On 21 January 2013 15:46, Łukasz Mierzwa <[email protected]> wrote: > 2013/1/21 Roberto De Ioris <[email protected]> >> >> >> > I get that, but having a cache that over which You have zero control >> > does >> > seems hackish to me. This is of course subjective and there might be >> > cases >> > when this is valid and solid solution. >> > It should be easy to add option that will inform uWSGI about worker max >> > lifetime, specially if this will make someone happy. >> >> >> That info is already available in last_spawn field of the uwsgi_worker >> structure. >> >> The master can simply check if uwsgi_now()-last_spawn > X and set >> manage_request to 0 + SIGWINCH to drop it. (SIGWINCH is required in case >> the worker is blocked in epoll_wait()/kevent()) > > > I'll make a pull request with such option, should be done soon.
Thanks a lot! That is very cool. FWIW, I share your general distaste that this is necessary, but in at least of couple of cases I am aware of it really is a good option. cheers, Yves -- perl -Mre=debug -e "/just|another|perl|hacker/" _______________________________________________ uWSGI mailing list [email protected] http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi
