Chapter 4. Testing the setup
Before the HA cluster setup is put into production, it needs to be thoroughly tested to verify that everything works as expected and also to allow the operators to get experience with how the HA cluster will behave in certain situations and how to bring the setup back to a healthy state in case a failure occurs.
At least the following tests should be carried out:
Perform a manual move of the primary SAP HANA instance via HA cluster commands.
Expected result: a takeover should be triggered on the SAP HANA side, promoting the secondary SAP HANA instance to become the new primary SAP HANA instance. Depending on the configuration of the
AUTOMATED_REGISTER
parameter of theSAPHana
resource, the HA cluster will either register the former primary instance, as the new secondary automatically, or an operator will have to determine what should happen with the former primary instanceCrash the HA cluster node where the primary SAP HANA instance is running.
Expected result: the HA cluster node should be fenced and a takeover should be triggered on the SAP HANA side, promoting the secondary SAP HANA instance to become the new primary SAP HANA instance. Depending on the configuration of the
AUTOMATED_REGISTER
parameter of theSAPHana
resource, the HA cluster will either register the former primary instance as the new secondary automatically, or an operator will have to determine what should happen with the former primary instance.Manually stop primary SAP HANA instance outside of HA cluster.
Expected result: a takeover should be triggered on the SAP HANA side, promoting the secondary SAP HANA instance to become the new primary SAP HANA instance. Depending on the configuration of the
AUTOMATED_REGISTER
parameter of theSAPHana
resource, the HA cluster will either register the former primary instance as the new secondary automatically, or an operator will have to determine what should happen with the former primary instance.Crash the node where the secondary SAP HANA instance is running.
Expected result: the HA cluster node should be fenced and the secondary SAP HANA instance should be started when the HA cluster node is back online and SAP HANA System Replication should resume.
Manually stop the secondary SAP HANA instance outside of HA cluster
Expected result: the secondary SAP HANA instance should be restarted by the HA cluster
Disable the network connection used by SAP HANA System Replication
Expected result: the HA cluster should detect that a failure of SAP HANA System Replication has occurred, but should keep the SAP HANA instances running on both nodes