This package is designed to deploy DUBBD to environments leveraging RKE2 as the Kubernetes distribution.
Warning This package is a WIP and not configured for production deployments yet. Notable items:
- You will need to satisfy dependencies of the cluster + storage class
- Loki will run in a single-pod setup with PVC backed log storage
- Common prerequisites
- RKE2 cluster -- K8s v1.27+
- Local K8s context is pointing to the above cluster
- RKE2 cluster has Zarf init package deployed (with
git-server
component)zarf init --components=git-server --confirm
- Existing S3 compatible object storage bucket for Velero backups with a credentials secret
The recommended way to configure DUBBD-RKE2 is via a zarf-config.yaml
file located in the same directory that you will be performing the deploy. The available zarf-config.yaml
configurations are shown below.
Note the keys that are not commented out are required to deploy DUBBD-RKE2.
package:
deploy:
set:
# -- Domain name for the cluster
domain: bigbang.dev
# -- TLS key
key_file: bigbang.dev.key
# -- TLS cert
cert_file: bigbang.dev.cert
# -- If set to true, force delete all resources on removal (i.e. object storage, PVCs, etc)
ephemeral: false
Velero is included by default. This requires that the cluster have access to an object storage provider (such as AWS S3, Minio, Nutanix or others). The default configuration used in CI deploys an instance of Minio and uses it to perform a simple cluster backup on a schedule. This default configuration can be adopted by setting the exposed velero bucket values, but in all likelyhood it will make more sense to provide a custom velero config at deploy time that overrides the udsbackup
schedule.