Il giorno 03/set/2010, alle ore 15.14, Jean-Baptiste Quenot ha scritto:
>
> Yes you are right, I should have mentioned that. I am performing
> warmup of the application before setting up the WSGI stuff, for
> ensuring high availability:
>
> def myapp(environ, start_response=None):
> start_response("200 OK")
> return "OK"
>
> def warmup(app):
> # Perform startup tasks, calls subprocess.Popen()
> return app
>
> cgiMode = False
>
> if __name__ == '__main__':
> cgiMode = True
> CGIHandler().run(myapp)
> else:
> print "Loading..."
> applications = {'': warmup(myapp) }
> --
> Jean-Baptiste Quenot
>
This is an interesting question, how the uWSGI master process has to
behave on signal/wait4() received by non-uWSGI died processes ?
I think your situation is not so rare.
I see a couple of solutions
-- Simply log them with their return code
-- Simply ignore them
-- before forking the first worker call waitpid() in a loop to remove the
eventual queue
-- other ideas ?
--
Roberto De Ioris
http://unbit.it
JID: [email protected]
_______________________________________________
uWSGI mailing list
[email protected]
http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi
