AIX > Administrator > High Availability

Replicating Success

Most IT managers know the fear. The datacenter is running smoothly and delivering business-critical applications when suddenly a natural disaster or catastrophic event strikes--perhaps the massive Northeastern U.S. blackout of August 2003 or a Florida hurricane--causing the entire complex to shut down. The staff is unharmed but the systems are out of commission, and it'll be days before the datacenter--with all of its critical applications and data--can be brought back online. This is a bad time to consider what you could've done to prepare for such a disaster.


If your applications are running under the AIX* 5L OS on IBM* eServer p5, IBM eServer i5 or pSeries* servers, you could've configured IBM's High Availability Cluster Multi-Processing (HACMP*)/XD new Geographic Logical Volume Manager (GLVM) to mirror all of your business-critical data to a remote site in real time and immediately restart your applications automatically at that remote site for high availability (HA) of your critical business systems.

HACMP has been IBM's premier HA solution for more than 12 years with more than 12,000 customers. The XD option provides several technologies to extend HACMP's failover capabilities to remote sites through remote data mirroring and automatic takeover of remote disks.

HACMP/XD GLVM is a new HA function that can mirror data across a standard IP network of potentially unlimited distance and provide automated failover and fallback of applications using this mirrored data. GLVM performs this remote mirroring of AIX logical volumes using AIX's native Logical Volume Manager (LVM) functions for optimal performance and ease of configuration and maintenance.

The new HACMP/XD GLVM significantly improves IBM's pSeries extended distance Disaster Recovery and Business Continuity solution, helping position it for anticipated growth in geographic disaster recovery.

What it Provides and How it Works

HACMP/XD GLVM provides two essential functions: remote data mirroring and remote failover and fallback. Together these functions provide HA support for applications and data across a standard TCP/IP network to a remote site.

Figure 1 shows how HACMP/XD GLVM's remote data-mirroring function integrates directly with AIX's LVM, extending it to support data mirroring to remote sites. The HACMP cluster consists of two nodes connected by a TCP/IP WAN: Node A is in the Tucson datacenter and Node B is in the St. Louis datacenter. Node A is the primary node for an HACMP resource group containing a logical volume with two copies, one local (on Node A) and one remote (on Node B). Node B serves as the backup node for this resource group. To the LVM on Node A, both copies of the data appear local, but the second copy is represented by a pseudo-device called a Remote Physical Volume (RPV). This RPV has two parts: a client operating on the local node and a server operating on the remote node. When data is written to the geographically mirrored logical volume on Node A, the local RPV client receives write requests as if it were a standard physical volume but then transfers those requests to the RPV server on the remote Node B, which writes the data to the physical volume residing on that remote node. In this way, a complete copy of the Tucson logical volume's data is maintained on the remote St. Louis node. (Note: To maximize performance, read requests on Node A are generally directed by the LVM to a local copy of the data to avoid the network latency associated with a remote-read operation.)

Andrew Naiberg is a product-marketing manager with IBM. Andrew can be reached at

Skip Russell is a senior software engineer with IBM. Skip can be reached at

comments powered by Disqus



2019 Solutions Edition

A Comprehensive Online Buyer's Guide to Solutions, Services and Education.

IBM Systems Magazine Subscribe Box Read Now Link Subscribe Now Link iPad App Google Play Store
IBMi News Sign Up Today! Past News Letters