How to restore part of a mailbox from an RDB

Solution: Publish mail-enabled BE account in the GAL, as per this document. I'll leave my edits in place for troubleshooting purposes of other people.

Check the status of the Exchange VSS writer. From a CMD prompt, type VSSADMIN LIST WRITERS. A list of writers (including Exchange) will output. They should all be in {1} stable condition.

On the restore, make sure that you have the "Restore to Server" field properly populated.

EDIT: From here:

Ensure that GRT is enabled before you run backups if you intend to be able to restore individual items. You can enable GRT for all backup jobs in Tools > Options. Or you can enable GRT for individual backup jobs on the Backup Job Properties dialog.

Back up your current or most recent GRT-enabled backup jobs to disk. It is more convenient to work with GRT-enabled jobs on the volumes that do not have file size limitations. You can create duplicate backup jobs and send copies of your backups to tape for archival purposes.

Use a backup-to-disk folder on a volume that does not have file size limitations as the destination for any backups that are enabled for GRT. An NTFS drive is an example of a volume without file size limitations. Some examples of volumes that have file size limitations include FAT and FAT32 volumes.

Review the requirements for staging locations in the Administrator's Guide.

You must use a staging location for GRT-enabled jobs in the following scenarios:

You back up to or restore from a volume with file size limitations.

You restore granular items from tape.

You run an offhost backup job.

Use a volume that is not your system volume for a staging location. The volume on which the staging location resides should have at least as much available space as the size of your largest GRT-enabled backup job. You can change the default staging locations in the default backup and restore option settings.

I'm going to have to look a little closer at BE2010 restore procedures, which will be later, but I would start with the above. Maybe it's because you're not staging it?

EDIT 2: Are you using incremental backups?

GRT restores will not work for individual mailboxes or items if the backup was to tape. From this document page 314:

Backup Exec must have access to a uniquely named mailbox within the Exchange organization for backup and restore of the Information Store.

See “Requirements for accessing Exchange mailboxes ” on page 1081. You cannot restore individual mailboxes and messages if both of the following conditions exist:

■ The incremental or the differential backup method was used.

■ The destination was a tape device.

If you create full, differential, or incremental backups, GRT-enabled jobs have the following restrictions:

■ The full, differential, and incremental job templates must be part of a policy.

■ The destination device must be a backup-to-disk folder.

■ The backup sets from the full, differential, and incremental jobs must be on the same volume.

One thing that may be slowing the process is the key_buffer_size, which is the size of the buffer used for index blocks. Tune this to at least 30% of your RAM or the re-indexing process will probably be too slow.

For reference, if you were using InnoDB and foreign keys, you could also disable foreign key checks and re-enable it at the end (using SET FOREIGN_KEY_CHECKS=0 and SET FOREIGN_KEY_CHECKS=1).