id summary reporter owner description type status priority milestone component version severity resolution keywords cc branch changelog apichanges internalchanges 4347 postgres connection limit exceeded nick.stone@… Jonas Borgström "I realise there seems to be loads of articles from the beginning of the year on this subject but nothing recently and we are experiencing the problem. We running with Trac 0.10.2, Postgres 8.1.5-1, Apache 2.2 and Python 2.4. Basically with Apache restarted there are no connections to Postgres but once a few pages are visited the number goes up and stays there. I notice in trac admin that there is a connection timeout which is set to 20 seconds (the default) which seems to have no effect. After a while the number of open connections exceeds the maximum number of available postgres connections and the sites fail. The only way to recover the situation is to restart apache. I notice from earlier mailing list entries that the problem also existed in tracd but nobody seemed to have posted a solution:-( We currently run something like 30 projects with each one being a schema within one Postgres DB. running ""ps ax"" gives: {{{ postgres: trac 127.0.0.1(36257) idle in transaction postgres: trac 127.0.0.1(36258) idle in transaction postgres: trac 127.0.0.1(36259) idle in transaction postgres: trac 127.0.0.1(36261) idle in transaction postgres: trac 127.0.0.1(36262) idle in transaction postgres: trac 127.0.0.1(36263) idle in transaction postgres: trac 127.0.0.1(36265) idle in transaction postgres: trac 127.0.0.1(36266) idle in transaction ... }}} If there's more info required please let me know - happy to assist as I really want to get a fix if possible. Thanks" defect closed high 0.11.1 general 0.10.2 critical fixed postgresql, mysql, schema, connection, pool pacopablo@… stephan.zeissler@…