High Availability, Backup and Disaster Recovery for SAP Landscapes

Keep this: job aid

HA, backup and DR sign-off card

• HA = component failure nearby, near-zero loss. DR = site or region lost, some loss accepted. Backups = logical errors and corruption.
• Single points of failure: ASCS (cluster with ERS), application servers (N+1), SAP HANA (system replication), storage and network.
• HSR modes: SYNC (secondary disk), SYNCMEM (secondary memory), ASYNC (no wait). logreplay = hot standby; logreplay_readaccess = read-enabled.
• Takeover moves the database role; applications, virtual names and interfaces must follow.
• RPO = point to which data must be recovered; RTO = tolerable time in recovery (NIST SP 800-34). Set them per process with owners.
• Check targets against the RISE contract: HA standby if contracted; DR optional; customer DR procedures are the customer's.
• DR test under RISE: SAP executes failover on request (one per year included in the optional service); customer runs functional tests; test postings are lost.
• Success criteria: time, data (known last transaction), processes, interfaces, output, jobs; record timestamps.
• Report: verdict, targets versus achieved, timeline, findings with owner and date, evidence.
• Sign-off: Basis, test lead, process owners, program lead; retest on a cycle and after major changes.