axis-c-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sam Carleton <scarle...@miltonstreet.com>
Subject Re: db connection pooling
Date Tue, 13 Jul 2010 19:18:09 GMT
On Tue, Jul 13, 2010 at 12:30 PM, Samisa Abeysinghe <samisa@wso2.com> wrote:
>
>
> On Tue, Jul 13, 2010 at 6:51 PM, Sam Carleton <scarleton@miltonstreet.com> wrote:
>>
>> Most servers which would host a SOAP engine like Axis2 have some method of implementing
connection pooling (for things like database connections).  How does one get access to connection
pooling in Axis2?
>
> We do not have a connection pool in Axis2/C
>
> However, you can plug any pool you want in your business logic implementation.
>
> By design, Axis2/C do not use any databases, hence we do not really need connections
for our engine.

Samisa,

The way Axis2/C stands today, I cannot plug any pool I want into my
business logic.  The pool I want is also in use by the hosting web
server.  To use this pool, I would need a token (pointer) to be passed
through your engine.

This is along the lines of my post from a week ago or so about the
Apache Portable Runtime. (The APR which happens to contain the
connection pool system I want to plug into).  There seems to be a very
important thing missing in your engine:  A way to server specific info
through to the business logic.  It would only seem logical to me that
on the env variable, there is a ??? variable which is the ??? from the
server.  Maybe the request object/variable.  In the end, the engine
treats it as a void*, just like engines such as eXpac send a void* to
all it's functions which is user defined. This would allow each server
module to use this void* to it's liking, if at all.  It would seem
logical that for both IIS and Apache this should be the request
pointer.

I fully understand and appreciate the fact that Axis2/C is server
agnostic, but I would be willing to bet dollars to donuts the business
logic running under Axis2/C normally is *NOT* server agnostic.  My
guess is that most folks are in my boat where the web services are
only part of the whole application, the whole application also depends
on the server, too.  So I believe many folks would find it of great
value to leverage server specific features in the business logic.

For the record, I am already hacking the mod_axis code to pass through
my own info, so I will be doing that with the connection pool, too.
My main point in asking the question and in the reply is to bring to
light the issue and how it can be of service to others.  Someday it
would be nice if my current hack of passing this info through the
engine was a standard.

Sam

---------------------------------------------------------------------
To unsubscribe, e-mail: c-user-unsubscribe@axis.apache.org
For additional commands, e-mail: c-user-help@axis.apache.org


Mime
View raw message