OpenStack Control Node Upgrade with Offline VM Replacement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cloud computing infrastructure upgrade methods, such as hot migration and disaster recovery switching, often result in service interruptions and pose risks to network-sensitive applications, failing to ensure complete service continuity.
Innovation Solution
An OpenStack-based cloud computing platform upgrading method that involves loading an upgraded control node, creating a second control node virtual machine offline, and replacing the first control node virtual machine with the second when complete, ensuring minimal service interruption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If offline upgrade method is used, then upgrade speed is fast, but service interruption time is prolonged
Solution Approach 1:
The patent applies preliminary action by creating a new control node virtual machine with the upgraded version before replacing the old one. The upgraded control node virtual machine is prepared in advance on the same host, and only after its creation is confirmed does the system proceed with replacement. This ensures that the upgrade process can start immediately without waiting for complex migration operations, thus reducing service interruption time while maintaining fast upgrade speed.
Solution Approach 2:
The patent uses copying by creating a new control node virtual machine that replicates the functionality of the old one but with the upgraded version. The new virtual machine is created as a copy of the existing control node configuration, allowing seamless replacement. This copying approach enables the system to maintain service continuity during upgrade by having a ready-to-takeover replica, effectively resolving the contradiction between fast upgrade and minimal interruption.
2Reliability
If online upgrade method with hot migration is used, then service continuity is maintained, but upgrade duration is relatively long
Solution Approach 1:
The patent applies the taking out principle by extracting the control node virtual machine from the running system and creating a new one independently on the same host. The old control node virtual machine is stopped and removed after the new one is created and verified. This extraction approach allows the upgrade to happen without requiring complex hot migration operations, thus reducing upgrade duration while maintaining service continuity through the brief replacement window.
Solution Approach 2:
The patent uses skipping by bypassing the complex hot migration process and directly creating the upgraded control node virtual machine on the same host. Instead of gradually migrating services during upgrade, the system rushes through the replacement process by creating the new virtual machine first and then immediately replacing the old one. This skipping approach maintains service continuity by minimizing the time the system is in transition, effectively resolving the contradiction between reliability and upgrade duration.
3Reliability
If disaster recovery switching is used, then service can be restored, but upgrade operations become complex with high manpower requirements
Solution Approach 1:
The patent applies self-service by implementing an automated control node replacement process that requires minimal manual intervention. The system automatically creates the new control node virtual machine, stops the old one, and performs the replacement without requiring complex manual disaster recovery procedures. This self-service approach maintains service restoration capability while significantly reducing upgrade operation complexity and manpower requirements, as the process is streamlined through automated virtual machine management.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
The present application relates to the technical field of cloud computing, and provides an OpenStack-based cloud computing platform upgrading method and a related device. The method comprises: when a control node runs a first control node virtual machine, loading an upgraded version of the control node; on a host where the control node is located, offline creating a second control node virtual machine corresponding to the upgraded version; and when the creation of the second control node virtual machine is completed, replacing the first control node virtual machine with the second control node virtual machine.