresql 6.3.1. Is there anything I should beware
> of, any good tips you have for setting up a web server on that
> platform. Are there benefits to be had by moving to 6.3.2 or 6.4 ?

sorry, can't help you with this. from watching the postgresql lists, i
would say that you always should use the latest release (and apply
patches). this makes even more sense, as you don't have to deal with a
production environment so far.

regards
oliver