Mirroring is an increasingly important storage management technique as storage volumes grow and backup windows shrink. In mirroring, data is written to more than one logical volume simultaneously, for increased read performance and/or backup. In modern systems the data is often written to three or more logical volumes so that one volume can be split off for backup while still maintaining a redundant active copy.
Veritas points out that mirrored systems are most vulnerable to failure between the time the volume is broken off and the time it is completely resynchronized with the other volumes. To minimize this vulnerability, and to improve performance, the company recommends scheduling resynchronization of the broken mirror during periods of low I/O activity. Resynchronization and regeneration are I/O intensive operations and doing them during periods of high activity not only can degrade performance for users, it takes longer and that stretches the window of vulnerability.
The company also recommends mirroring critical files to more than three volumes for added redundancy.
Veritas has a white paper on best practices for managing Windows volumes online at its website at http://www.veritas.com/us/products/whitepapers.html.
Rick Cook has been writing about mass storage since the days when the term meant an 80K floppy disk. The computers he learned on used ferrite cores and magnetic drums. For the last twenty years he has been a freelance writer specializing in storage and other computer issues.
Did you like this tip? Whether you did or not, why not let us know. Drop us an email and sound off.
Data Replication: Tools and Techniques for Managing Distributed Information
Author : Marie Buretta
Publisher : John Wiley & Sons
Published : Feb 1997
Today's distributed databases increasingly require the creation and management of multiple copies of the same data. Successfully managing this data is the complex task explored here. After an overview of data replication terms and technologies, this book reviews approaches to data replication offered by new software tools from IBM, Sybase and others. There are alternative design strategies for enterprise-wide databases that can handle replicated data. And finally work plans provide detailed guidelines on selecting the right mix of replication tools and verifying that they work successfully.