Good question. We had a discussion awhile back on whether SSHing into an instance broke the line between black and white box testing. We had people on both sides, but the discussion was tabled for a bit while we dealt with other issues. It's definitely something I'd like to talk about again. I'm doing some SSHing in the dev branch that I work from and find it fairly useful.
Daryl On Jan 31, 2012, at 10:07 AM, David Kranz wrote: > The discussion about Libvirt File Injection on the main list reminded me that > I am not clear on what we mean that a functional test passes. For example, in > one version of the stress test we are going to check in, the test assigns a > floating ip and ssh's to the created server to make sure it is, by some > definition, working. It looks like the Tempest test of server creation passes > if the API response says it did. Is this adequate? Another example is that if > we are testing that a volume is attached to a server, don't we have to ssh to > the server and make sure that the volume is accessible? Perhaps these aspects > are tested by the nova unit tests but at first glance it did not seem so. > > -David > > -- > Mailing list: https://launchpad.net/~openstack-qa-team > Post to : openstack-qa-team@lists.launchpad.net > Unsubscribe : https://launchpad.net/~openstack-qa-team > More help : https://help.launchpad.net/ListHelp -- Mailing list: https://launchpad.net/~openstack-qa-team Post to : openstack-qa-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack-qa-team More help : https://help.launchpad.net/ListHelp