Our platform consists of two servers with database postgresql v9.3, v3.3 pgpool for high availability on debian OS v7, and we have many connections webserver servers. The problem is presented when down two servers in weblogic that are clustered, and the pgpool connection status discard all remain, making it impossible to back up both servers WebLogic, to solve it we must lower the pgpool of each nodes so that sessions that were to discard all canceled status and weblogic can start and open connections without error.
We would like to know if there are any configuration parameter that can kill pgpool sessions remaining discard all status.
Alexander Angel
DAB Postgresql
Venezuela