On 06/14/2013 11:20 AM, Steve Baker wrote: > Enabling heat on an otherwise default devstack will increase the > keystone catalog so that the resulting v3 token header exceeds > eventlet.wsgi's limit [1] of 8192 bytes. > > This is currently preventing Heat from being enabled in gating [2][3], > and others have hit this limit too [4]. > > The catalog-optional blueprint [5] might help, but it would probably > need to become the default fairly soon, since the current token size > is not far under the limit. > > One possible (unverified) fix would be to patch every project's > wsgi.py to raise eventlet.wsgi. MAX_HEADER_LINE to provide some > headroom (32768? that leaves another 32768 for other headers) > I can confirm that adding this to nova/wsgi.py allows the failing tests to pass eventlet.wsgi.MAX_HEADER_LINE = 32768 > Is there any objection to this? I can start with nova and keystone to > see if it fixes the failing tempest tests. > > [1] https://github.com/eventlet/eventlet/blob/master/eventlet/wsgi.py#L19 > [2] https://bugs.launchpad.net/keystone/+bug/1190149 > [3] https://review.openstack.org/#/c/32002/ > [4] https://bugs.launchpad.net/keystone/+bug/1186177 > [5] https://blueprints.launchpad.net/keystone/+spec/catalog-optional > -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20130614/82f4d02d/attachment.html>