I think there's little to add to what Aaron said.
This mechanism might end up generating long-running sql transactions which
have a detrimental effect on the availability of connections in the pool as
well as the threat of the deadlock with eventlet.
We are progressively removing all the controlle
In my experience doing any kind of http request inside a of a db
transaction kills performance vastly (and can lead to situations where
deadlock often occurs due to eventlet+sqlalchemly). This topic also was
recently discussed here:
http://lists.openstack.org/pipermail/openstack-dev/2013-November/0
Have a question regarding calling an external SDN controller in a plugin.
The ML2 model brings up the fact that it is preferred not to call an
external controller within a database session by splitting up each call
into two calls: *_precommit and *_postcommit. Makes sense.
Looking at the existin