regardless of if the server the db is on is clustered or not, the same principles apply
backup database dbname to file = 'c:\backup.bak'
same goes with the restore
here is an article that will help:
http://www.sqlteam.com/item.asp?ItemID=26689
|||Hi Tim,
Yes the db is on clustered.. do i need to backup each db on each node? or just backup the active node's db?
|||If the cluster is active/active (two instances, one on each node), and there are databases on the 2nd node that you want to backup, then you'll likely need to back that db up as well. Otherwise, you can backup the database on the active node.
The thing about a clustered environment is that it can host multiple instances. These instances are mutually exclusive from each other, so you can treat them like they are different server instances, even though at the time they may be on one node. So, just backup the databases as you normally would as if it was a nonclustered environment.
HTH,
Tim
|||To take this to the simplest case:
Take a 2 node cluster with 1 instance, and 1 user database:
There is only one physical copy of that database. At any point in time, it may be mounted on node "A" or node "B".
The instance name is always the same, regardless of which physical machine hosts it.
So, to back up the database, you run a backup job on whichever node is currently hosting the instance.
Ideally, your backups go to media which is accessable from all cluster nodes, so that restores are clean as well. (another shared disk, or a network share if perf isn't as critical)
Restore is just the same: Go to the node which has your instance currently hosted (or move the instance to the node that you're currently on), and run restore.
For more complex configurations, just take the above and iterate for each instance.
There are 3rd party backup products which operate over the network, which will automatically go to the node currently hosting a database (via the instance's virtual hostname) and do the backup/restore on the appropriate machine.
No comments:
Post a Comment