Check the code on gw/dlr_mysql.c: pconn = dbpool_conn_consume(pool);
<do your stuff here with pconn> dbpool_conn_produce(pconn); Hope it helps, -- Alejandro Guerrieri [email protected] On 28/01/2010, at 23:17, Nikos Balkanas wrote: > No. There is no point in having a different pool/thread, otherwise it is a > buffer, not a pool. When using the DB (insert, select, update, etc.) use > gwlist_consume and when finished with the SQL return it with gwlist_append. > Use a common pool for all threads. > > If you need more help check source code. > > Nikos > ----- Original Message ----- > From: Esteban Cacavelos > To: Nikos Balkanas > Cc: devel Devel > Sent: Thursday, January 28, 2010 10:52 PM > Subject: Re: dbpool problem > > Do you mean use gwlist_consume instead gwlist_get (in dbpool_check) and > gwlist_append instead gwlist_produce (in dbpool_increase) ? > > or to use directly to get connections ? > > > My SOLUTION (for the moment) is to hold one pool per thread and it works ;). > > > -- > Esteban L. Cacavelos de Amoriza > Cel: 0981 220 429
