Our Exchange 2010 deployment consists of three mailbox servers in a DAG, to "Primary" servers and one "Standby" server. Each primary server hosts several information stores with user mailboxes divided more-or-less evenly across them. The standby server hosts replicated copies of all of the information store databases hosted on the two primary servers.
We use Backup Exec to back up Exchange, targeting the DAG. Our BE environment consists of a single-server deployment.
Until recently, all three mailbox servers were located at our HQ site, but recently we relocated the "standby" server to a colo site. Since the servers that hold my "active" databases are located in the HQ location along with the BE server, I did not anticipate a problem with the backup server at the site running it's regular backup jobs, since I assumed that BE would pull data for it's backups from the active replicas.
However, when the backup job on the BE server located in the HQ ran, it saturated the network link to the datacenter, ostensibly pulling data from there. I verified that all of my information stores were MOUNTED in the HQ mailbox servers, and HEALTHY on the datacenter mailbox server.
So my question is, when a DAG is targeted, where does Backup Exec pull the data from? This of course begs the question of whether or not Backup Exec can be configured to only pull data from specific IS replicas?