Third option would be to modify global socket timeout just for
this request and set it back to default after request is processed and/
or
timeouts are handled.
Same problems can actually hit you with Plone RSS portlet. Feedparser
doesn't support setting the timeout for the request even though it
Jukka Ojaniemi wrote
Third option would be to modify global socket timeout just for
this request and set it back to default after request is processed and/
or
timeouts are handled.
Sorry but this is a pretty stupid idea.
Don't touch global timeout..it will break your neck earlier or
It could be useful for somebody else, so here's a possible solution:
http://opensourcehacker.com/2011/08/02/ajax-proxy-view-with-python-urllib-and-plone/
Checkout the socket.setdefaulttimeout(10) section.
It's not a great solution (as pointed out by ajung) but it works, and the
finally block
Yes, that's what happens, all threads busy and waiting for a response.
Any idea is really appreciated.
Thanks,
Marco
--
View this message in context:
http://plone.293351.n2.nabble.com/Timeout-decorator-on-external-service-call-tp7063682p7066429.html
Sent from the Product Developers mailing list
You have two choices.
1. either handle the request using another framework that is much
lighter weight such as pyramid, grok. pyramid running on gevent is a
good choice.
2. create a 2nd set zope instances which more threads per process and
lower zodb cache so you can have more to handle
Hi,
I use YouTube Data API (
http://code.google.com/apis/youtube/1.0/developers_guide_python.html
http://code.google.com/apis/youtube/1.0/developers_guide_python.html ) to
fetch videos and other youtube informations.
Sometimes (not that often, but it happened) the python call freezes while
waiting
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
It's truly not recommendable calling external webservice from
within a worker thread. If all of the (four) worker threads of Zope run
into a timeout then you are doomed. Your complete system will be blocked
and modifying the socket time out is also a