-
Feature
-
Resolution: Done
-
Critical
-
None
-
Strategic Product Work
-
False
-
-
False
-
OCPSTRAT-1408Comprehensive Automation of HCP Backup and Restore Processes for Enhanced Scalability and Reliability
-
100% To Do, 0% In Progress, 0% Done
-
7
-
0
-
Program Call
Feature Overview (aka. Goal Summary)
Enable the recreation of an HCP Agent provider cluster without triggering the reprovisioning of worker nodes, ensuring minimal downtime and operational efficiency. This feature introduces a mechanism to prevent changes or removal of associated resources during the deletion process.
This is crucial to unblocking backup and restore flows. Without this feature, customers would have to carry the burden of backing up not only the control planes and the management cluster but also the workloads, which sometimes might be infeasible!
Goals (aka. expected user outcomes)
- Seamless Recreation: Users can delete and recreate an Agent provider cluster without affecting the worker nodes. This will require changes in CAPI-agent provider.
- Efficient Restoration: Restored Agent provider cluster and AgentMachine resources continue to manage the agents and related resources seamlessly.
Deployment Considerations
Scenario | Needs |
Self-managed, managed, or both | Self-managed |
Classic (standalone cluster) | N/A |
Hosted control planes | Applicable |
Multi node, Compact (three node), or Single node (SNO), or all | All |
Connected / Restricted Network | Applicable |
Architectures, e.g. x86_x64, ARM (aarch64), IBM Power (ppc64le), and IBM Z (s390x) | Applicable |
Operator compatibility | Must ensure compatibility |
Backport needed (list applicable versions) | Applicable versions to be listed |
UI need (e.g. OpenShift Console, dynamic plugin, OCM) | UI adjustments may be required |
Other (please specify) | N/A |
Use Cases (Optional)
- Customer backs up the control planes using OADP and process outlined here: https://issues.redhat.com/browse/HOSTEDCP-1484 and expects the cluster to continue to be functional.
- An administrator deletes an Agent provider cluster, and the associated resources are marked with annotations to prevent their removal. The administrator then recreates the Agent provider cluster, which manages the same worker nodes without requiring their reprovisioning.
- Alternative Flow: An administrator attempts to recreate an Agent provider cluster without the correct annotations, leading to a fallback mechanism or alert.
Background
The current process for recreating an Agent provider cluster involves reprovisioning worker nodes, which is inefficient and leads to unnecessary downtime. This feature aims to streamline this process by allowing the recreation of Agent provider clusters without affecting the worker nodes.
This will make it feasible for customers to backup and restore their hosted control-planes without the additional burden.
Customer Considerations
- Ease of Use: The feature must be easy for administrators to understand and use.
- Documentation: Clear guidance on how to use the annotations and recreate Agent provider clusters without reprovisioning nodes.
- is related to
-
OCPSTRAT-989 Backup/restore for Hosted Clusters for Self-Managed HCP Part I
- Closed
- links to
- mentioned in
-
Page Loading...