You can create a high availability group from a pair of IBM® MQ Appliances to give your IBM MQ system a degree of resilience.
See High availability for an overview of the high availability solution.
When you plan a high availability implementation, consider the following points:
Appliances:
A high availability group requires two IBM MQ Appliances.
Both appliances should be running the same level of appliance firmware. (Appliances can operate
at different levels to allow time to upgrade the appliances separately, but you should avoid
configuring HA queue managers, including adding or removing queue managers, during this
period.)
You can run queue managers on both appliances (there is no concept of an 'active' and a
'standby' appliance).
An appliance can run high availability queue managers or disaster recovery queue managers, and
queue managers that do not belong to either group. You can also run high availability queue managers
that belong to disaster recovery configurations.
An appliance can belong to only one HA group.
You can create the HA group from either of the appliances.
The date and time settings must be sychronized between the two appliances. You can achieve this
by configuring both appliances to use the same NTP server (see Configuring the locale, date, and time).
For optimal performance, use the same model appliance for both members in the HA
pair (for example, two M2003 models). It is possible to pair older and newer appliances, but the
pair will perform to the constraints of the older model.
Queue managers:
You specify that a queue manager belongs to an HA group when you create the queue manager.
You can create an HA queue manager on either of the two appliances in the HA group.
The appliance that you create the queue manager on is the preferred appliance for that queue
manager. The queue manager runs on its preferred appliance so long as that appliance is
available.
A queue manager that belongs to a high availability group can also be set to belong to a
disaster recovery configuration. The disaster recovery configuration can itself be an HA pair.
You can specify a floating IP address for individual HA queue managers. Applications can use the
floating IP address to connect to a queue manager regardless of which appliance it is actually
running on. You specify an interface name to connect on when you create the floating IP address (for
example eth22). That interface must be a physical interface
configured with a static IP address on both appliances. You can also specify an aggregate interface
that exists with the same name on each appliance. The name of the aggregate interface must obey
certain rules to be used for this purpose, see link-aggregation.
You can specify that the HA replication link is secured by using TLS.
Physical configuration:
The appliances in the HA group synchronize by replicating queue manager data across a 10 Gb
Ethernet link. Use the eth21 interface for the replication link (advanced users
can, if required, use one of the appliance's other links, or an aggregated link, for
replication).
The appliances in the HA group are also connected by a primary and a secondary interface that
both use 1 Gb Ethernet links. These interfaces are used to monitor the presence of the other
appliance in the group and detect a failover situation. Use the eth13 interface for the primary
link, and the eth17 interface for the secondary link.
For the best performance of synchronization and failover, the two appliances need to be as
physically close as possible, ensuring short connecting cable lengths. However, the two appliances
should not be in the same rack, in case the rack fails.
If you connect the two appliances by using a network switch, you should use a separate switch
for each of the three connections.
![[MQ 10 2026]](../ng10cd.gif)
You can specify that the HA replication link is secured by using TLS. 