Yesterday I had some issue with docHarbor that made me think of a similar situation in CMOD.
Suppose for whatever reason, I decide to have 2 object servers at the same geogrophical location.
Now for load balancing, backup, etc the same reports are loaded into both the servers.
So practically object server 1 and object server 2 are each other's replica.
This means there would be 2 copies of each document and each would have a unique document ID.
When a user will search for a document CMOD would return 2 rows pointing to duplicate documents.
I understand that no one would practically do such an implementation in CMOD
But when the user selects one of the documents and hits view, how would CMOD know which object server would have the document since during document retrieval, the Library server is bypassed.
Assume that the documents are still in cache of the Object servers.