3714096 - SAP HANA scale-out high availability: takeover intentionally prevented due to topology uncertainty (split-brain protection) resulting in both primary and secondary systems being unavailable

• In a SAP HANA scale-out environment using System Replication together with a cluster framework (for example Pacemaker/SAPHanaSR), a failover or takeover attempt does not complete successfully.
• Both the former primary and secondary systems may become unavailable after the incident.
• The database services stop without signs of a crash (no SIGSEGV, assertion failure, OOM kill, or kernel panic in HANA traces).
• HANA traces show internal communication timeouts between hosts or services.
• The nameserver coordinates topology decisions and interacts with the System Replication hook and the cluster framework.
• Customers may suspect a database revision defect or operating system issue, although no product error is recorded in HANA traces.

SAP HANA DB Version 2.0

SAP HANA, System Replication, scale-out, Pacemaker, split brain protection, takeover prevented , KBA , HAN-DB-HA , SAP HANA High Availability (System Replication, DR, etc.) , Problem