On 21 Jun 2001 15:16:19 +0800, Rajeev Rumale wrote:
> Hi,
>
> I need to know if there is any easiest way to keep session data or object
> accross the scripts.
>
> Basically I would like to pool Database connections so that Parrallel
> running scripts don't open multiple connection with the database.
>
> with regards
>
<snip />
The only way I can think of to achieve this would be to write a daemon
process in perl (or any other language for that matter) that would be
responsible for accessing the database based on requests (through IPC,
BSD style sockets, files being placed in certain directories, smoke
signals, whatever) and returning the data (again through some
communication method). I have seen production systems (not that I
recommend this) that had a special set of directories named /work/in and
/work/out. Shell scripts would print sql statements to files in the
/work/in dir and a C daemon would: pick them up, see if they were from
the right owner, discard the invalid files, run the valid ones and put
the results in the /work/out dir. Filenames were based on the pid of
the shell script. The shell script would then sit waiting for a file
with its pid to show up in the /work/out directory.
In general, if you are having to create hacks like this the problem is
most likely you choice of RDBMS. Enteprise level databases generally
don't have a problem with thousands of concurrent connections.
--
Today is Boomtime, the 26th day of Confusion in the YOLD 3167
Or is it?
