diff options
author | Russell Bryant <rbryant@redhat.com> | 2013-06-18 14:33:29 -0400 |
---|---|---|
committer | Russell Bryant <rbryant@redhat.com> | 2013-06-18 14:44:15 -0400 |
commit | ee5d9ae8d376e41e852b06488e922400cf69b4ac (patch) | |
tree | 29a45d56c71db9db252955ef5069c88891ad884e /nova/utils.py | |
parent | 880ddfb124428116600ad66de06c817c9f9de486 (diff) | |
download | nova-ee5d9ae8d376e41e852b06488e922400cf69b4ac.tar.gz nova-ee5d9ae8d376e41e852b06488e922400cf69b4ac.tar.xz nova-ee5d9ae8d376e41e852b06488e922400cf69b4ac.zip |
Revert "Delegate authentication to quantumclient"
This reverts commit dd9c27f999221001bae9faa03571645824d2a681.
When this patch was merged, we were suspicious about whether it was safe
to cache the client object and have it used by multiple greenthreads.
We decided it was safe and merged it. After thinking about it and
discussing it further, it is really a bad idea. Sharing httplib2
connections is not considered thread-safe.
quantumclient.client.HTTPClient inherits from httplib2.Http. The
following page says sharing instances of this object between threads is
not safe:
https://developers.google.com/api-client-library/python/guide/thread_safety
"The google-api-python-client library is built on top of the httplib2
library, which is not thread-safe. Therefore, if you are running as a
multi-threaded application, each thread that you are making requests
from must have its own instance of httplib2.Http()."
Potentially fix bug 1192131. Even if it doesn't fix that bug, this
needs to be reverted anyway.
Change-Id: I2e4bf5e7b6458cd7b76e30847fe07f06b25c34f7
Diffstat (limited to 'nova/utils.py')
0 files changed, 0 insertions, 0 deletions