Progress and evidence register¶
Updated: 2026-10-15 (Africa/Nairobi).
This register records user-shared terminal output and completion reports. It is not an independent live audit of the Hyper-V host. Selected original screenshots are embedded in the installation and first-boot guides. They show only the personal lab; credentials, cookie-bearing HTTP output and unrelated personal storage-cleanup material are not included.
The screenshot index lists all 51 included images, their purposes and the reasons other supplied images were not published.
| Milestone | Status | Evidence / limit |
|---|---|---|
| Windows source repository | Verified | main branch, .gitignore and README committed and pushed |
| CRC cleanup | Verified | Uninstall reported; Test-Path for .crc returned False |
| Hyper-V lab switch | Verified | New-VMSwitch returned reservoir-lab-switch / Internal |
| Lab adapter | Reported verified | User reported it listed with status Up |
| Gateway address | Verified | 192.168.200.1, prefix 24, ActiveStore Preferred |
| NAT | Verified | reservoir-lab-nat, 192.168.200.0/24, Active=True |
| ISO | Verified | RHEL 9.8 x86_64 DVD SHA-256 matches vendor value |
| Control-plane VM creation | Reported verified | User reported New-VM succeeded |
| CPU configuration | Verified command completion | Set-VMProcessor -Count 2 returned without errors |
| Secure Boot, ISO and boot order | Reported verified | User reported each command succeeded |
| VM boot | Verified | User reported RHEL installer boot menu visible |
| Installer software/time settings | Verified in screenshot | Server, English (US), Africa/Nairobi |
| Administrator account | Verified | Console login succeeded and sudo whoami returned root |
| Hostname | Verified in screenshot | Current host name lab-control-01 |
| Installer IPv4 settings | Reported verified | User reports 192.168.200.10 and DNS servers; summary shows eth0 connected |
| Installation disk selection | Verified in screenshot | 40 GiB sda; Automatic; final summary has no destination warning |
| RHEL installation | Verified boot | Screenshot shows installed RHEL 9.8 login prompt after reboot |
| Post-install address and routes | Verified in screenshot | eth0 UP, 192.168.200.10/24, default via 192.168.200.1 |
| DNS and HTTPS | Verified in supplied output | getent returned IPv4; curl received HTTP/2 301 from Red Hat |
| SSH service and TCP reachability | Verified | sshd active; Windows TcpTestSucceeded=True on port 22 |
| SSH login | Reported verified | Initial password failures resolved after private password change |
| RHEL registration | Reported verified | User confirmed interactive registration succeeded |
| BaseOS and AppStream | Reported verified | User confirmed both enabled in dnf repolist |
| Repository access | Reported verified | Metadata refresh and update check succeeded, including after subscription warning |
| First worker VM | Created and verified | Generation 2, two CPUs, fixed 5120 MiB, 40 GB disk; RHEL 9.8 installed and booted |
| RHEL update | Verified | Running kernel 5.14.0-687.57.1.el9_8.x86_64 |
| VM memory | Reported verified | Dynamic Memory disabled, startup 3072 MiB; guest about 2.6 GiB |
| Swap | Verified after reboot | Swap 0B and zero active swap units |
| Time sync | Verified | chronyd enabled/active, selected source and clock synchronised |
| Security baseline | Updated | SELinux Enforcing, cgroup v2; firewalld subsequently disabled for Calico, with backup |
| Kernel network settings | Verified applied | overlay and br_netfilter loaded; three sysctl values 1; reboot persistence not yet retested |
| CRI-O | Verified | crictl reports runtime 1.36.7, CRI v1, RuntimeReady true |
| Kubernetes packages | Verified | kubeadm/kubelet/kubectl 1.36.5; cri-tools 1.36.0; kubernetes-cni 1.9.1 |
| Kubelet boot enablement | Verified | systemctl is-enabled returned enabled |
| Runtime API checks | Verified | RuntimeReady true; NetworkReady false before CNI |
| Cluster creation | Verified initialisation | Control-plane Pods Running; initial NotReady/Pending resolved after CNI |
| Calico host preparation | Verified | NetworkManager exclusions, VXLAN loaded, firewalld disabled/inactive |
| Calico operator | Verified | 3.33.0 manifests; rollout succeeded and Pod 1/1 Running |
| Calico networking | Verified ready | Node Ready; Calico Available True, Progressing False, Degraded False; VXLAN/BGP/CIDR configuration observed |
| CoreDNS | Verified | Both Pods 1/1 Running, zero restarts, Pod addresses 10.244.242.3 and .4 |
