Manual System Failover For ActiveCluster

ActiveCluster
Audience
Public
Product
FlashArray
FlashArray > Purity//FA
Source Type
Documentation

The SAP HANA environment is configured as follows:

  • 2 separate distinct SAP HANA Systems.
    • Compute 1 and Compute 2
    • Each has a separate volume mounted to /hana/shared from the localized array. i.e /hana/shared for compute 1 comes from array 1.
    • The exact same SID is set on both systems either at installation time or post install.
    • The exact same version of SAP HANA is installed on both systems.
  • 2 separate distinct FlashArray storage platforms.
    • Array 1 and Array 2
    • Less than 1ms round trip latency between the arrays for the replication data path.
    • The arrays can be of a different model - but the number of supported nodes is established by the lower FlashArray model.
    • The volumes on both arrays in the SAP-HANA-POD are connected to both the initiators for Compute 1 and Compute 2, but only mounted to one system at a time.
    • The preferred path the volumes on Array 1 are to the initiator for Compute 1. The preferred path for the volumes on Array 2 is to the initiator for Compute 2.
  • Initiators on Compute 1 and Compute 2 have access to both Array 1 and Array 2.

  • 3 separate distinct SAP HANA Systems
    • Compute 1, Compute 2 and Compute 3
    • Each has a separate volume mounted to /hana/shared from the localized array. i.e /hana/shared for compute 1 comes from array 1.
    • The exact same SID is set on all systems either at installation time or post install.
    • The exact same version of SAP HANA is installed on each system.
  • 3 separate distinct FlashArray storage platforms.
    • Array 1, Array 2 and Array 3
    • Less than 1ms round trip latency between the arrays for the replication data path.
    • The arrays can be of a different model - but the number of supported nodes is established by the lower FlashArray model.
    • The volumes in the SAP-HANA-POD are connected to the initiators for both Compute 1 and Compute 2, but only mounted to one system at a time.
    • The preferred path the volumes on Array 1 are to the initiator for Compute 1. The preferred path for the volumes on Array 2 is to the initiator for Compute 2.
  • Initiators on Compute 1 and Compute 2 have access to both Array 1 and Array 2. Compute 3 only requires access to Array 3 but no volumes are connected until a failover occurs to it.