Deploy and validate UDS Core on the stable RKE2 platform #10

Closed
opened 2026-08-29 00:22:50 -05:00 by portegumi · 3 comments
Collaborator

Description

Deploy UDS Core only after the underlying cluster prerequisites have been documented and validated. Keep UDS lifecycle work distinct from base RKE2 bootstrap so failures have a clear ownership boundary.

Scope

  • Select and record the initial UDS Core configuration.
  • Deploy UDS Core using the approved networking, storage, DNS, and certificate integrations.
  • Validate component health, persistence, access, and restart behavior.
  • Document routine operations and rollback or removal procedures.

Acceptance criteria

  • The selected UDS Core version and non-secret configuration are committed.
  • Required UDS components report healthy after deployment.
  • Storage and network integrations pass the agreed validation checks.
  • Restart and recovery behavior is tested and documented.
  • Rollback or clean removal steps are documented and tested to the agreed extent.

Dependencies / blockers

Blocked by completion of the UDS prerequisite/readiness issue and stable base networking and storage.

## Description Deploy UDS Core only after the underlying cluster prerequisites have been documented and validated. Keep UDS lifecycle work distinct from base RKE2 bootstrap so failures have a clear ownership boundary. ## Scope - Select and record the initial UDS Core configuration. - Deploy UDS Core using the approved networking, storage, DNS, and certificate integrations. - Validate component health, persistence, access, and restart behavior. - Document routine operations and rollback or removal procedures. ## Acceptance criteria - [x] The selected UDS Core version and non-secret configuration are committed. - [x] Required UDS components report healthy after deployment. - [x] Storage and network integrations pass the agreed validation checks. - [x] Restart and recovery behavior is tested and documented. - [x] Rollback or clean removal steps are documented and tested to the agreed extent. ## Dependencies / blockers Blocked by completion of the UDS prerequisite/readiness issue and stable base networking and storage.
gitchegumi added this to the UDS install milestone 2026-08-29 01:58:39 -05:00
Author
Collaborator

Deployment kickoff

Issue #9 passed its readiness gate and is closed. UDS Core 1.11.1 deployment can begin.

Pinned inputs

  • UDS Core 1.11.1
  • UDS CLI 0.35.1
  • RKE2 / Kubernetes 1.35
  • UDS-managed tenant and admin Istio gateways allocated by MetalLB
  • truenas-iscsi as the default StorageClass; truenas-nfs only when ReadWriteMany is required
  • External S3-compatible storage for Loki and Velero
  • External PostgreSQL 18 for Keycloak
  • Runtime secrets sourced from ~/.config/uds/secrets/ and never committed

Reusable configuration and documentation will use only S3_ENDPOINT, LOKI_CHUNKS_BUCKET, LOKI_ADMIN_BUCKET, VELERO_BUCKET, KEYCLOAK_DB_HOST, KEYCLOAK_DB_NAME, KEYCLOAK_DB_USER, tenant/admin domain placeholders, and secret references.

Deployment phases

  1. Build and review a sanitized UDS Core configuration.
  2. Materialize private runtime values from the external secret files.
  3. obtain the pinned bundle and confirm its metadata.
  4. Deploy while monitoring package and Kubernetes health.
  5. Validate gateways, TLS, storage, Keycloak, Loki, Velero, and workload recovery behavior.

Deployment will proceed one verified step at a time.

## Deployment kickoff Issue #9 passed its readiness gate and is closed. UDS Core 1.11.1 deployment can begin. ### Pinned inputs - UDS Core 1.11.1 - UDS CLI 0.35.1 - RKE2 / Kubernetes 1.35 - UDS-managed tenant and admin Istio gateways allocated by MetalLB - truenas-iscsi as the default StorageClass; truenas-nfs only when ReadWriteMany is required - External S3-compatible storage for Loki and Velero - External PostgreSQL 18 for Keycloak - Runtime secrets sourced from ~/.config/uds/secrets/ and never committed Reusable configuration and documentation will use only S3_ENDPOINT, LOKI_CHUNKS_BUCKET, LOKI_ADMIN_BUCKET, VELERO_BUCKET, KEYCLOAK_DB_HOST, KEYCLOAK_DB_NAME, KEYCLOAK_DB_USER, tenant/admin domain placeholders, and secret references. ### Deployment phases 1. Build and review a sanitized UDS Core configuration. 2. Materialize private runtime values from the external secret files. 3. obtain the pinned bundle and confirm its metadata. 4. Deploy while monitoring package and Kubernetes health. 5. Validate gateways, TLS, storage, Keycloak, Loki, Velero, and workload recovery behavior. Deployment will proceed one verified step at a time.
Owner

Restart and recovery testing — PASS. Representative Keycloak, Loki, and Velero pods were manually deleted. Kubernetes automatically recreated all workloads without intervention. Keycloak recovered and returned HTTP 200 through the external Istio route; Loki restored its expected replica count; Velero recovered with its default BackupStorageLocation reporting Available. All cluster nodes remained Ready and no unexpected non-Running workloads remained after testing.

Restart and recovery testing — PASS. Representative Keycloak, Loki, and Velero pods were manually deleted. Kubernetes automatically recreated all workloads without intervention. Keycloak recovered and returned HTTP 200 through the external Istio route; Loki restored its expected replica count; Velero recovered with its default BackupStorageLocation reporting Available. All cluster nodes remained Ready and no unexpected non-Running workloads remained after testing.
Owner

Closing as complete

Closing as complete
Sign in to join this conversation.
No description provided.