Home › Docs › Reset & reinstall

Reset: wipe and reinstall from scratch

If a VM gets into a bad state, or you want a guaranteed-clean baseline, tear the whole thing down and start over. There are three levels, from lightest to a full nuke.

1. Rebuild the VM, keep the image cache

klimax destroy deletes the VM, every cluster, and the host route — but with the default host-backed cache it leaves ~/.klimax/registry-cache/ in place, so the rebuild pulls from local cache instead of the internet:

$ klimax destroy        # VM + clusters + host route (sudo)
$ klimax up             # fresh VM, warm registry cache

This is the same destroy && up you run after a klimax upgrade — a rebuilt VM with the image cache intact.

2. Also drop the registry cache

To force cold pulls again — say you want to reclaim disk or test a clean network path — clear the mirror caches too:

$ klimax registry clean-cache   # remove mirror containers + cache dirs
$ klimax destroy
$ klimax up

3. Full clean install — remove all state

The nuclear option. This deletes klimax's entire state directory, including config.yaml, the Lima instance, and the registry cache — the equivalent of a first-ever install:

$ klimax destroy              # tear down VM, clusters, host route (sudo)
$ rm -rf ~/.klimax            # config, Lima instance, registry cache
$ rm -rf ~/.kube/klimax       # per-cluster kubeconfigs (optional)
$ klimax up                   # regenerates a default config + fresh VM

See also

• Upgrading klimax — the everyday destroy && up you run after a new release.
• Configuration — what lives in ~/.klimax/config.yaml before you delete it.
• Registry mirrors — what the image cache holds and how clean-cache works.