| Pod DNS and cleanup | Reported verified | Internal and external lookups succeeded; temporary dns-check Pod deleted |
| GitHub runner and local registry | Planned | Selected co-location on supporting RHEL VM outside Kubernetes |
| Worker networking | Verified repaired | Dedicated lab-worker-eth0 profile active, gateway reachable, DNS and HTTPS successful; reboot persistence user-confirmed |
| Worker time sync | Verified | Selected NTP source, Leap Normal, clock synchronised yes |
| Worker registration | Verified in private screenshot | Second attempt succeeded; identifying output excluded from publication |
| Worker OS update | Completed | DNF Complete; kernel 5.14.0-687.58.1 installed; reboot checks user-confirmed |
| Worker swap | Reported verified after reboot | Swap stayed off |
| Worker modules and runtime | Verified | overlay, br_netfilter, vxlan; sysctls 1; CRI-O startup and RuntimeReady true |
| Worker cluster join | Verified in private screenshot | Join succeeded; token-bearing capture excluded; both nodes Ready user-reported |
| Cross-node Pod HTTP | Verified in screenshot | Both directions returned the destination Pod hostname |
| Cross-node Service and DNS | Verified / reported | Worker DNS visible; both-node DNS user-confirmed; corrected worker-only EndpointSlice and Service HTTP verified |
| Network test cleanup | Verified | network-check namespace deleted; final two-node Ready state user-reported |
| Data/AI services and MCP | Planned | No deployments in the rebuilt lab |
Second worker and reboot recovery - 2026-10-15¶
| Milestone | Status | Evidence / limit |
|---|---|---|
| Existing two-node reboot recovery | Verified | Both Ready; Calico available; worker 01 services enabled/active and swap off |
| Worker 02 VM and RHEL | Verified | Gen 2, 2 CPUs, fixed 5120 MiB, 40 GB disk; installed RHEL 9.8 boot |
| Worker 02 network repair | Verified | Named profile, 192.168.200.12/24 and gateway .1; retained after reboot |
| Worker 02 registration and update | Verified | Registration and cache completed privately; DNF Complete and updated running kernel captured |
| Worker 02 swap and networking | Verified after reboot | Swap 0B; three modules loaded and sysctls 1; SELinux Enforcing, cgroup v2 |
| Worker 02 join and readiness | Verified | Three Ready nodes; kubelet/CRI-O enabled/active; runtime and network ready |
| Worker-to-worker HTTP | User-confirmed | Both direct Pod-IP directions passed |
| Worker DNS and Service | Verified / reported | Both-node DNS user-confirmed; worker 02 DNS and uniquely selected Service endpoint/response visible |
| Test cleanup | User-confirmed | worker-network-check deleted; all three nodes remain Ready |
| Worker 02 post-join reboot | Pending | Earlier prerequisite reboot is verified, but not a reboot after joining |
Resolved symptom, unknown cause¶
Switch creation initially failed with 0x800700B7. Read-only switch, adapter
and Plug and Play inspections did not identify a conflicting lab device.
A retry later succeeded. A restart was suggested but not confirmed. The
root cause must remain unknown rather than being presented as a proven fix.
Decisions and open questions¶
- Use Hyper-V as the local substitute for infrastructure-provided VMs.
- Run upstream Kubernetes inside the RHEL guests, not on OpenShift Local.
- RHEL 9.8 chosen for this lab; reference environment version still unknown.
- Server selected without GUI or optional add-ons; administer through terminal/SSH.
- Guest address 192.168.200.10/24 with gateway 192.168.200.1; public DNS for the home lab.
- Root password set privately; root password SSH was instructed off and needs verification.
- Disk layout verified: EFI, XFS boot, LVM root and swap; swap disabled without deleting its volume.
- CRI-O 1.36 selected; Kubernetes 1.36.5 packages installed. Calico 3.33.0 selected with VXLAN and BGP disabled.
- Podman is intended for image workflows, not as the Kubernetes CRI runtime.
- One control plane initially; high availability is not provided.
- Add services incrementally within laptop capacity.
- Public evidence must use synthetic data and omit credentials/internal documents.
- Satellite/package mirroring was discussed for understanding, not added to scope.
- Read-only MCP is deferred until Kubernetes is operational.
Next step¶
Verify worker 02 post-join reboot recovery. The three-node cluster and worker networking tests are complete; the GitHub runner/local registry VM remains planned. See second worker.