It would help if we could isolate the problem. How are you measuring it? make sure you are not running cron jobs in background (with <2.1 cron was always enabled) moreover web2py's memory usage may grows for the first 100 requests than some stuff is garbage collected and it stabilizes.
massimo On Monday, 15 October 2012 15:39:30 UTC-5, Bill Thayer wrote: > > That's all I have right now. Haven't tried it with SQLLite. > > > > On Monday, October 15, 2012 3:35:25 PM UTC-5, Massimo Di Pierro wrote: >> >> I this only work Oracle? >> >> On Monday, 15 October 2012 14:56:44 UTC-5, Bill Thayer wrote: >>> >>> >>> Perhaps someone can check session.connect with Oracle for a rocket >>> server memory leak? >>> To reproduce on windows 7 using web2py 2.1.1 from source: >>> 1) Create a new simple application called simpleapp where simpleapp is >>> the name of an application with an Oracle table for storing sessions >>> 2) Create a new 0.py model file & >>> set:settings.database_uri='oracle:username/password@server:1521/servicename' >>> 3) in db.py: >>> line 14: >>> >>> db = DAL(settings.database_uri, check_reserved=['oracle']) >>> >>> >>> line 15: & 16: >>> ## store sessions and tickets there >>> session.connect(request, response, db=db, migrate= >>> 'db.web2py_session_simpleapp') >>> >>> >>> >>> line 29: >>> >>> auth.define_tables(username=True, signature=False, migrate=False) >>> >>> Takes about an hour or less for rocket server to run out of memory so >>> that it will just just spin and not respond to the stop server button click. >>> >>> >>> --
