How high is high? If you are considering climbing, then Everest is pretty high after all at 8848M above sea level. Although without the right equipment, team and preparation, trying to climb just 2M can be impossible. But ‘high’ is used in other contexts as well. Like when you are trying to keep a business running these days. If you are then it’s likely the high you may be thinking about is High Availability. Without the right approach, tools and infrastructure you may be trying to solve a problem that can seem to be the same scale as Everest.
With business becoming more global, and being more responsive to events, and with mobile or web traffic coming direct from partners, customers or suppliers, downtime has to be avoided. How do you keep your systems up, your applications running and your data available all the time? Even when, inevitably, there are failures?
IBM MQ is a critical part of your business connectivity. It provides a reliable, secure, scalable and robust middleware layer connecting applications, systems and services and exchanging data between them. Making use of IBM MQ ensures your applications can be simpler and more agile, yet more reliable, and also easier to shift between deployment environments. Your applications will rely on IBM MQ persisting their messages, ensuring that messages are never lost. How do you reap these rewards of simpler applications unless the MQ middleware is highly available to ensure the applications can keep running?
Having been around for 25 years, IBM MQ understands this need very well. As such it provides a variety of ways to configure and manage High Availability. And the most recent innovation, based on the High Availability approach used in the MQ Appliance is designed to not only offer extremely robust and effective high availability, but at the same time ensuring it is simple to set up and maintain, without additional external complexity: Replicated Data Queue Managers.
Many clients were facing the same set of problems: they didn’t like the costs and complexity of providing and maintaining network attached storage, which was a common way of providing high availability for MQ. The request was high availability that was more self-contained, without external dependencies. A way to deploy MQ in highly available configurations without the requirement for an environment that needs lots of setup, with highly skilled resources and additional costs.
With IBM MQ V9.1, our new Long-Term Support release, customers can now take advantage of Replicated Data Queue Managers, which offer a 3-node configuration, making use of replicated local storage, which make the MQ messages available on each of 3 MQ systems, instead of relying on a single copy of data on network storage.
Instead of requiring lots of setup, and ongoing extensive maintenance, MQ itself will do almost all the setup during the initial MQ install. Then, when you are creating a Queue Manager, you simply request it as a RDQM resource, and that’s pretty much all that’s needed. And it’s not just simple in the configuration of the Queue Managers. As it supports Floating IP, when one Queue Manager fails, and another instance automatically starts up on one of the other 2 nodes, the original Queue Manager IP address will move with it, meaning the applications are essentially unaware of the move, and the workload is uninterrupted as the messages and logs had been kept up to date synchronously on all 3 systems.
With an additional option allowing for manual startup in a replicated pair of systems by choosing either synchronous or asynchronous replication to provide Disaster Recovery configurations, this new approach to HA really goes a long way to make it much simpler to reach the highest peaks of high availability.
There are already a few places to look for more information on this exciting new development. There is a technical blog entry by John Colgrave, along with a GitHub community, and of course the Knowledge Center.
Suitable for customers on RedHat Linux on the x86 platform, you need MQ Advanced licensing on just one system node, and MQ Advanced High Availability Replica licensing on the other 2 nodes. Also, this can’t be used with container deployments – but virtual machine images or bare metal is fine. With RDQM now part of the Long-Term Support release of MQ V9.1, you can scale the highest peaks of availability. You are not starting at base camp. You are already close to the summit. Let’s get to the top.