Currently we first stop the zk connection and afterwards the providers. This leads to a race condition where the providers may still want to communicate with zk while the connection is already dead. Instead stop the zk connection after stopping the providers. This also fixes some occational and misleading zk connection exception logs at the end of test cases. Change-Id: Ibb13f74a1130190403dd79922ccd6c2b75daffc9
Nodepool
Nodepool is a service used by the OpenStack CI team to deploy and manage a pool of devstack images on a cloud server for use in OpenStack project testing.
Developer setup
Make sure you have pip installed:
Install dependencies:
sudo pip install bindep
sudo apt-get install $(bindep -b nodepool)
mkdir src
cd ~/src
git clone git://git.openstack.org/openstack-infra/system-config
git clone git://git.openstack.org/openstack-infra/nodepool
cd nodepool
sudo pip install -U -r requirements.txt
sudo pip install -e .
If you're testing a specific patch that is already in gerrit, you will also want to install git-review and apply that patch while in the nodepool directory, ie:
Create or adapt a nodepool yaml file. You can adapt an infra/system-config one, or fake.yaml as desired. Note that fake.yaml's settings won't Just Work - consult ./modules/openstack_project/templates/nodepool/nodepool.yaml.erb in the infra/system-config tree to see a production config.
If the cloud being used has no default_floating_pool defined in nova.conf, you will need to define a pool name using the nodepool yaml file to use floating ips.
Export variable for your ssh key so you can log into the created instances:
Start nodepool with a demo config file (copy or edit fake.yaml to contain your data):
All logging ends up in stdout.
Use the following tool to check on progress: