Performance with fork and Perl's DBI

[Matthew Fong]

I have an application that forks to handle requests.

I've read through several web documents claiming that I cannot share the same connection between the forked processes and that the correct way to handle this is to create a New DBI class within each child fork.

Each of my fork processes only do a single, very simple query. It seems highly ineffiecient that I perl must close and open a new mysql DB connection for each fork. My program performs at least 50+ forks a second.

Does DBI have some internal way of performacing turning and using some sort of persistent connection sharing? Are there any work arounds, and what are my options. Thanks.


Matt