Single-Node Orchestration Updates Using a Temporary Control Plane Copy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Single node deployments in orchestration platforms face challenges with disruptive updates, as they cannot perform rolling restarts and result in cluster headlessness and disruptions during planned outages, leading to increased costs and availability issues.
Innovation Solution
A method is implemented where a processor set generates a copy of the control plane node, allocates workloads to this copy, applies the disruptive update, and then reallocates workloads back to the original node upon completion, using network interfaces and logical partitions to ensure continuous operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single node deployment is used in an orchestration platform, then device complexity is reduced and cost is lowered, but the system cannot perform rolling restarts and experiences disruptions during disruptive updates
Solution Approach 1:
The system segments the single node into two functional parts: the original control plane node and a dynamically created copy. This segmentation allows the node to be updated without complete downtime by separating update operations from workload execution, resolving the contradiction between simplified deployment and update availability.
Solution Approach 2:
The system creates a copy of the control plane node to handle workloads during disruptive updates. This copy mechanism enables the original node to be updated while maintaining service continuity through the replica, directly addressing the availability issue during updates while preserving the single-node deployment simplicity.
2Reliability
If workloads are migrated to a copy during update, then availability is maintained, but device complexity and resource overhead increase
Solution Approach 1:
The system dynamically creates and manages the node copy only when needed for disruptive updates, rather than maintaining it permanently. This dynamic approach maintains availability during updates while minimizing device complexity and resource overhead during normal operation, as the copy mechanism is activated only when required.
Solution Approach 2:
The system performs preliminary actions by creating the node copy before the disruptive update begins, allowing seamless workload migration. This preliminary preparation maintains availability while managing complexity through automated orchestration of the copy creation and workload transfer process.
3Reliability
If a copy is created for disruptive updates, then continuous operation is maintained, but resource consumption and overhead increase
Solution Approach 1:
The node copy is dynamically provisioned only during disruptive update operations and deactivated afterward, allowing continuous operation during updates while minimizing resource consumption during normal operation. This dynamic resource allocation resolves the contradiction between maintaining continuous operation and reducing resource overhead.
Solution Approach 2:
The system discards the temporary node copy after the disruptive update is complete and recovers the computational resources. This approach maintains continuous operation during the necessary update window while minimizing long-term resource consumption and overhead.
Data Source
AI summary
A computer-implemented method for updating a node in an orchestration platform. A processor set receives a patch of a disruptive update for the node in the orchestration platform. The node is a control plane node that is configured to run workloads. The processor set generates a copy for the node in the orchestration platform. The processor set allocates the workloads for the node to the copy for the node. The processor set updates the node using the patch of the disruptive update. The workloads are running on the copy for the node when the node is updating. The processor set allocates the workloads from the copy for the node back to the node upon finishing the disruptive update.