Re: [PHP4BETA] Database Connection Pooling

Thus spake Manuel Lemos (mlemos@acm.org): > >> AFAIK all supported databases have functions to establish persistent > >> connections. They work on a per server process basis. That means > >> persistent connections are kept until a server process lives, and > >> each server process manages their own pool of persistent connections. > > >Assumption is the mother of all screw-ups. While PHP does support > >persistent connections, those are per Apache process. What it needs is > >a common connection pool for all Apache processes. > > If that is what it was meant, I don't think it is a good idea. One > persistent connection per process is better than have Apache processes > competing for database connections allocated from a common pool. The way I > see it, that would impose needless latency in database connections. > I don't see why you would think that a connection per process is better. Except for the case where EVERY hit requires a DB connection for handling of the request, you will always have extra, resource consuming connections. (Note that unless you have your MaxRequestsPerChild set very low, you will eventually have a db connection in every Apache child.) IMO, it would be better to have a pool of connections that you could tune much the way the pool of Apache children is tuned. Cary -- Cary Collett cary@ratatosk.org http://ratatosk.org/~cary "To me boxing is like ballet, except that there's no music, no choreography, and the dancers hit eachother." -- Jack Handy