I recently benchmarked my two apps with apache benchmark just to see 'req/s' and found out(accidently) in task manager, that both leak memory somehow. e.g. in one of my app python interpreter has grown from 70kB to 110kB just after 10K requests. Second app was sightly better but still it eats memory constantly. I tested on latest web2py trunk version. I bechmarked just app home page, its was common page with some text and login form. Migration was set off for both. My common observation was that more leaks were in app with more models.
Are there any best practices to avoid memory leaks in web2py apps? Any experience? David --
