SAP NetWeaver on SQL Server 2012 AlwaysON for HA and DR

[mamartins]

Scenario:

Main Data center + DR

3 physical servers for SQL Server (2 on main DC + 1 DR) AlwaysON

SAP Application Servers hosted under a VMWARE cluster

SAP Netweaver 7.40 SR 2 ABAP, media list:

51047515        SQL Server 2012

51048524_1    NW 7.4 SR2 Installation Export

51048524_2    NW 7.4 SR2 Upgrade Export 1/2

51048524_3    NW 7.4 SR2 Upgrade Export 2/2

51048524_4    NW 7.4 SR2 Language 1/2

51048524_5    NW 7.4 SR2 Language 2/2

51049350_10  SAP Kernel 7.42 Windows Server on x64 64bit

51045590       NW AS ABAP for Suite/NW 7.4 - Enhancement Packages

Software minimum requisites are:

Windows 2008R2

SQL Server 2012

Also, a valid Active Directory Domain is mandatory because the Failover Cluster resources and the SAP distributed AS.

Server names:

SRVDB1 (SQL Server node 1 - main DC)

SRVDB2 (SQL Server node 2 - main DC)

SRVDB3 (SQL Server node 3 - DR)

SRVSAS1 (Central Instance)

SRVSAS2 (Primary Application Server)

SRVSAS3 (Additional Application Server installed on DR)

Relevant SAP notes:

http://service.sap.com/sap/support/notes/965908

http://service.sap.com/sap/support/notes/1365796

http://service.sap.com/sap/support/notes/1772688

The SAP High Availability is done inside the main DC using the HA features of the VMWARE cluster. Also the VM hosting the CI will be replicated to the DR site where another VMWARE cluster is running.

After OS installed and Failover Cluster nodes configured, install SQL Server.  On the main DC, the 2 SQL Server nodes can be pure AlwaysON AG or a FCI. On the first case you have independent storage; on the second one the storage is shared. The replication to the DR will be done with an Availability Group (AG). Microsoft support both scenarios: pure AlwaysON AG or a mix of FCI and AG. The differences between them are:

https://msdn.microsoft.com/en-us/library/ff929171.aspx?f=255&MSPPError=-2147217396

	Nodes within an FCI	Replicas within an availability group
Uses WSFC cluster	Yes	Yes
Protection level	Instance	Database
Storage type	Shared	Non-shared1
Storage solutions	Direct attached, SAN, mount points, SMB	Depends on node type
Readable secondaries	No2	Yes
Applicable failover policy settings	WSFC quorum FCI-specific Availability group settings3	WSFC quorum Availability group settings
Failed-over resources	Server, instance, and database	Database only

1 While the replicas in an availability group do not share storage, a replica that is hosted by an FCI uses a shared storage solution as required by that FCI. The storage solution is shared only by nodes within the FCI and not between replicas of the availability group.

2 Whereas synchronous secondary replicas in an availability group are always running on their respective SQL Server instances, secondary nodes in an FCI actually have not started their respective SQL Server instances and are therefore not readable. In an FCI, a secondary node starts its SQL Server instance only when the resource group ownership is transferred to it during an FCI failover. However, on the active FCI node, when an FCI-hosted database belongs to an availability group, if the local availability replica is running as a readable secondary replica, the database is readable.

3 Failover policy settings for the availability group apply to all replicas, whether it is hosted in a standalone instance or an FCI instance.

Using SWPM, the CI is installed on the appropriate VM. No connection to DB is needed at this time. Once finished, the SAPMNT share will be visible as \\SRVSAS1\SAPMNT.

Then we should logon on the first SQL Server node and start SWPM to install the DB. On this step, the SWPM will only load the SAP DB (tables, views, SP, jobs) and will not create any extra folders or files. The SWPM will need the path to the SAPMNT share to read and update the DEFAULT profile. After the process is completed, the profile will point to the SQL Server node 1 physical name.

After the DB is created, then the AG can be created and the SAP DB replicated. After the listener created, the profile need to be adjusted accordingly (please see the SAP notes posted here). To help create the AG, please check this link: http://www.brentozar.com/archive/2011/07/how-set-up-sql-server-denali-availability-groups/

The next process is the PAS installation and additional AS if needed. The AS installed on DR will be part of the logon group, but will be parametrized to have very low capabilities ensuring that will not be used on normal conditions.

In case of main DC failure, the VM with the CI is booted on the DR site, the IP address is changed and the DNS record reconfigured.

SCN relevant discussions:

http://scn.sap.com/thread/3676897

[former_member430198]

Hi Manuel,

I just saw your answer in this thread (https://archive.sap.com/discussions/thread/3695808) because right now I am performing a SolMan 7.2 installartion in a MS SQL cluster, and I also have the doubt about how this should be installed. For what you wrote, I understand that  you used Distributed Option for the DB Instance right? In my case we have three nodes in MS SQL cluster, I am executing SWPM in node 3, so could you kindly answer if there is there anything else I have to do in the other nodes? and in the Application Server? How will the application server know where to connect if for any situation other node of the db cluster has to take-over?

Thank you,

Regards

[former_member406028]

Hi manuel.martins ,

I want to setup DR for my current SAP Landscape please find attached my scenario and our current SAP Landscape diagram. I Would highly appreciate if you guide and propose best solution for my SAP Landscape.



 

[mamartins]

From your picture, I see that you are planning only a DR for the ERP.

Because you haven't shared storage, the simple and cheap way to replicate the DB is to use AlwaysON with Availability Groups.

Then you can install a new SAP Application Server on the DR server (Distributed mode on SAP language) and keep it turned off. Or keep it on, but modify the SMLG configuration in order to let the users connect to it only in last resort.

[imaad_vasiq]

Hi Manuel Martins

 

Thanks for sharing useful blog. We are also trying to set up the same scenario where two nodes in primary datacenter running in high availability with shared storage where our ASCS and DB are running in cluster environment. We have third node which is independent running at DR site with independent storage. We have successfully set up the AAG and able to replicate the SQL DB. We have even tested failing over the DB to DR site and making DR site as primary for DB and started SAP application from Primary site (I mean ASCS ,CI,AAS) running from primary and DB alone from DR (since DB failed over to DR).

Our requirement is we need the SAP application server also to operate from DR in case of full disaster at primary. Our issue is we are unable to set up SAP application  server at DR while trying to setup this, the swpm process detects already running ASCS at primary and does not allow proceed further. To achieve this do we need to replicate the drive where our ASCS running to the DR site and then attempt to set up SAP application server( should this be swpm- (Relese> DB> System Copy> Target System> Distributed > Based on ABAP> Primary Applicatin Server Instance)  on DR? Your help in this regard is highly appreciated.

Thanks....