Centralized Workflow System for HCI Cluster Configuration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Configuring a hyper-converged infrastructure (HCI) cluster is a tedious and time-consuming process due to the need for serial operations and extensive manual input across multiple nodes, often resulting in inefficiency and potential inconsistencies in network configurations.
Innovation Solution
Implementing a centralized workflow system that allows multiple configuration clients to invoke and manage batch-configuring operations through a cluster manager, which automates and simplifies the configuration of multiple cluster nodes by storing and managing workflow sessions in a centralized database, enabling parallel execution of workflow operations and reducing repetitive tasks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual configuration operations are performed on each cluster node individually, then configuration flexibility is maintained, but configuration time and operational complexity increase significantly
Solution Approach 1:
The configuration process is segmented into discrete workflow operations that can be independently defined, executed, and managed. Each workflow operation represents a specific configuration task that can be applied to one or more cluster nodes, allowing the overall configuration process to be broken down into manageable, reusable units that reduce operational complexity and time
Solution Approach 2:
The system transforms the configuration approach by changing the operational parameters from individual node manipulation to batch workflow execution. Configuration parameters are standardized into workflow templates that can be applied across multiple nodes simultaneously, fundamentally altering how configuration operations are performed and reducing the time required
2Reliability
If configuration operations are performed serially on each node, then configuration consistency is maintained, but productivity and efficiency decrease
Solution Approach 1:
Configuration workflows are defined and validated in advance before execution. Workflow templates capture the required configuration steps, parameters, and validation rules beforehand, ensuring that when workflows are executed on multiple nodes, consistency is maintained through pre-established guidelines rather than ad-hoc serial operations
Solution Approach 2:
A centralized workflow management system acts as an intermediary between the administrator and the cluster nodes. This intermediary orchestrates the execution of workflows across multiple nodes, ensuring consistent application of configuration operations while enabling parallel execution to improve productivity without sacrificing reliability
3Adaptability or versatility
If multiple administrators configure the cluster independently, then operational flexibility is improved, but configuration inconsistencies and errors increase
Solution Approach 1:
The workflow system provides universal templates that can be used by multiple administrators across different operations. These standardized workflows serve multiple functions: they guide administrators through proper configuration steps, ensure consistency across different administrators' work, and maintain accuracy while still allowing flexibility in when and how workflows are invoked
Solution Approach 2:
The system implements feedback mechanisms where workflow execution status, results, and any errors are tracked and communicated back to administrators. This feedback loop ensures that multiple administrators working independently can see the state of configuration operations, validate their work against established workflows, and maintain configuration accuracy through continuous monitoring and validation
Data Source
AI summary
Example methods are provided for configuring a hyper-converged infrastructure (HCI) cluster managed by a cluster manager. The method may comprise retrieving, by a workflow session from the cluster manager. The workflow session may include a plurality of workflow operations, and a first workflow operation selected from the plurality of workflow operations is marked as incomplete. The method may transmit o the cluster manager a request to invoke the first workflow operation. The cluster manager may be configured to perform the first workflow operation by batch-configuring a plurality of nodes in the HCI cluster. In response to a determination that the first workflow operation is completed, the method may request the cluster manager to mark the first work operation in the workflow session as completed.