First Existing Host Install
Use the existing-host path when the target RHEL machine already exists and the Cockpit host can reach it over SSH.
Check Access
Run these checks from the Cockpit host before opening the UI:
ssh -i <ssh-key> <ssh-user>@<target-host> 'cat /etc/redhat-release'
ssh -i <ssh-key> <ssh-user>@<target-host> 'uname -m'
ssh -i <ssh-key> <ssh-user>@<target-host> 'sudo -n true'
Expected result:
- the target is RHEL 9 or RHEL 10
- the architecture is
x86_64 - the SSH user can run passwordless sudo
Choose Repository Access
Use one of the supported package access modes:
| Mode | Use it when |
|---|---|
| Preconfigured repositories | the target host is already registered and exposes microshift and openshift-clients |
| RHSM activation key | the UI should register the target with organization ID and activation key inputs |
Run The Install Flow
- Open
https://<cockpit-host>:9090. - Navigate to
MicroShift. - Select
Install MicroShift. - Choose
Use an existing RHEL host. - Enter the target host, SSH, registry authentication, repository, network, and MicroShift config fields.
- Continue to review.
- Inspect
config.yaml, request summary, and install plan. - Start deployment only after the generated plan matches the intended host.
After Install
Return to the inventory entry and confirm:
- deployment status is recorded
- target pattern is
existing-host - kubeconfig access details are present
- recent output does not show package, service, or readiness failures
Existing remote hosts are not treated as disposable targets. Destroy and clean
rebuild are intentionally focused on local create-host deployments.