openshift-client-runner¶

Runner provides a fedora-base for openshift commands

The openshift-client-runner context is a wrapper for the fedora/latest container.

This context is required for running the OpenShift CLI tools for managing OpenShift clusters.

Secrets handling¶

The container mounts a tmpfs (RAM-backed, 6GB cap, mode 0700) over /root/cluster. All installer state, kubeconfigs, and PXE build artifacts are therefore never written to the container's disk layer and are destroyed when the container stops. The on-disk kubeconfig is also removed at the end of every run, success or failure, via the cluster prelude's EXIT trap, so an idle container holds no secret files outside the install window. A failed install is the exception by design: kubeadmin-password and .openshift_install_state.json remain on the tmpfs until the retry completes, since the installer needs them on the next attempt. After a container restart, tasks that need the kubeconfig rematerialize it from the openshift/kubeconfig param twice over: the cluster prelude rewrites it at the start of every consumer task, and those tasks also declare the openshift-refresh-kubeconfig prerequisite. A restart during the install window (between image generation and install-complete) loses installer state; re-run the install pipeline from openshift-cluster-prep in that case.

Objects that Reference openshift-client-runner¶

Not used in current content packs

Objects used by openshift-client-runner¶

Not used in current content packs