Skip to content

Claudie v0.2

Due to a breaking change in the input manifest schema, the v0.2.x will not be backwards compatible with v0.1.x.

Deployment

To deploy the Claudie v0.2.X, please:

  1. Download the archive and checksums from the release page

  2. Verify the archive with the sha256 (optional)

    sha256sum -c --ignore-missing checksums.txt

    If valid, output is, depending on the archive downloaded

    claudie.tar.gz: OK

    or

    claudie.zip: OK

    or both.

  3. Lastly, unpack the archive and deploy using kubectl

    We strongly recommend changing the default credentials for MongoDB, MinIO and DynamoDB before you deploy it. To do this, change contents of the files in mongo/secrets, minio/secrets and dynamo/secrets respectively.

    kubectl apply -k .

v0.2.0

Features

  • Unify the naming schema in the input manifest #601
  • Deploy MinIO in multi-replica fashion #589

Bugfixes

No bugfixes since the last release.

Known issues

  • Workflow fails to build when a user makes multiple changes of the input manifest, regarding the API endpoint #606
  • Longhorn pod longhorn-admission-webhook stuck in Init state #598
  • Deletion of config fails if builder crashes after deleting nodes #588

v0.2.1

Features

  • Improve management of Longhorn volume replicas #648
  • Improve logging on all services #657

Bugfixes

  • Fix unnecessary restarts in Wireguard playbook #658

Known issues

  • Certain change in nodepool configuration forces replacement of VMs #647

v0.2.2

Features

  • Cluster Autoscaler integration #644
  • Drop using grpc-health-probe in favour of Kubernetes native grpc health probes. #691
  • Centralized information about the current workflow state of a cluster in the frontend service #605

Bugfixes

  • Certain change in nodepool configuration forces replacement of VMs #647