[freetds] TDsPool

ZIGLIO, Frediano, VF-IT Frediano.Ziglio at vodafone.com
Tue Jun 10 04:26:36 EDT 2008

> Federico Alves wrote:
> > Any chance we could update tdspool? 
> I myself have no intention on working on tdspool.  Patches 
> are of course
> welcome!  

I can confirm, personally I have very few time to improve it.
However to solve the problem personally I would work on odbc pooling.
The reasons are
- you are using odbc
- it's easier to implement cause 
 - you have just to reset connection, put in the pool and retrieve on a
new connection
 - you have the clear password to check (tdspool require network parsing
which is difficult using ntlm or kerberos)
- it's faster (you don't need an extra process which add latency)
The only problem I see is that odbc pooling work on a single process,
but with current multi-thread applications it's not a big problem. It
could be fixed duplicating file descriptors across processed but I don't
know a way to do it under *nix.


More information about the FreeTDS mailing list