A large number of companies are using high-availability (HA) cluster solutions to ensure the increase in uninterrupted mission-critical SAP usage. A reference architecture specified by SAP now defines clustering guidelines, which indicates a trendsetting standardization. The HA extension by SUSE Linux has already been certified in the SAP Linux Lab Walldorf.
If an outsider hears the description "SAP Netweaver High Availability Cluster 730 Certification" for the first time, this will mean little to them. However, the significance behind this SAP certification for high-availability cluster solutions
is in fact a trendsetting development.
And there's more. In a sense, it represents a game changer for the structure, functions, and usage of HA clusters in the SAP environment. This is because SAP is providing a cluster reference architecture, for the first time, includes important clustering guidelines, solution providers are required to fulfill.
Formerly, each provider cooked up their own "cluster soup". A standard? Not really - which surely also provides user-side opinion leaders in the world of clustering with enough fodder to give a wide berth to the use of SAP high-availability cluster solutions, or to postpone its implementation indefinitely.
You don't need to be a prophet: The SAP Netweaver High Availability Cluster 730 certification will certainly provide a significant boost for HA Linux cluster solutions in the SAP environment. This is due to the recent SAP standardization; but it is also because the reference architecture means that the structure and use of HA cluster solutions are more easily represented, are faster to implement, and above all, are associated with cost efficiency benefits.
An HA Linux cluster solution of this type, such as the already SAP-certified SUSE HA extension as part of the "SLES for SAP Application Priority" package, covers all central monitoring, messaging, and cluster resource management functions, is managed and controlled using SAP Solution Manager, and has a sophisticated support service. In fact everything from a single source.
In addition, this type of solution should support physical AND virtual environments (heterogeneous clusters) in a wide-ranging, practice-oriented form. In the case of the SUSE HA Extension, SUSE (together with VMWare) has developed a fully automatic process, which monitors and controls all processes. The SUSE Linux HA extension performs monitoring of the SAP applications, and VMware HA controls the hardware outage.
Through the complete independence of the server levels in the cluster, it is possible to move virtual machines from one physical host to another at runtime. This eliminates even the short downtime that occurs in redundant servers in a redundant design in the event of a hardware failure, and takes place even before the other server begins its tasks.
Also important to know and fact: The first step to the optimized solution based on fully independent levels is usually the relocation of important SAP applications, such as the central instance or database, to different physical hosts. This migration is also a powerful HA cluster solution thanks to the complete separation of the virtual and physical level at run-time. Through this measure alone, SAP workloads are already better protected against a possible hardware failure. The SUSE Linux HA extension is also responsible for the monitoring of these functions in an environment such as this. If a message is received from an inactive virtual machine, this is automatically relocated to another host.
If, for example, the database is now moved and as a result, a very high load arrives at the new database, you are using the SUSE HA Extension, this is also registered and the application server is moved (again) to a different physical server. In this way, the HA cluster solution provides a meaningful distribution of SAP workloads, while VMware HA is responsible for monitoring and control of the hardware and the virtual machines containers. Since all the components of the cluster solution communicate with each other across all providers, the smooth operation of all processes is ensured.
One crucial point as to why the new reference architecture is associated with a kind of paradigm shift, is the increased user convenience. Or in other words, the complexity that is often proffered in the field as an argument against HA cluster solutions has imploded.
While it is still recommended that more complex clusters are designed by an experience system architect, basic setup tasks can now be performed by virtually any IT administrator without detailed knowledge of Linux clustering. At least, the required templates and wizards to enable this are provided within a modern and powerful solution.