note 12319 deleted from function.mysql-pconnect by jimw
| From: | jimw@php.net | Date: | Fri, 07 Dec 2001 01:45:43 +0000 |
| Subject: | note 12319 deleted from function.mysql-pconnect by jimw | ||
| References: | 1 | Groups: | php.notes |
| Request: | Send a blank email to php-notes+get-21770@lists.php.net to get a copy of this message | ||
We run a heavily loaded site on our systems (>2,5 million hits a day) and make extensive use of
PHP4 (several versions tested) and MySQL 3.23 (also several versions tested, currently running
3.23.33). I was always using mysql_pconnect() instead of mysql_connect() until someone suggested I
ran some benchmarks because pconnect might actually be slower.
I don't have any numbers to back this up anymore but the results were stunning,
mysql_pconnect() was actually MUCH SLOWER than the normal mysql_connect(). Also the memory
requirements of MySQL were higher when using pconnect. The server the DB runs on has 1GB RAM so
swapping is not an issue here. We have only 3 different databases on our server (totalling about
200MB data) and about 100 to 200 open mysql connections at a time, so theoretically this is an ideal
situation to use persistent connections. Also, all connections use the same user/password
combination.
In the past few months I have been testing this on different sites and different servers and
mysql_connect() is consistently faster than mysql_pconnect(). So I'd recommend anyone to stay
away from this function or at least do some thorough benchmarks of your own site with or without
pconnect because it might help you gain some performance.