Need recommendation for dash copy mode (nw optmized / disk read optimized) in grid environment.
Q1 : If I have 2 nodes grid at primary site, each with one partitioned of ddb, and 2 nodes grid at secondary site, what is the recommended dash copy mode is?
Q2 : the same as Q1, but single node MA at secondary site, what is the recommended dash copy mode is?
I take this into consideration, as ddb in each node of grid may not in sync, cause from one node may be down for maintenance for a period, which makes that node cannot keep up with ddb in other node for some of ddb entries.
My concern is....
If I auxcopy by using "nw optimized" mode, on the node with less ddb entries, it will read the ddb to determine which chunks should be replicated (which may not equal to another ddb in other node in grid). Could it replicate some chunks which are already replicated to other site in the past? which can create unnecessary network load between sites.
Also, if both nodes' ddb in grid is not in sync, how I can sync it to be equal manually, or it have to be left unsync forever (or it have internal mechanism which sync them together periodically, or they are not designed to be in sync from the start, which I don't know)