Re: [OT] Persistant MySQL connections in Apache 2

[Stas Bekman]

Issac Goldstand wrote:

This is really a C API thing, but I was wondering if Apache2::DBI people can
help me shed some light on this...

I'm not sure what's the Apache2::DBI question, but we (Tim, Hugo, Eric at el.) have discussed the Apache::DBI for threaded apps at the last TPC and the conclusion was that we need to write a generic Perl Thread::DBI module to work with threaded Perl and then use Apache::DBI on top of it.
Some of this was discussed on the dbi-dev list, but as far as I know nothing was developed so far. If you want to discuss the design, which is not trivial, and implementation, it's probably the best to do that on the dbi-dev list.

----- Original Message -----
From: "Kent Fitch"
To: <[EMAIL PROTECTED]>
Sent: Monday, November 18, 2002 5:37 AM
Subject: Re: [apache-modules] Persistant MySQL connections in Apache 2

Hi Issac,

From: "Issac Goldstand"
Subject: [apache-modules] Persistant MySQL connections in Apache 2

Hi all,
 I'm a rookie C API programmer (mod_perl was just so much *easier*

most of

the time :-)), and am working on coding a module to with in my

front-end

Apache 2/mod_proxy server.  The idea of the module is to provide

access to

the front-end server with a list of "blacklisted" IPs which are stored

in a

MySQL database.  So I'm thinking the module's child_init should open

the

connection to the MySQL database and then the access_handler can grab

the

MySQL connection from the child-pool.

Yes, but beware that in a threaded Apache run-time, there could
be many threads fo execution processing many requests "simultaneously"
in your child process, so you'll have to arrange to lock the
connection whilst a thread is using it and possibly have a pool of them.

See "Thread Safety"
http://httpd.apache.org/docs-2.0/developer/thread_safety.html

But the problem is that I would *like* for the connections to the

database

be closed before the pool is simply swept clean upon destruction.  I

would

think that that would be a "good thing to do" - maybe I'm wrong...

Yes - you should register a cleanup function on the
pool you allocate the data structures for the connection(s)
from in your "init". See these notes on pooling (which may be
a bit 1.3-ish and hence out of date, but the ideas are relevant):

http://httpd.apache.org/docs-2.0/developer/API.html#pools

and see the apr_pool_cleanup_register api

Kent Fitch

--


_____________________________________________________________________
Stas Bekman             JAm_pH      --   Just Another mod_perl Hacker
http://stason.org/      mod_perl Guide   http://perl.apache.org/guide
mailto:[EMAIL PROTECTED]  http://ticketmaster.com http://apacheweek.com
http://singlesheaven.com http://perl.apache.org http://perlmonth.com/