## Do

- ops resources: flag to choose bulk vs per-ns stats strategy, pluggable transport beyond podman exec
- replace kubectl/crictl exec chains with podman API and client-go for cluster operations
- Skip reapply option for resources already present, keyed by GVK + namespaced name
- force-conflicts flag on server-side apply for configure (needs apply.go refactor)
- StorageCluster configuration: expose key fields (sizing, device class, placement) as configure flags
- Hub and spoke: multi-cluster topology with a hub managing spoke MicroShift clusters
- Infrastructure configuration: node labeling, machine config, kernel module management
- ODF Disaster Recovery: Regional-DR and Metro-DR using Ramen/ACM on MicroShift

## Done

- Cluster lifecycle: create, start, stop, delete, list, exec, kubeconfig
- LVM thin pool backend for TopoLVM with loop device, force cleanup of dm devices on delete
- Podman network per cluster, IDMS mirror support with generated registries.conf and policy.json
- ODF addon: install (shim CRDs/CRs/RBAC/oauth, catalog, subscriptions), configure, uninstall
- ODF configure: SINGLE_NODE patch via jsonpath discovery, node labeling, resources/ applied
- Shim and resource files numbered for predictable apply order
- perms create/delete: sudoers for passwordless sudo including modprobe and install
- addon odf modules load/unload: writes modules-load.d conf and loads/unloads rbd/ceph/nbd for current session
- support.RunPrivileged/WritePrivileged: single OS-aware privileged execution abstraction
- Custom network subnet via --network-subnet (default 172.20.0.0/24, avoids common VPN ranges)
- Kubeconfig port rewrite: server URL updated to match --api-server-port after copy from container
- docs command: CLI reference, examples, project skill, and hidden devlog
- embedded examples document via --examples
- ops resources (experimental): CPU/mem requests limits and live usage per container, per-ns totals, node summary, filterable by ns and node
- MicroShift power tuning on create, opt-out via --no-power-tuning
- ExposeKubeAPI: generate cert with all host IPs and hostname as SANs, merge all kubeconfigs into single file with multiple contexts per client
- LoadBalancer services: node acting as load balancer working
- cluster: test running multiple named clusters simultaneously to verify config dirs, networks, and loop devices don't collide
- VM access: cluster running in a VM should be reachable from the local machine without manual kubeconfig or cert fixups
- Overlayfs deadlock: fix kubeconfig copy race condition that triggers machine restart on rapid file access (antivirus was the issue)
- Multus configuration: secondary network attachment for storage traffic isolation in various groups via VLANs
- Network listing: show names of saved dfmicro bridge networks
- Multi-node clusters: worker onboarding, readiness gating, and node slot reuse
- Whereabouts IPAM: cluster-wide secondary network allocation across worker nodes
- Build metadata and examples: UTC version timestamp and tested MicroShift image defaults
- cluster peering for non-overlapping service and cluster ip ranges
- cluster-specific route base domains under `<cluster>.dfmicro.io`
- ODF storage topologies: Rook Ceph provider with external consumer, and ODF provider with client-only consumer
- Test suite levels 4-6: single-node ODF, external storage, and multi-node provider/client PVC read-write coverage