This group is locked. No changes can be made to the group while it is locked.
Re: Running tempest tests -- using local vm (from Vagrant) -- cont.
Flavio Fernandes <ffernand@...>
toggle quoted message Show quoted text
I created a wiki page  which may be easier to refer to as well as improve on this subject.
Please change or add anything that is not good enough. Goes w/out saying, but if something is not clear,
don’t be shy to reach out!
On Mar 24, 2015, at 1:41 PM, Mathieu Lemay <mlemay@...> wrote:Sounds good thanks Flavio...On Tue, Mar 24, 2015 at 1:38 PM, Flavio Fernandes <ffernand@...> wrote:Sorry folks; one important detail I failed to mention.If you use optional step 2 (using odl outside the vm), you will also need to set the port used by neutron.By default, the all-in-one option changes that to 8087 to avoid a conflict with the swift service  in the same vm.Otherwise, the port ODL uses is 8080 and that is what we would need to set it to:ODL_PORT=8080I will be putting all this in a wiki page, so we can refer to it.— flavioOn Mar 23, 2015, at 8:37 PM, Flavio Fernandes <ffernand@...> wrote:Running tempest tests -- using local vm
This is a short an simple set of steps you can take to get going on tempest tests.
Big thanks for enabling this goes to Kyle and Swami!
git clone git@...:mestery/odl-openstack-ci-1.git && cd odl-openstack-ci-1/centos
# look in Vagrantfile file and make sure private_network has an address that is
# valid for your system. In my system, I run ODL locally, which is reachable by
# the vagrant vm if I give it an address of 192.168.50.x/24
vagrant up && vagrant reload && vagrant ssh
# at this point, you are ssh’ed into the centos7 vm
# I like to have control over doing the stacking, as well as running the
# tests manually. Because of that, some tweaks are needed.
# Comment out the following lines:
trap archive-logs EXIT
# to see the diff, look at: https://gist.github.com/79c7a6289242489534d6
# at this point, run the modified script
# OPTIONAL 1 of 2 -- start
# let's say you wanted to use a Lithium version of ODL. For that, these are the
# variables you would set in local.conf.
# OPTIONAL 1 of 2 -- end
# OPTIONAL 2 of 2 -- start
# Note that the above would still give you an ODL running inside the VM
# if you wanted to have ODL runinng outside the VM, you can change local.conf
# to look like this:
# while running ODL 'externally', make sure to load the features and set logging
log:set TRACE org.opendaylight.neutron
log:set TRACE org.opendaylight.ovsdb
log:set INFO org.opendaylight.ovsdb.lib
log:set DEBUG org.opendaylight.ovsdb.openstack.netvirt.impl.TenantNetworkManagerImpl
log:set INFO org.opendaylight.ovsdb.plugin.md.OvsdbInventoryManager
log:set error org.opendaylight.openflowplugin.applications.statistics.manager
# OPTIONAL 2 of 2 -- end
# stack using defaults in local.conf. That will give you a Helium ODL running inside the vm.
cd /home/vagrant/jenkins-openstack-gerrit-1/devstack/ && time ./stack.sh
# FYI: my first stack.sh took 45 minutes. Following stacks takes about 4 minutes.
# Make sure stack finishes cleanly. If it did, you will someting like:
# Horizon is now available ....
# At this point, you should have a stacked Openstack running in the VM. Nice! :)
# You may consider 'freezing' the pkgs used by modifying local.conf:
# to unstack:
cd /home/vagrant/jenkins-openstack-gerrit-1/devstack/ && ./unstack.sh
# to run tempest; starting with a subset that should give you no failures while using ODL:
./run_tempest.sh tempest.api.network.test_networks.NetworksTestJSON \
# Here is a test that is known to fail on Helium:
cd /opt/stack/tempest && ./run_tempest.sh tempest.api.network.test_networks.NetworksIpV6TestJSON
# However, this maybe a tomcat being too slow thing, as running the test that failed on its own is not
# a problem:
cd /opt/stack/tempest ; \