locked
Always On High Availability RRS feed

  • Question

  • I know that when you are trying to setup Always on you need to do a few things:

    1) You need to setup windows failover cluster; give a name to the cluster, give an ip to the cluster you just named and add the machines.

    2) you need to check the box for always on in configuration manager on all the nodes

    3) you need to create a shared folder on node1 where you create backups of the databases you want in the group

    4) you run the wizard and you add the replica for the failover

    5) you create a listener with a name and static IP.

    *from the tutorials that I've seen, this is pretty much what you need to do in a nutshell

    Here are my questions:

    From number 3, why do we need a shared folder? Is it just for accessing the backups from the initial setup, or is it something where node2 creates the mdf and log files it needs, like a central repository in a SAN for 2008 clustering?

    From number 5, why do we need to create a listener with a static ip, isn't that why we created the windows clustering with a name and ip to redirect us to the primary/secondary server?

    Wednesday, August 28, 2013 2:40 PM

Answers

  • Hi There!

    Number 3 is only used for the initial replication. The wizard will do a full backup of the DB and will restore it on the replica server. Once that is done, you won't need the shared folder anymore.

    For number 5: when you create the listener, a new computer object with the listener name gets created and DNS entry are created as well. Creating the listener is not mandatory BUT if you need to setup SQL Routing to leverage the "ApplicationIntent" parameter in the connection string you WILL need the listener.

    I hope this helps.


    • Marked as answer by Fanny Liu Wednesday, September 4, 2013 9:45 AM
    Wednesday, August 28, 2013 2:56 PM
  • Hi thanks!  So if I don't create the listener, and lets say the Primary goes down, does the Windows Failover handle the failover to the secondary?

    The Failover will be handled for the AG/FCI.

    But you will not be able to connect automatically to the now new Primary replica, because you will not know where it is. This is what the listener is for. Like an automatically updated alias, pointing to the Primary.

    You can of course use a DNS alias instead.


    Andreas Wolter | Microsoft Certified Master SQL Server
    Blog: www.insidesql.org/blogs/andreaswolter
    Web: www.andreas-wolter.com

    • Marked as answer by Fanny Liu Wednesday, September 4, 2013 9:45 AM
    Wednesday, August 28, 2013 6:37 PM

All replies

  • Hi There!

    Number 3 is only used for the initial replication. The wizard will do a full backup of the DB and will restore it on the replica server. Once that is done, you won't need the shared folder anymore.

    For number 5: when you create the listener, a new computer object with the listener name gets created and DNS entry are created as well. Creating the listener is not mandatory BUT if you need to setup SQL Routing to leverage the "ApplicationIntent" parameter in the connection string you WILL need the listener.

    I hope this helps.


    • Marked as answer by Fanny Liu Wednesday, September 4, 2013 9:45 AM
    Wednesday, August 28, 2013 2:56 PM
  • Hi thanks!  So if I don't create the listener, and lets say the Primary goes down, does the Windows Failover handle the failover to the secondary?
    Wednesday, August 28, 2013 4:06 PM
  • Yes it does, with or without listener.

    As a warning, remember that, as per article http://msdn.microsoft.com/en-us/library/ff929171.aspx:

    Restrictions on Using The WSFC Failover Cluster Manager with Availability Groups

    Do not use the Failover Cluster Manager to manipulate availability groups, for example:

    • Do not add or remove resources in the clustered service (resource group) for the availability group.

    • Do not change any availability group properties, such as the possible owners and preferred owners. These properties are set automatically by the availability group.

    • Do not use the Failover Cluster Manager to move availability groups to different nodes or to fail over availability groups. The Failover Cluster Manager is not aware of the synchronization status of the availability replicas, and doing so can lead to extended downtime. You must use Transact-SQL or SQL Server Management Studio.

    This is just to make you aware of the fact that the Alwayson cluster role is a very special one.


    Wednesday, August 28, 2013 4:21 PM
  • Hi thanks!  So if I don't create the listener, and lets say the Primary goes down, does the Windows Failover handle the failover to the secondary?

    The Failover will be handled for the AG/FCI.

    But you will not be able to connect automatically to the now new Primary replica, because you will not know where it is. This is what the listener is for. Like an automatically updated alias, pointing to the Primary.

    You can of course use a DNS alias instead.


    Andreas Wolter | Microsoft Certified Master SQL Server
    Blog: www.insidesql.org/blogs/andreaswolter
    Web: www.andreas-wolter.com

    • Marked as answer by Fanny Liu Wednesday, September 4, 2013 9:45 AM
    Wednesday, August 28, 2013 6:37 PM
  • Cool!  Thanks guys!  This is really helping me clear some of the confusion I was having.
    Wednesday, August 28, 2013 7:17 PM
  • For the Listener, the DNS name can be anything I want to name it in the wizard right?

    So when I'm trying to set this up at my work place, how many IPs, do you think I should request from the Network Admins?

    • Edited by Diango Wednesday, August 28, 2013 8:23 PM
    Wednesday, August 28, 2013 8:13 PM
  • From BOL:

    Each availability group listener requires a DNS host name that is unique
    in the domain and in NetBIOS. The DNS name is a string value. This name
    can contain only alphanumeric characters, dashes (-), and hyphens (_),
    in any order. DNS host names are case insensitive. The maximum length is
    63 characters, however, in SQL Server Management Studio, the maximum
    length you can specify is 15 characters.

    http://technet.microsoft.com/en-us/library/hh213080.aspx

    The Number of IP's depends on the exact configuration and numer of nodes that you have.

    You should first read through all of the documentation before you start a "High Availability" Project. You don't want to miss anything.

    You can start here for example http://technet.microsoft.com/de-de/library/hh510230.aspx


    Andreas Wolter | Microsoft Certified Master SQL Server
    Blog: www.insidesql.org/blogs/andreaswolter
    Web: www.andreas-wolter.com

    Wednesday, August 28, 2013 8:44 PM