SCDPM 2007 SQL Self Recovery with clusters
The past week our MS SQL environment got a health check my Microsoft, comparable with ADRAP for AD. The downside of being a SCCM/SCOM/SCDPM administrator is that you get involved in whatever happens with other products in your company, because you deploy them, monitor them and back them up.
There are two things that make a DPM admin's life easier when it comes to recovery: enabling end-user recovery - which is a simple AD schema update if you're post-Vista (on XP you need to install an agent on the clients), and enabling SQL self recovery for SAs. The latter is slightly more complicated.
Click on the Protection tab
Go to the Action menu and click Configure self service recovery for SQL server.
And there you have the DPM Self Service Recovery Configuration Tool for SQL Server
As I said earlier configuring this is a straightforward wizard-driven process like many of the other wizards an Administrator encounters. You'll have to create users, add AD users or groups to them, define the databases they can recover and also to where they can recover databases.
It really doesn't need any explanation, but I'd like to stress however that in DPM 2007 you can only recover SQL DBs (or any other kind of data) to a server that has an agent connected to the same DPM servers as the machine you are trying to recover data from.
In this whole process there's a point where you have to choose the databases that the SA can recover. You enter a SQL server in the following format:
Whatever it says, the SQLServer van be both hostname or FQDN/
Easy, right? Yes, until you want to add a clustered server. You're very likely to run into the following error message:
It took me a while to figure out, but you need to useThe specified SQL instance <YourCluster> is invalid or cannot be found (ID 32608).
If a SQL instance exists with the specified name, use Get-Datasource -ProductionServer <ProductionServer> -Inquire cmdlet to refresh the information. Then retry the Operation.
If you're not used to clustering - and I'm not - then the cluster network name might be a bit puzzling. The easiest way to find that out is to simply look at your protection group.
In this example, the cluster network name is BRUKE.WILSON-FAILOVER.WSOUZA.LOCAL, the SQL server is WILSON-SQL2K8R2 and the instance is BRUKE.
Using this syntax - and here you really do need to use the FQDN - you will be able to add clustered SQL servers to the SQL self recovery tool too.
I used images from Wilson Souza to illustrate the clustering as I can't take a screenshot of my own environment without censoring, but censoring in this case would make the whole picture rather useless.