HCI Host Scoring for Faster Maintenance Mode Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The time required for maintenance mode in hyper-converged infrastructure (HCI) clusters is prolonged due to the need to shut down and migrate virtual machines (VMs) and resynchronize data, which increases downtime and operational inefficiencies.
Innovation Solution
A method is employed to determine a score for each host in the HCI cluster based on working memory and persistent storage sizes of active VMs, select a host for maintenance using these scores, and migrate VMs to optimize the maintenance operation, thereby reducing the overall time required for entering maintenance mode.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all hosts in the HCI cluster undergo maintenance mode simultaneously, then maintenance operations can be performed on all hosts, but the overall maintenance time and downtime increase significantly
Solution Approach 1:
The patent divides the cluster of hosts into multiple groups and performs maintenance operations on hosts in sequential batches rather than simultaneously. This segmentation allows maintenance to be performed on a subset of hosts at a time, reducing the overall maintenance window and minimizing impact on the cluster while still completing maintenance on all hosts.
Solution Approach 2:
The patent calculates and determines a score for each host based on VM workload characteristics before initiating maintenance operations. This preliminary scoring and selection process enables the system to identify which hosts should be prioritized for maintenance first, allowing for optimized maintenance sequencing that reduces total maintenance time while ensuring all hosts receive necessary maintenance.
2Reliability
If virtual machines are shut down on hosts entering maintenance mode, then maintenance operations can be performed, but productivity and service availability decrease
Solution Approach 1:
The patent segments the cluster into multiple host groups and performs maintenance on only one group at a time. This allows VMs on other groups to continue running uninterrupted, maintaining productivity while still completing maintenance operations on the selected hosts. The segmentation ensures that not all VMs need to be shut down simultaneously.
Solution Approach 2:
The patent performs preliminary scoring of hosts based on their VM workload characteristics before maintenance. By selecting hosts with lower scores (indicating less critical or more suitable candidates for maintenance), the system minimizes the impact on ongoing VM workloads while still achieving necessary maintenance objectives.
3Reliability
If data resynchronization is performed after host exits maintenance mode, then data consistency is restored, but additional time and operational complexity are required
Solution Approach 1:
The patent performs preliminary actions during the host selection and scoring phase to prepare for efficient maintenance execution. By pre-calculating scores based on VM characteristics and selecting optimal hosts beforehand, the system reduces the time and complexity required for subsequent data resynchronization operations after maintenance mode exits.
Data Source
AI summary
An information handling system may include at least one processor and a memory. The information handling system may be configured to perform a maintenance operation involving a plurality of hosts of an information handling system cluster by: determining a score for each host based on a sum of working memory sizes for all active virtual machines executing on such host plus a sum of persistent storage sizes for all virtual machines stored on such host; based on the determined scores, selecting a first host for upgrading; migrating at least a portion of all virtual machines stored on the first host from the first host to one or more other hosts; and causing the first host to perform the maintenance operation.