Recommendations for Implementing Geographic Logical Volume Manager (GLVM) On-Premises and on the Cloud

Abstract

This IBM Redpaper introduces the AIX Geographic Logical Volume Manage and examines DR options for AIX, both on premises and in the Cloud. It also offers recommendations designed to ensure a smooth implementation and covers good practices around ongoing maintenance and monitoring. The main driver for writing this document was to address the varied experiences we have seen with the implementation of GLVM both standalone and when managed by PowerHA. While many implementations go smoothly and the customer experiences very few issues in production there are others that have been, shall we say, far from smooth. The aim is to document what has lead to a successful GLVM implementation and to provide warnings for where "there be dragons." Most of the recommendations will generally apply to any implementation of GLVM, but we will highlight the specific requirements dictated by PowerVS. It is also vital that the OS and applications teams understand that they will be working in a clustered environment, so practices and processes particularly change control must be modified to reflect this.

At the time of writing this document Global Replication Service (GRS), which provides disk mirroring capability within IBM Power Virtual Server (PowerVS) was only available between a limited number of Data Centers and its support by IBM PowerHA SystemMirror Enterprise Edition (PowerHA) had not been announced. Hence we are focusing only on IBM AIX Geographic Logical Volume Manager (GLVM) which provides data mirroring capabilities between AIX systems. We will cover using GLVM as a tool to migrate from on-premises systems to PowerVS and as a DR solution. The DR solution could be solely in the cloud, or between on-premises and the cloud.

It is assumed that the reader is familiar with the IBM Redpaper by Ravi Shankar et al. Asynchronous Geographic Logical Volume Mirroring Best Practices for Cloud Deployment, REDP-5665..