Click here to monitor SSC
SQLServerCentral is supported by Red Gate Software Ltd.
 
Log in  ::  Register  ::  Not logged in
 
 
 
        
Home       Members    Calendar    Who's On


Add to briefcase

Always ON SQL Server 2012 Expand / Collapse
Author
Message
Posted Thursday, August 7, 2014 8:49 AM
Grasshopper

GrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopper

Group: General Forum Members
Last Login: Thursday, October 9, 2014 8:44 AM
Points: 12, Visits: 58
Hello:

I just configured Always ON with two nodes in the same site dbprod1a and dbprod1b. For Windows FC, we are not using shared storage and hence there is no quorum drive specified. Both nodes are in the same site. I have configured a couple of DAGs with listeners and everything seems to be working as expected. The two nodes on each DAG are setup for synchronous replication and automatic failover.

I have this part figured out!

The question I have is about the third node that I plan on adding to this cluster.

The third node dbprod1c will be in a different site and will be setup for asynchronous replication. The question I have is around configuring the listener service and understanding how failover takes place during a real DR situation.

Scenario ---

Nodes a and b are on the same site and therefore on the same subnet (example: 10.2.0.x). The listener service is on the same subnet as the two nodes.

Node c is in a different site and therefore on another subnet (example: 10.4.0.x).

Note that all systems will be on the same VLAN.

I can easily add node c to the Windows FC.

What do I need to do to add it to the database availability group? How should the listener be configured or does it matter? Are applications connecting to the "listener"?

What will happen if my primary site fails and the link between the two sites is down? Should the application be configured to connect to read-only copy and connect to the endpoint in secondary site directly?

I would like some tips on inter-site failover during real DR situation. We have web/app servers that will need to connect to these databases and uptime is crucial.

Please advise.

Thanks,
Geeta
Post #1600682
Posted Thursday, August 7, 2014 9:50 AM


SSCarpal Tunnel

SSCarpal TunnelSSCarpal TunnelSSCarpal TunnelSSCarpal TunnelSSCarpal TunnelSSCarpal TunnelSSCarpal TunnelSSCarpal TunnelSSCarpal Tunnel

Group: General Forum Members
Last Login: Today @ 9:29 AM
Points: 4,436, Visits: 6,337
I strongly recommend you bring Allan Hirt (http://sqlha.com) on board to help you understand what your options are, what you need and how to get the two to match up. Then configuration, testing and monitoring setup. IMNSHO there is nobody in the business better than him at what you are asking about, and you just don't want to be setting up your company's HA/DR based on some forum posts!!!




Best,

Kevin G. Boles
SQL Server Consultant
SQL MVP 2007-2012
TheSQLGuru at GMail
Post #1600723
Posted Friday, August 8, 2014 4:02 AM


SSCertifiable

SSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiable

Group: General Forum Members
Last Login: Yesterday @ 7:46 AM
Points: 6,619, Visits: 14,185
Firstly, have a look at this.

http://msdn.microsoft.com/en-gb/library/ff878716.aspx


Libby1981 (8/7/2014)
Hello:The question I have is about the third node that I plan on adding to this cluster.

The third node dbprod1c will be in a different site and will be setup for asynchronous replication. The question I have is around configuring the listener service and understanding how failover takes place during a real DR situation.

Scenario ---

Nodes a and b are on the same site and therefore on the same subnet (example: 10.2.0.x). The listener service is on the same subnet as the two nodes.

Node c is in a different site and therefore on another subnet (example: 10.4.0.x).

Note that all systems will be on the same VLAN.

I can easily add node c to the Windows FC.

What do I need to do to add it to the database availability group? How should the listener be configured or does it matter? Are applications connecting to the "listener"?

You can either utilise a single IP address within the VLAN that all 3 nodes will be connected to, with SQL server 2012 on however, you are also able to use multiple subnet IPs instead. If you look at the network name resource in a cluster group and go to the dependencies tab, you'll see you have the option to select multiple IP resources and then set the name resource to require all IPs online or a single IP online, this is all detailed in the link i posted above. Have a good look at it and do some testing then post back if you're stuck




Libby1981 (8/7/2014)
What will happen if my primary site fails and the link between the two sites is down? Should the application be configured to connect to read-only copy and connect to the endpoint in secondary site directly?

I would like some tips on inter-site failover during real DR situation. We have web/app servers that will need to connect to these databases and uptime is crucial.

Please advise.

Thanks,
Geeta

Have you given thought to what quorum configuration you'll use when the 3rd node is added, will you want the 3rd node on the DR site to have a vote and potentially influence cluster behaviour?


-----------------------------------------------------------------------------------------------------------

"Ya can't make an omelette without breaking just a few eggs"
Post #1601098
« Prev Topic | Next Topic »

Add to briefcase

Permissions Expand / Collapse