Using Node classification to control/config hazelcast operations in production environment in 722
Hi
We are looking for documentation/info around this scenario... if we can use node classification and group a specific set of jvms/nodes in one data center and configure the prconfig's of those jvms ( and open only the relavant ports needed for hazelcast if needed) there by be able to avoid any hazelcast related pinging/communication to the other nodes (that does not belong to the specific classification)
Thanks
Manoj
***Updated by Moderator: Marissa to update categories***
Pegasystems Inc.
US
Hi Manoj,
The node list is governed by pr_sys_statusnodes. Are you saying you have a cluster sharing the same PegaRULES DB that you want to prevent all the nodes from participating in the same hazelcast cluster?
Let me know,
Thanks,
jim
Pegasystems Inc.
US
Hi Manoj,
If you have multiple clusters sharing the same machine (IP) then segment the hazelcast ports to prevent cross talk (ultimately they will stop talking to each other). This only makes sense if they are not part of the same Pega cluster, but are sharing machine resources.
type 1
<env name=“cluster/hazelcast/ports” value=“PORT_RANGE” /> for example <env name=“cluster/hazelcast/ports” value=“5701-5710” />
type 2
<env name=“cluster/hazelcast/ports” value=“PORT_RANGE” /> for example <env name=“cluster/hazelcast/ports” value=“5711-5721” />