SLES for SAP Applications 15 is now Certified for SAP S/4HANA and NetWeaver HA-Interface Certification
The SAP Application Server High Availability Interface Certification, a.k.a. SAP Application Server HA-Interface Certification, is to ensure the communication between SAP Applications and 3rd-party cluster solutions. A solution with this certification will avoid a split-brain situation where a cluster will not try to interfere with planned maintenance tasks carried out by the SAP administrators. Without the HA Interface, when an SAP administrator stops an SAP application that’s running in a cluster, the cluster will interpret it as a failure and thus try to remediate. The result is disruptive to the planned maintenance tasks.
The communication is enabled between the HA-interface (cluster API) of SAP Start Service (sapstartsrv) and 3rd party cluster connector, as illustrated in the image below.
SUSE’s Distinguished SAP Architect Fabian Herschel has explained the technical details in this blog: Using sap_vendor_cluster_connector for interaction between cluster framework and sapstartsrv
SLE15 has been certified both NW-HA-CLU_750 and S/4-HA-CLU 1.0 for SUSE Linux Enterprise Server for SAP Applications 15.
There are slight differences between the two certificates:
|Certificate||Certified OS Version||Certified SAP Kernel||Supported Databases||Standalone Enqueue Server||Number of Nodes in the Cluster|
|NW-HA-CLU_750||SUSE Linux Enterprise Server for SAP Applications 15||Netweaver 7.50+||SAP ASE, Oracle, IBM DB2, MaxDB, HANA DB||Standalone Enqueue Server 1 (ENSA1)||Only 2 nodes|
|S/4-HA_CLU 1.0||SUSE Linux Enterprise Server for SAP Applications 15||S/4HANA 1809||HANA DB 2.0||Standalone Enqueue Server 2 with the Enqueue Replicator 2 (ENSA2)||2 nodes or more|
Both certifications are applicable to multiple platforms, from bare metal, virtualized environment, to public cloud, where SUSE Linux Enterprise Server for SAP Applications 15 and High Availability Extension are supported.
Customers now benefit from a certified start/stop infrastructure within the HA setups, can establish a unified management interface, and simplify the management of the maintenance tasks such as a rolling kernel switch (RKS).
Without this HA-Interface, it’s difficult to perform maintenance to SAP instances running inside a cluster. For example, if SAP sys admin needs to restart the SAP ASCS/ERS instances, there are two options to do it:
- Use pacemaker cluster management interface crm or crm_gui to stop or start the instances through the resources.
- Put the cluster into Maintenance mode, then stop and start the instances using SAP MMC/MC, or sapcontrol, or other tools. After maintenance tasks are complete, exit the cluster Maintenance mode.
The problems with either methods are that, in a lot of datacenters, the cluster and SAP instances are managed by different teams, respectively, thus in order to stop/start SAP instances, the SAP sys admin has to go through a complex process to get the support from the cluster administrator. With this HA-Interface in place, when the SAP sys admin stops the SAP instances, cluster will standby and not to interfere. So, it eases the operations.
For example, the ASCS resource, that’s managed by a cluster, has a status “Started”:
Now the SAP sys admin stops the ASCS instance through SAP Management Console:
The ASCS instance is stopped:
In the cluster, the ASCS resource has a “Stopped (disabled)” status because the cluster recognized that it’s an intended operation thus didn’t interfere.
As you can see, because of the HA-Interface integration, by using the same SAP Management Console, SAP sys administrators can independently perform the maintenance tasks that require stop/start the SAP application server instances that are running in a cluster. This feature greatly simplifies the operational process and provide flexibility and stability to the environment.
For details on how to enable the HA-Interface in your cluster, please refer to the following documentations:
- SAP NetWeaver Enqueue Replication 1 High Availability Cluster – Setup Guide for SAP NetWeaver 7.40 and 7.50, section 3.4 https://documentation.suse.com/sbp/all/html/SAP_NW740_SLE15_SetupGuide/index.html
- SAP S/4 HANA – Enqueue Replication 2 High Availability Cluster – Setup Guide, section 3.4 https://documentation.suse.com/sbp/all/html/SAP_S4HA10_SetupGuide-SLE15/index.html
If you have any questions, please feel free to ask questions in the comments section.
Thanks for the post Sherry,
I went through the documentation in the post that explain about the way set-up has to be done.
However, I want to understand how does HA-Interface differentiate the two cases -
1. Planned shut down (requiring the cluster resource to be 'disabled' .
2. Unplanned shutdown(requiring the cluster to remediate).
Do we have any documentation for understanding it. Fabrian's blog captures most of the aspects of the cluster but does not seem to have HA-Interface.
For planned shut down, the command needs to be sent from a client that supports the SAP Start Service (sapstartsrv), e.g. MMC, sapcontrol. Cluster will not interfere.
For unplanned shut down, because it’s not triggered by any client through the SAP Start Service (sapstartsrv), cluster will try to remediate.
As you can see, for any 3rd party toolset, it’s important to integrate with the SAP Start Service (sapstartsrv) in order to coordinate with clusters that are HA-Interface certified.
Am I understanding correctly?
The HA-Interface itself does not distinguish and act differently for planned shutdown (by bringing the cluster in standby) or unplanned one (by triggering remediation). It relies on the kind of the information it gets from the client through sapstartsrv.
That's correct. That's why there is the HA-Interface.
I assume it's only for application server and not for the HANA database?
Hi Bobby, the certification is designed to test Application Servers not the database. The "supported database" column means that the certified Application Server layer is supported on the list of databases, but the databases are not tested using the particular certification tests.
Thank you for clarifying.
I'd be interested on HANA integration as database maintenance in system replication involves many parties and different OS/DB crm_xxx/hdbnsutil commands.