Replicating DynaCache service in a container

This section describes how to replicate Dynacache service in a containerized HCL DX (v9.5 CF_172 container release or later) by customizing timeout properties in the Resource Environment Provider.

Previous versions up to HCL WebSphere Portal 8.5 and 9.0 uses Dynamic Cache (DynaCache), a cluster-aware cache service that allows all HCL WebSphere Portal cluster members to know any name/value pair changes made by any member. Having a cluster-aware cache does not make sense for HCL WebSphere Portal 9.5 onwards, as the versions are on a Portal farming or containerization setup. To work around this issue, an administrator can customize the timeout properties in the WAS Resource Environment to replicate the DynaCache service.

Configure the following timeout parameters in the WAS Resource Provider (REP):

• Time between 'reads' of the database table containing invalidation messages. By default, timeout is 1 minute (60000 milliseconds) . Practically, this means this is the longest delay before any dynacache invalidation is 'known' on each member (e.g. Kubernetes, POD, farm worker).
  db.cache.invalidation.read.freq = (timeout in milliseconds)

• Age before an entry in the database table containing invalidation messages is deleted. By default, timeout is 10 minutes (600000 milliseconds). Generally, the number should be much larger than the 'red' frequency to ensure that invalidation messages are read before they are deleted.
  db.cache.invalidation.cleanup.freq = (timeout in milliseconds)

To replicate DynaCache service in an HCL WebSphere Portal farm deployment, see Tuning the servers in the environment in Roadmap: Portal farm.

Parent topic: Customization