-
Sub-task
-
Resolution: Unresolved
-
Medium
-
None
-
None
-
None
-
None
we had the idea of adding an initial set of health checks to some first suite where if anything
failed, we could just abort the job right then and there instead of letting the rest of the suites
spin in loops checking on failures that have no chance to succeed because something is
seriously broken. For example, if neutron-server is in FAILED state we could know that
straight away, abort the job and make it very clear.
some other ideas:
- is there br-int and it's connected
- is there netvirt:1
- are all the relevant openstack services running
- do we see all of the happy patch types of logs in karaf.log, openstack logs
- patch ports
- default flows
- tunnels
if there was a way to not need openstack involved here, maybe we could do this
kind of test with the netvirt project's karaf distro, instead of the int/dist karaf. There
has also been an ask to validate that distro as recently there was a bug introduced
when the aaa-cli.jar was added to the netvirt project which broke things. We never
noticed becuase that bug was not present in the int/dist version of things. just a
thought.