A cached resource that exists in a TIBCO Data Virtualization (TDV) cluster will normally depend on a unique clusterid value found in its corresponding cache status table.  In a non-clustered, stand-alone single TDV instance, the clusterid value will be NULL.

The clusterid value is generated upon creation of a cluster, when the cluster is initially declared in the first TDV node. 
If the cluster nodes are all dissociated and the cluster declaration entity is then deleted, all cached resources will have their corresponding row in their cache status table deleted, and the cached resource's status will become: NOT LOADED
The caches will then need to be refreshed at a later time.

Sometimes during events when the cluster nodes have either unexpectedly dissociated or need to brought down and added back in due to TDV issues, you want to preserve the last remaining node in the cluster and not drop the declared cluster entity, if possible.  One node must still belong to the previously declared cluster, in order to preserve that unique clusterid.  Otherwise, a new unique clusterid will be generated when a new cluster is declared.

Any issues with the individual nodes should be managed separately while they are not clustered together, and if the original clusterid remains in existence on one node, the rest of the nodes can join back and preserve the state of the caches.



 

Why does recreating the cluster in TIBCO Data Virtualization render all caches as "NOT LOADED"?