Il giorno 06/apr/2011, alle ore 12.49, Pavel Kolla ha scritto:
> Hello
> I am using nginx/0.8.54 + uwsgi/0.9.6.6 + python/3.1
> (SQLAlchemy/Postgres) to power my backend and once in a while i
> encounter some issue (usually infrastructure related) which presents
> itself as a uWSGI 502/nginx 500 error with very little info available
> even in debug mode -
>
> pkolla:88 [pid: 32734|app: 0|req: 14/15] 192.168.1.160 (test) {66 vars
> in 1150 bytes} [Wed Apr 6 13:07:55 2011] POST
> /query/ListPersonsByLastnameLike?nocache=1302080877506 => generated 0
> bytes in 1 msecs (HTTP/1.1 500) 0 headers in 0 bytes (0 async switches
> on async core 0)
>
> How can I increase verbosity of these message to the level where i can
> figure out root of the problem easily?
> Current approach is relatively cumbersome and includes remote
> debugging sessions, aggressive application level logging, netcat and
> tshark...
>
>
Do you use some WSGI adapter in your custom app (like werkzeug or similar) ?
The only things that come to my mind is the exception_info usage (that has been
added
only in 0.9.7.2) not being used in 0.9.6
Can you check if your app/middleware make use of this argument ? (it is passed
as the third argument of start_response)
--
Roberto De Ioris
http://unbit.it
_______________________________________________
uWSGI mailing list
[email protected]
http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi
