Configuring failover settings

DNS failover

Identity failover

Failover Type indicates what name resolution method will be used to redirect users to the target SQL server in the event of a source failure. By default, DNS Failover is selected.

Note:

For clustered environments, only DNS failover is supported.

DNS failover

DNS Failover is the recommended method for failover. Use this option if you want to failover by updating the DNS records associated with the source. This will modify all source server A, MX, and PTR-type DNS resource records to point to the target.

In DNS Failover, the DNS records for the source server are modified to point to the target server’s IP address. This allows clients to resolve the source SQL server name to the target server’s network name and IP address at failover time. DNS Failover eliminates duplicate server name and IP addresses on your network.

After you select the DNS Failover option, click Configure. The Configure DNS Failover window will appear.

Configure the following information for DNS failover:

1. To add additional DNS server IP addresses, type the IP address into the DNS Server field, then click Add.
2. The list box under the DNS Server field contains all DNS IP addresses for the source and target servers. The label after the DNS IP address indicates whether the DNS IP address belongs to the source, target, or both.

Note:

To remove an IP address from the DNS server list, select the IP address, then click the Delete button.

3. In the Source IP column, select the checkbox next to the source IP address(es) to be monitored for failover.
4. In the Target IP column, select the target IP address to be used when failover occurs.

Note:

If the target is monitoring multiple source IP addresses for failover, each monitored source IP address must be mapped to a unique target IP address.

5. To specify the value the Application Manager will establish for Time to Live (TTL) on the source’s affected DNS records, select the Update TTL checkbox, then enter the desired update interval (in seconds). The default is the current maximum TTL of all the source’s A records. The recommended value is 300 seconds (5 minutes).

Note:

If the server that the Application Manager is running on has a problem connecting to the DNS server through WMI, the TTL record will not be updated. This will cause the Application Manager to return an error that the RPC server is unavailable.

6. In the Username field, enter the user name that will be used to access/modify DNS records. The login account MUST be a member of the DNSAdmins group for the domain in which the DNS server resides. For more information about permissions, see Application Manager Credentials.

You may enter a user name for a different domain by entering a fully-qualified user name. The fully-qualified user name must be in the format domain\username or username@domain. If you enter a non-qualified name, the DNS domain will be used by default. The domain name is obtained from the DNS server name, provided that reverse lookup in DNS is enabled. For more information about enabling reverse lookup, refer to your Microsoft documentation.

7. In the Password field, enter the password that will be used to access/modify DNS records.
8. Click the Test button to validate that DNS failover is configured correctly for the selected DNS server(s) and that the specified credentials are sufficient to update DNS.
9. When the DNS configuration is complete, click OK to save your entries and return to the Configure Protection window.

Note:

If you are running Windows Server 2000 on the primary DNS server hosting zones or domains that contain source and/or target resource records, you must have the DNS WMI Provider installed on that DNS server. The Dynamic updates setting for the DNS zone should be set to Secure only. Otherwise, you must disable dynamic registration on the source server in order to prevent the source from reclaiming its DNS record. If a hosts file entry for the source server exists on the client machine, errors may occur during a failover and failback. Reverse lookup in DNS should be enabled. For more information about enabling reverse lookup, refer to your Microsoft documentation. DNS registration for the private (devoted to Double-Take Availability) NIC IP should be disabled. If dynamic updates are enabled on a standard primary zone, the source server will be able to update its DNS records after failover. To prevent this, configure DNS to use an Active Directory-integrated zone. In a clustered server in an environment with more than one NIC, if the heartbeat network IP is at the top of the binding order, a socket error occurs when you attempt to select or configure the server. This error can be avoided by changing the binding order on all the source and target servers/nodes involved so that the domain IP address should be at the top of the order. Then, reboot the servers involved. For more information about using the DNS Failover utility, access the dfo.exe help by typing dfo.exe /? from a command prompt.

Identity failover

Select this option if you want to failover by transferring the source IP address and name to the target. When using identity failover, it is possible that a name and/or IP address conflict can occur either during failover or when the original source server comes back online. To avoid this conflict, use DNS Failover.

Note:

Identity failover is not supported in clustered environments. Application Manager cannot be used for failover or failback when the Identity failover method is selected. However, the Failover Control Center can still be used to initiate failover. DNS failover reduces downtime and provides other benefits. It is recommended for most environments.

In Identity Failover, the target’s physical identity is modified to match the source during a failover. This includes the target adopting the source server’s name, primary IP address, and drive shares during a failover. Identity failover may be required in the following situations:

• Access to the domain controller or DNS server is not available (for example, due to permissions) from the account that Double-Take Availability runs under on the source/target servers.
• If you determine that the time it takes to propagate the necessary DNS or Active Directory changes to the rest of your environment is not acceptable. The time needed to propagate these changes depends on your Active Directory Replication and DNS server settings.
• If you have client applications that are configured to connect to an IP address rather than a server name.

After you select the Identity Failover option, click Configure. The Configure Identity Failover window will appear.

Enter the following information for Identity failover:

1. In the Source IP column, select the source IP address(es) to be monitored for failover.
2. In the Target NIC column, select the target NIC to be used when failover occurs.
3. The Target IP Addresses area displays the IP address(es) of the selected target NIC.
4. Select the IP Address checkbox if you want the specified source IP address to be monitored (Default = selected).

Note:

If your source and target servers are on different subnets, you should NOT failover the IP address. Instead, if the server name is required you should choose DNS failover with the Advanced switch and select the Server Name. For more information, see Advanced settings.

5. Select the Server Name checkbox if you want the source name to be added to the target when failover occurs (Default = NOT selected).
6. Select the Shares checkbox if you want the source file shares to be added to the target when failover occurs (Default = NOT selected).
7. Select the Active Directory Hostname (SPNs) checkbox to remove the host SPN (Service Principle Name) from Active Directory for the source automatically and add it to Active Directory for the target during failover (Default = NOT selected).
8. Click OK to save your entries and return to the Configure Protection window.

Application Manager will determine the appropriate SQL services to start/stop based on your operating system/SQL configuration. You should only modify this selection if there are additional services that need to be started along with SQL during the failover/failback process.

Modifying the default configuration for services may affect whether data can be successfully replicated. Do not modify the services to start/stop unless you are very familiar with Double-Take Availability and SQL.

1. To add a service, click Add. The Add Service window will appear.

2. Select the Service name from the drop-down box, then click Add; or, type the name of a service, then press Enter.
3. If available, select the Service must be stopped on target checkbox. Most services must be stopped on the target for replication to occur properly.
4. When you are finished entering services, click Close to return to the Failover configuration window.
5. To remove a service, select one or more services, then click Remove.

You can remove all of the services besides the required SQL server service (MSSQLSERVER for a default instance, or MSSQL$<instance_name> for a named instance).

6. Services will be started in the displayed order (top-to-bottom starting with the left column) and shutdown in reverse order (bottom-to-top starting with the right column). To re-arrange the order in which services are started/stopped:

1. Add any additional services.
2. Select a service in the list, then use the up and down arrow buttons to move the service to the appropriate position in startup/shutdown order.

If you are using clustered SQL servers, you will select resources instead of services to bring online and offline during failover. The Resource selection works exactly like the service selection functionality.