I have come to like postgresql and I have written many utilities for
internal usage, mostly related to my mailing list manager. The actual
mailinst list program is written in c and opens and closes only a single
connection for each email received. The problem is mostly with the
utilities which are shell scripts using psql between 1 and 5 times per
script. Needless to say, I can expect to waste many resources if/when
traffic becomes more important.
The solution to this problem could be to write a daemon which would leave
a connection open to a postgresql database. Then, when data is needed, I
could pass the query using shared memory and perhaps semaphores to avoid
concurrent access to the daemon therefore avoiding the overhead of
creating a connection each time. I could then use this same method for my
mailing list manager and each utility (which would have to be re-written
in c) all of which would use the same connection provided by the daemon.
Unfortunately, my understanding of the subject is somewhat limited. I can
manage to write the daemon using books and sample source code, but I'm not
in a position to weigh the pros and cons of such a solution. Therefore, I
would like to sollicit your feedback to gain a better understanding of the
problem at hand.
Looking forward to your insight,
Marc