Closed sjpb closed 1 month ago
CI clusters are deleted if the OpenTofu provisioning fails or the workflow is cancelled, but are left up if the Ansible fails to allow debugging.
This PR adds a workflow nightly to clean up any remaining CI clusters, unless the control node has been tagged with "keep", e.g. using:
openstack server set slurmci-RL9-1374-control --tag keep --os-compute-api-version 2.26
It also amends the deployment/reimage workflow so that it is not triggered by changes to workflows, other than itself.
CI clusters are deleted if the OpenTofu provisioning fails or the workflow is cancelled, but are left up if the Ansible fails to allow debugging.
This PR adds a workflow nightly to clean up any remaining CI clusters, unless the control node has been tagged with "keep", e.g. using:
It also amends the deployment/reimage workflow so that it is not triggered by changes to workflows, other than itself.