Q Apply parameters
The two types of Q Apply operating parameters are replication queue map parameters and Q Apply instance-level parameters.
- Replication queue map parameters
- These parameters affect the way a Q Apply program processes messages
on a single receive queue, and are most important for performance
tuning. These parameters are:
- num_apply_agents (number of agent threads)
- maxagents_correlid (number of agent threads for batch jobs with the same job name))
- memory_limit (memory buffer for the receive queue)
The following figure illustrates how these Q Apply parameters help control the program's interaction with IBM® MQ and DB2®.
Figure 1. Important tuning parameters for the Q Apply program: 1. num_apply_agents and 2. memory_limit. This figure shows how some of the Q Apply program's activities are controlled by operating parameters that you can tune: A browser thread gets messages from the receive queue and rebuilds transactions from them. The memory_limit parameter controls the size of the memory buffer that is used to hold these reconstructed transactions. Next, the browser schedules which transactions can be applied in parallel by a number of agents that you set with the num_apply_agents parameter. Then the agents take available transactions, rebuild SQL statements, and execute them.
- Q Apply instance-level parameters
- These parameters affect overall program operation, but have little
impact on performance. These parameters are:
- monitor_interval (frequency of inserts into the monitor table)
- prune_interval (frequency of deletes from the monitor and trace tables)