On 01/12/2014 08:00 AM, Gary Kotton wrote:
Hi,
I was looking into SSH failures and stumbled upon the following tempest
issue. The validations would not take into account the neutron agent
heartbeat timeout. This would cause tests to occasionally fail. In short
there was a window of at most 4 second for the test test_list_agent to
run in order to pass. Please
see: https://bugs.launchpad.net/tempest/+bug/1268293
Thanks
Gary
Yeh, I saw a related review like that the other day. I think a bunch of
the checks currently do dictionary compares, which is terrible for
things that have timestamps.
I think we need to route out any tests which do that and unroll them to
attribute compares for specific attributes that we know are stable.
This is a broader issue, and I'm actually trying to figure out if there
is a good way to enforce this in the code programatically, as we've had
to clean a few of these up before. And it's not entirely straight
forward in review that these things would cause an issue.
-Sean
--
Sean Dague
Samsung Research America
s...@dague.net / sean.da...@samsung.com
http://dague.net
_______________________________________________
OpenStack-dev mailing list
OpenStack-dev@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
