Bootstrap GitOps stack (ArgoCD, ESO, Longhorn) onto existing Talos cluster
Deploys GitOps stack (ArgoCD, ESO, Longhorn) onto existing Talos clusters with automated bootstrap workflow
/plugin marketplace add basher83/lunar-claude/plugin install omni-scale@lunar-claudeDeploy the GitOps stack onto an existing talos-prod-01 cluster.
Note: This is a WIP command that will be refined as bootstrap procedures evolve.
Before starting:
/omni-scale:status$INFISICAL_CLIENT_ID and $INFISICAL_CLIENT_SECRET set in environmentmcp__plugin_omni-scale_kubernetes__kubectl_get(resourceType: "nodes")
All nodes should show Ready. If not, stop and troubleshoot.
The one manual secret for Infisical authentication:
kubectl create namespace external-secrets
kubectl create secret generic universal-auth-credentials \
--namespace external-secrets \
--from-literal=clientId="$INFISICAL_CLIENT_ID" \
--from-literal=clientSecret="$INFISICAL_CLIENT_SECRET"
If command fails → Stop → Inform user env vars are not properly set.
mcp__plugin_omni-scale_kubernetes__kubectl_apply(manifest: "https://raw.githubusercontent.com/basher83/mothership-gitops/main/bootstrap/bootstrap.yaml")
This deploys the App of Apps which manages everything else.
mcp__plugin_omni-scale_kubernetes__kubectl_get(resourceType: "applications", namespace: "argocd")
Expected wave order:
mcp__plugin_omni-scale_kubernetes__kubectl_get(resourceType: "applications", namespace: "argocd")
All apps should show Synced and Healthy.
After Longhorn is healthy:
argocd-ha applicationThis is intentionally manual as a safety gate.
ESO not creating secrets:
Longhorn volumes pending:
Apps stuck OutOfSync:
If bootstrap fails mid-way, delete and retry:
mcp__plugin_omni-scale_kubernetes__kubectl_delete(manifest: "https://raw.githubusercontent.com/basher83/mothership-gitops/main/bootstrap/bootstrap.yaml")
Fix the issue, then reapply Phase 3.