Hi,
I've been looking into postgres and was wondering if I could pose a few
questions.
1) Has any use postgres within a web scenario where you might have 50+
httpd processes, each requiring a db connection? Using postgres, wouldn't
that 50 corresponding backend processes have to be started? Furthermore,
if the web processes weren't using persistent db connection, wouldn't the
postmaster get worked?
2) Has anyone used postgres in conjunction with a ramdisk and how did you
go about maintain synchronicity between the db on disk and the copy in
memory? Did it make a difference?
3) I was under the impression that RDBMS transactions all include a write
to a log file in addition to the db insert/update/etc which allows you to
rebuild the db should anything happen. If that were the case, then what
purpose does fsync mode serve?
4) I've read in a lot of places that the postgres' speed can be reduced,
even dramatically reduce, by using query optimization, stored procedures
and what not. What is the learning curve associated with that?
tia.
r.