Shared file system
On Mulitplatforms, a multi-instance queue manager uses a networked file system to manage queue manager instances.
A multi-instance queue manager automates failover using a combination of file system locks and shared queue manager data and logs. Only one instance of a queue manager can have exclusive access to the shared queue manager data and logs. When it gets access it becomes the active instance. The other instance that does not succeed in getting exclusive access waits as a standby instance until the queue manager data and logs become available.
The networked file system is responsible for releasing the locks it holds for the active queue manager instance. If the active instance fails in some way, the networked file system releases the locks it is holding for the active instance. As soon as the exclusive lock is released, a standby queue manager waiting for the lock attempts to acquire it. If it succeeds, it becomes the active instance and has exclusive access to the queue manager data and logs on the shared file system. It then continues to start.
The related topic, Plan file system support describes how to set up and check that your file system supports multi-instance queue managers.
A multi-instance queue manager does not protect you against a failure in the file system. There are a number of ways to protect your data.- Invest in reliable storage, such as redundant disk arrays (RAID), and include them in a networked file system that has network resilience.
- Back up IBM MQ linear logs to alternative media, and if your primary log media fails, recover using the logs on the alternative media. We can use a backup queue manager to administer this process.
Parent topic: Multi-instance queue managers