Agent Terminal User Interface (agent-tui)
This tool aims to improve the installation outcomes where:
- the provided manifests do not match the situation found at runtime and re-configuration is necessary, or
- an interactive installation is being performed (not yet implemented)
Thus, it is only aimed at agent-based installation and not for the managed service.
In the first case, it will perform some checks to determine if the installation can proceed. If it all the checks come
back successful, the tui will quit and installation will proceed. If some check fails, it will present the user with
information relative to the failed checks and give the user a chance to correct those.
How is it built?
It is built using the tivo/tview golang library
How do I run it locally?
The tool has a dependency on the nmstate library. Install the library for your OS. For CentOS, the library is in the nmstate-libs package:
dnf install nmstate-libs
Then build and run the agent-tui.
./hack/build.sh
RELEASE_IMAGE=quay.io/openshift-release-dev/ocp-release:4.12.0-rc.7-x86_64 ./bin/agent-tui
RELEASE_IMAGE is the URL to a OpenShift release image.
Will this grow to be an entire agent based TUI for interactive installation?
It is not likely
Where can I get support for it?
The right place is in https://issues.redhat.com. Filing an OpenShift bug (OCPBUGS project) for the "Installer / Agent
based installation" component