You may lose access to mailboxes because the information store has crashed, is corrupt, or is otherwise inaccessible.

Front-end Microsoft Exchange Server 2007 servers are not a concern for data corruption because they do not store data; they can be added and removed as necessary should a server fail. It is the back-end Exchange servers which provide the data storage. Depending on the type of data loss experienced, there are a variety of mechanisms used to recover data.

If a single node of a back-end cluster is lost, data loss and recovery is not an issue, since the virtual server that was running on that server will be able to fail over to the inactive node. The failed node would then be rebuilt and added back into the cluster.

Note:
A single Exchange Standard Edition server can support up to five storage groups. Enterprise Edition servers can support up to 50 storage groups. We recommend you allocate a seprate database for each storage group.

If an Exchange information store crashes, there are two options for restoring service to users:

• Option 1: Remove Affected Database Files and Mount Affected Store - The first method enables fast restoration of service to users to with previous data becoming available after some time. This method involves removing the damaged database files and mounting the affected store (but not deleting the object from Active Directory). This causes a new empty store to be created. Users will then be able to access their mailboxes; however, old mail will not be available. Depending on your customers this may not be an issue (for example, POP clients downloading mail). A restore then can be performed to a recovery storage group and the data transferred back to the mailboxes of the users. This process does not affect the link between the store's GUID in Active Directory and the GUID in the MPS Resource database.
  
• Option 2: Full Restore from Backup - The second option is to perform a full restore of the data from backup. During this time the users will not be able to access their mailboxes because the store will not be mounted until the restore completes. In the case where a single mailbox needs to be restored, the store can be restored to the recovery group and then the mailbox extracted from that store. This process does not break the critical link between the user, store, and MPS database.
  

  Caution:
  If a store is deleted from Active Directory and subsequently recreated, it will be allocated a new GUID. The generation of a new GUID will break the link between the store and the MPS database. Should this situation occur, you must contact Microsoft Support to help resolve the issue.

The mailbox store and public folder store data in your Exchange Server 2007 databases and transaction log files are the most important data to back up in your Exchange organization. You can use an Exchange database backup to restore damaged mailbox or public folder stores to a functioning server that is running Exchange Server 2007.

Exchange Server 2007 Disaster Recovery gives complete information on using Exchange backup to perform regular full and incremental backups, including transaction logs used to bring the restored mail store up to date to the point at which it was lost or corrupted.

When you lose all access to mailboxes because of an information store failure, you will need to restore the information store from backup, and then replay transaction logs to resynchronize the mail store.

When you use Exchange backup to restore Exchange databases, calls are made to the Exchange Extensible Storage Engine (ESE) to restore Exchange database files and their associated log files. You can use Exchange database backups to restore one or more damaged mailbox or public folder stores.

In a disaster recovery scenario that involves rebuilding a server that is not part of a cluster, use Backup to restore your Exchange databases after you run Exchange Setup and any Exchange service packs in Disaster Recovery mode.

Exchange Server 2007 Disaster Recovery gives complete information on using Exchange Backup to restore mail stores to existing or newly provisioned Exchange servers.

AlpineSkiHouse.com backs up its Exchange data nightly at 3:00 A.M. and does not have circular logging enabled. At 9:00 A.M. users report they are not able to access mail.

The results of a brief investigation determine that the BusinessStore1 has failed and can not be mounted. The decision is made to restore the database from the previous night's backup. The administrators restore the database and allow the transactions that have occurred to be replayed.

The database is operational again at 9:30 A.M. and all data is present prior to the failure. The administrators investigate further the cause of the failure.

An alternative option would have been to remove the damaged store and then restore to the recovery storage group. However, this method requires the logs to be present to be replayed.