Automating cascading power management procedures for VMWARE VSPHERE servers using network infrastructure

Abstract

The article investigates the current problem of ensuring the stability of information systems in the conditions of an energy crisis. The need to protect VMware vSphere virtualized environments is substantiated, where sudden power outages without prior controlled shutdown lead to irreversible damage to virtual disk file systems, violation of database integrity and metadata at the storage level.

Analysis of existing solutions has shown that industrial uninterruptible power supply systems (UPS) with network management cards (NMC) have a high cost and limited software flexibility, which makes their use economically inefficient for small and medium-sized cluster segments.

The aim of the work is to develop a flexible automated cascade power management system based on available network infrastructure, in particular MikroTik routers, and high-power household charging stations. Within the framework of the study, a classification of virtual resources was carried out and it was proposed to distribute all resources into several service levels, for example: minimal (vCenter, AD servers, Web server), basic (file servers, RDP) and full (VMware Horizon, UAG). The authors described the main stages of creating the relevant system: distribution of the virtual environment, study of electrical power consumption, configuration of the virtual environment, as well as development of procedures (scripts) for controlling the start/shutdown of the VMware vSphere virtual environment. Based on experimental data, time delays for cascading start and shutdown are proposed, which ensures correct synchronization of system components. The technical implementation of the system is described, which includes setting the server BIOS for automatic power recovery and supporting Wake-on-LAN. A scheme for distributing power consumption through two power supplies is proposed, which allows optimizing charging of stations without stopping the servers. Also considered are methods for optimizing the VMware Horizon environment through the ParentVMs parameter control to reduce the number of active service virtual machines and the features of developing start/stop control procedures using the example of a MikroTik router.

The practical significance of the results tested in the laboratory "IT Solutions Center" of Taras Shevchenko National University of Lviv is the creation of a budget and reliable infrastructure protection tool that simulates the functionality of corporate NMC systems. The use of the SSH protocol with certificates for command automation on ESXi hosts provides a high level of autonomy and security of the management process without the need for manual administrator intervention.