image_pdfimage_print

This blog is about one of the most interesting PoC where the customer wanted to protect their mission-critical SAP HANA®  instance using both SAP HANA system replication and Pure Storage®  ActiveCluster. The customer wanted to check the log latency for SAP HANA primary system using SAP hardware configuration check tool with both system replication and Active cluster turned on.

Pure Storage® Purity ActiveCluster is a fully symmetric active/active bidirectional replication solution that provides synchronous replication for RPO zero and automatic transparent failover for RTO zero. ActiveCluster spans multiple sites enabling clustered arrays and clustered hosts to be used to deploy flexible active/active datacenter configurations.

With this kind of double protection, if SAP HANA system replication fails due to some reason and primary SAP HANA crashes, the Pure Storage ActiveCluster will make sure your data is synchronously replicated safely to the secondary Pure Storage FlashArray volumes. In this scenario, these volumes (Data, log and shared volumes) replicated using ActiveCluster can be mounted to recover the secondary SAP HANA instance.

The idea about this PoC is to test the log latencies for SAP HANA when with primary instance enabled for both Pure Storage ActiveCluster and SAP HANA system replication at the same time.

In order to test it:

  • SAP HANA System replication was set up between SAP HANA instances
  • A custom Java multi-threaded load generator was connected to SAP HANA via JDBC. This was populating primary SAP HANA with 12000 records per minute on to 10 tables with size 500 varchars long. This generates completely random data.
  • Pure Storage ActiveCluster was set up for SAP HANA Data, log, shared volumes were stretched to the secondary array with 0ms RTT.
  • SAP HANA HWCCT load generator was running on the primary instance to measure the log latencies.

activecluster and system replication.png

As shown above, the Java load generator was committing 12000 records per minute on to 10 tables. This is getting replicated to the secondary SAP HANA system via SAP HANA system replication. In parallel SAP HANA HWCCT load generator is measuring log latencies on the primary system. HWCCT very long runs were configured to measure the log latencies for the SAP HANA primary system.

The CPU consumption of SAP HANA from Java load generator and system replication was around 5-9%.

The results were quite surprising, the log latencies were still meeting well within the  SAP HANA TDI KPIs with both system replication and Pure Storage ActiveCluster running in parallel.

“For more information on SAP and Pure Storage ActiveCluster, click here.”     (Hyperlink to  https://www.purestorage.com/content/dam/purestorage/pdf/whitepapers/pure-storage-activecluster-with-sap-hana.pdf)