2012-06-22 04:53 AM - edited 2015-12-18 01:18 AM
For some reason I can't select database to be backed up in my DAG configuration. I think it might be the way I have the DAG configured but I am told the configuration is correct by Microsoft.
It may also be because I have 2 databases on the same LUN which NetApp says they do not support.
2012-06-22 06:13 AM
/vol/db1_vol/db1_lun ----> db1.edb
/vol/db2_vol/db2_lun ----> db2.edb and so on...
Please refer TR - http://media.netapp.com/documents/tr-4033.pdf
2012-06-22 07:34 AM
Microsoft also recommends that during a failover that a database have the same drive letter on the second memer of the DAG. So you can run 2 databases on the same drive letter if there is a failover of databases in the DAG. This would mean that there is 2 databases on the same drive letter thus on the same LUN.
See below. Maybe I am doing something wrong in the config.
2012-06-22 12:30 PM
So if I am reading all this correctly and I have a 2 member dag and 7 databases (4 on one, 3 on the other) I need to assign a database to a different drive letter so basically, (H, I, J, K for the 1st server) (L, M, N for the second server) for Databases and then Send all Snapinfo and transaction logs to O in order to use snapmanager? I know Microsoft usually recommends seperating the transaction logs as well for performance which adds another 7 drives to a server.
Or my other solution is to create extra large Databases kind of like putting all my eggs in one basket to lessen the number of databases in total.
2012-06-22 01:53 PM
The database and the transaction logs need to be on separate LUNs to support VSS hardware provider snapshots that SME uses.
Exchange databases can share a transaction log LUN and SnapInfo directories can be located in the transaction log LUN.
Exchange 2010 DAGs provide Exchange high availability through having multiple copies of databases. So if you have a 2-node DAG, you will create a copy of each database on each node. If you lose a database, Exchange will automatically start using the surviving copy without disruption to users (if they are using Outlook in cached mode).
I don't know how big your databases are, but I suggest that you use a few large databases instead of a lot of smaller databases. With SME, you can leverage a large database design because the time to restore a 2TB database is basically the same amount of time to restore a 200GB database. Combining the benefits of Exchange HA (multiple database copies) with the rapid restore capabilities of SME/DataONTAP, you are reducing your risk.
In addition, a side benefit of using fewer databases is that your backups will go faster.