Hi all,
I'm new to the list as well as to Postgres. At present we're using MySQL,
on a BSD platform,
and migrating the databases to AIX, keeping BSD on the frontend. The
frontend is powered
by Php and Apache and a few Perl scripts since I just love GD::Graphs3d :-).
I hope I can contribute in the near future, but my guess is, I'll be askin'
a lot first :-). Especially
since it's quite a different setup as opposed to MySQL.
We're considering Postgres at least for some more demanding webapplications
and therefore
I've setup a little test environment at home on a similar box (well -
software/OS then).
As for the first questions:
It's customary for us to run web-accessible services chrooted. However,
postgres needs to
change user context and chroot requires the root user. There are two things
bugging about
the current Postgres implementation, with respect to a chrooted environment:
1) Postgres requires me to have 'su' in the chrooted env
2) Postgres makes a shell call to the 'cp' command when creating new databases.
The first I've overcome, with the not-so-elegant script below, but the
second is not so easy.
So my questions:
-- Does anybody run PostgreSQL chrooted and so, how?
-- Is there a way to use PostgreSQL and not have `cp' in the environment?
Many thanx, in advance.
The startupscript:
$ cat /chroot/pgsql/startpgsql.sh
#!/bin/sh
cp -p /usr/bin/su /chroot/bin/su
/usr/sbin/chroot /chroot /bin/su postgres -c "/pgsql/bin/postmaster -i -D
/pgsql/data &"
sleep 2
rm /chroot/bin/su
--
Best regards,
Melvyn Sopacua
---------------------------(end of broadcast)---------------------------
TIP 4: Don't 'kill -9' the postmaster
