d8a03801c2
We will now run the bootstrap-ansible and bootstrap-aio scripts in the Zuul pre-run job phase. This has several benefits: - Job stability will be higher. When we occasionally receive nodepool instances with unreliable networking, the job will often die off during the role clone due to git or pip internet resource unavailability. Jobs that fail during pre-run will be retried several times by Zuul using different nodes/clouds. - Jobs which inherit the integrated role's job definition may add pre-run playbooks of their own, which allow child jobs to configure OSA parameters in /etc/openstack_deploy (such as custom user variables or conf.d definitions). Child job pre-run playbooks will run after the parent's, so the child job can configure itself with the benefit of an already bootstrapped OSA AIO configuration. - The Zuul pre-run phase runs on a separate timeout from the run phase, so our run playbook will have more time available to it due to the bootstrapping being done before the run timeout begins. Environment variables are used to skip the bootstrap scripts when executing gate-check-commit during the job's run phase. Change-Id: I0e77423a1f28d9a53215ae68c1f9f3e2af00f340 |
||
---|---|---|
.. | ||
upgrade-utilities | ||
ansible-role-requirements-editor.py | ||
bootstrap-aio.sh | ||
bootstrap-ansible.sh | ||
fastest-infra-wheel-mirror.py | ||
federated-login.sh | ||
gate-check-commit.sh | ||
gen-config.py | ||
get-ansible-role-requirements.yml | ||
get-pypi-pkg-version.py | ||
inventory-manage.py | ||
keepalived_haproxy_check.sh | ||
keepalived_haproxy_notifications.sh | ||
keepalived_pingable_check.sh | ||
log-collect.sh | ||
openstack-ansible.rc | ||
openstack-ansible.sh | ||
os-cmd | ||
pw-token-gen.py | ||
rabbitmq-test.py | ||
release-yaml-file-prep.py | ||
run-upgrade.sh | ||
scripts-library.sh | ||
sources-branch-updater-lib.sh | ||
sources-branch-updater.sh | ||
update-revision.sh |