2012-12-02 07:16 AM - edited 2015-12-18 01:16 AM
We are in the process of upgrading to Exchange 2010 and have a streched dag with the following set up.
(a) Site A (Main Site) has two servers. One running CAS and Hub Transport role and other running Mailbox Role.
(B) Site B (DR Site) has two servers. One running CAS and Hub Transport role and other running Mailbox Role.
We have installed SnapManager for Exchange version 6.04 on both mailbox servers in main site and DR site. We want to be able to backup the DAG so we are able to do a point in time restore if required.
(a) Do we need to back up both sites or would backing up just the DR site would work?
(b) If the Snapmanager is set up to backup the DAG, if one of the site goes down, how would the other site get backed up by SnapManger?
(c) If we need to do a gapless restore on the main site, and we go with the option of just backing up the DR site, how would that work.
Please guide us in the right direction and all help/feedbacks would be appreciated.
2012-12-04 04:53 AM
Answering your questions by letter....
(a). You don't need to back up both sites. You can select the server in the DR site and schedule "full" backup jobs and keep the number of days or number of snapshots that you need. When SME takes the backup it tells Exchange a backup has been done and that server node tells the node where the database is active. The log truncation takes place from the active and comes to the passive copy(s) as appropriate. Obviously the LUN size will be the same in primary and secondary site but remember that the DR site will have larger FlexVol sizes.
(b) If you do connect to the DAG (cluster) rather than the server (node) then you will decide whether to backup the active or the passive and if you selected the passive you get the option on which activation preference to back up. If you want to do backups in the DR site I guess you're likely to want to create backups on the passive copy at activation preference 3. However, in that case, if the servers in that site do go offline you won't have a backup. Personally speaking I would advocate just connecting to the node in DR and doing a backup of the databases on that server.....
(c) The gapless feature is pretty cool but I would suggest you think about simply creating jobs on the nodes in the primary site and selecting the "Copy Backup" option (the gapless thing you talk about). Do one a day, maybe two and at a time that suits you. That's a great way of eliminating a reseed operation if those copies ever go bad and you end up activating a different mailbox copy. Rather than have Exchange do the reseed you can use PowerShell to do a volume snap restore of a previous database snapshot and hand the database back to Exchange for it to feed in the necessary logs. You might have to manually copy some logs out of the SnapInfo directory on the copy in the DR location (where you are doing the backups) before issuing the Resume command for the DAG copy but you will know which logs you need by reading through the directory. It's easy enough to read the HEX names of the log files and decide which, if any, you need to drop into the directory path.
Hope this helps.