> Could we refactor this so that we use it in tests/run.py to clean up
> *just* the security groups that were created in the test run (minus
> precious)?
That's part of what I'm doing in lp:~vila/ubuntu-ci-services-itself/integration and I'd rathe avoid duplicating the work especially since we don't yet know who/why these groups fail and a fix can come magically from juju or juju-deployer.
In the mean time, I think cleaning up after a test run while collecting data is the minimum we should do.
> Could we refactor this so that we use it in tests/run.py to clean up
> *just* the security groups that were created in the test run (minus
> precious)?
That's part of what I'm doing in lp:~vila/ubuntu-ci-services-itself/integration and I'd rathe avoid duplicating the work especially since we don't yet know who/why these groups fail and a fix can come magically from juju or juju-deployer.
In the mean time, I think cleaning up after a test run while collecting data is the minimum we should do.