jprknight
Mohawk is right, Archive is a logical representation of the data archived (savesets) in the Vault store partitions, savesets can be stored in many different partitions (due to several reasons) but the archive would still remain the same. When the items get archived, each item would have a transaction ID, saveset id which is unique to the items and where the items are localed and in which partition. Therefore Enterprise Vault can easily identify for each user how many items have been archived and where each item is located and in which partition. This can be obtained easily through SQL query.
For example say you had an open Vault Store partition (Partition 1) and there are 1000 items archived from TSO user's archive, then today you close the Vault store partition and you open a new partition (Partition 2), TSO archives manually/scheduled archiving 50 items. So in total
TSO's archive contains 1050 items
1000 archived items will be stored in Vault Store Partition 1 - Which is closed now
50 Items will be stored in Vault Store Partition 2 - Currently open partition.
When the user's perform searches (uisng search.asp page or search02k page) all they get back is results of the number of items indexed from their archive it does not matter which partition the items are stored in.
There is no rule or limit in Enterprise Vault that says Archive A will have all its items stored in Vault Store A and if you roll over to partition B the user A will have a second Archive created because we have closed partition A and open partition B.
I hope that helps!
