Using 6.0.5.2 we have 3 gateway processes to send various regions over WAN to other location.
In our setup that has been working for about a year we enable gateway on the regions in cache server that actually holds the data without specifying hub id. Then in configuration for the gateways we specify a subset of the regions and we set the partitions to be empty.
We also have another gateway process that forwards some events onto JMS topic and it receives the events for the regions that it has in its XML configuration.
As a result it appears that we can control what events go to what gateways by specifying the regions in the gateway processes.
Are there any downsides to this approach? Are the gateways that do not have a given region listed in the XML configuration really getting an update event? Would that event just get discarded?
We have fairly complex monitoring and queue size on Gateway A for example increase as we make bulk updates to the region that is defined in that gateway. The queues stay 0 in the other gateways but it's possible that the gateway receives those events and then discards them.
Finally, when using hub-id in the region attributes, does the gateway process need to specify the region in its XML configuration?
Thanks,
David