I am sorry for resurrecting this post but I am facing the same problem. Using the web2py -F profile option and then reading that dump with runsnakerun I see that fcntl.flock is where more time is spent. I have disabled migrations in the appconfig.ini and in the DAL constructor but I am still having fcntl.flock slowing my requests.
What fcntl.flock does? Is it directly related to migrations? If so how can I definitely get rid of this bootleneck? Thank you. El lunes, 5 de septiembre de 2016, 10:25:12 (UTC-5), Scott Hunter escribió: > > I found the problem; migrations weren't being disabled because, while I > had the correct settings in appconfig.ini, they were not being used > properly in creating the db object. > > - Scott > > On Saturday, September 3, 2016 at 1:01:39 PM UTC-4, Scott Hunter wrote: >> >> I have an app which runs *much* slower than expected on pythonanywhere. >> >> I enabled the profiler, and one thing stood out: on pythonanywhere, the >> function the most time was spent in, by far, is fcntl.flock from inside >> portalocker while loading the models. In fact, when run locally, >> portalocker isn't even being called: it goes straight from save_dbt to >> cPickle.dump. >> >> Is there some configuration specific to web2py that might address this, >> or is it something I should take up w/ the people at PA? >> >> - Scott >> > -- Resources: - http://web2py.com - http://web2py.com/book (Documentation) - http://github.com/web2py/web2py (Source code) - https://code.google.com/p/web2py/issues/list (Report Issues) --- You received this message because you are subscribed to the Google Groups "web2py-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to web2py+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
