Hyperconverged Cluster Upgrade Parallelization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lifecycle management (LCM) methods in hyperconverged infrastructure (HCI) clusters are limited by the need to upgrade hosts sequentially due to failure tolerance constraints, resulting in prolonged upgrade times, especially in large clusters where dozens of hosts require upgrading without interrupting business services.
Innovation Solution
The method involves identifying the minimum number of hosts required to satisfy operational requirements, removing excess hosts from the cluster, upgrading them in parallel, and re-adding them once upgraded, thereby reducing the overall upgrade time without causing data loss or service disruption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If hosts are upgraded sequentially to satisfy FTT constraints, then reliability is maintained, but productivity deteriorates due to prolonged upgrade times
Solution Approach 1:
The cluster is segmented into two groups: active hosts that remain operational and hosts being upgraded. This segmentation allows parallel upgrading of multiple hosts while maintaining a sufficient number of active hosts to satisfy FTT constraints, thus improving upgrade speed without compromising reliability
Solution Approach 2:
The system dynamically adjusts the number of active hosts during the upgrade process. Hosts are temporarily removed from the active cluster, upgraded in parallel, and then re-added. This dynamic adjustment allows the system to maintain reliability thresholds while maximizing parallel upgrade capacity
2Productivity
If more hosts are upgraded in parallel, then productivity improves, but reliability deteriorates due to increased risk of data loss
Solution Approach 1:
The system continuously monitors the number of active hosts and compares it against the minimum required to satisfy FTT constraints. This feedback mechanism ensures that parallel upgrading proceeds only when reliability thresholds are maintained, preventing data loss while maximizing upgrade speed
Solution Approach 2:
The system changes the operational state parameter of hosts during upgrading, temporarily removing them from the active cluster. This parameter change allows parallel upgrading of multiple hosts while maintaining the necessary number of active hosts to satisfy reliability constraints
3Reliability
If sequential upgrading is performed, then reliability is maintained, but loss of time increases significantly
Solution Approach 1:
The system performs preliminary checks to determine the minimum number of hosts required to satisfy FTT constraints before initiating parallel upgrades. This preliminary action enables the system to safely upgrade multiple hosts in parallel while ensuring service continuity, thus reducing upgrade duration without compromising reliability
Data Source
AI summary
An information handling system may include at least one processor; and a non-transitory memory coupled to the at least one processor. The information handling system may be configured to manage an upgrade of a cluster of N host systems by: receiving an indication of a number M of host systems of the cluster that are needed to satisfy an operational requirement of the cluster; removing N-M of the host systems from the cluster; causing the removed host systems to be upgraded in parallel; and adding at least one of the removed host systems back to the cluster after the at least one of the removed host systems has completed the upgrade.