必须在备节点数据库和主节点数据库已经同步完成的条件下，集群的备节点才会接管数据库。

如果数据库未同步完成，主节点数据库故障时，备节点HA不会接管主节点数据库。

1. 执行SAPHanaSR-showAttr命令查看SAP HANA System replication的主备当前状态。

   在主节点执行命令SAPHanaSR-showAttr。

   显示结果如下：

   sync_state：

   PRIM：表示该节点为主节点。

   SOK：表示该节点已经同步OK。

   SFAIL：表示该节点还未同步OK。

   参考打印如下：

   HW00001:/home # SAPHanaSR-showAttr
   Host \ Attr clone_state remoteHost roles           site   srmode sync_state vhost  lpa_sle_lpt
   ---------------------------------------------------------------------------------------
   HW00001 PROMOTED HW00002 4:P:master1:master:worker:master HW00001 sync PRIM  HW00001 1416991408
   HW00002 DEMOTED  HW00001 4:S:master1:master:worker:master HW00002 sync SOK   HW00002 30

2. 如果数据同步未完成时，HA cluster不会自动切换，如需手工切换，请参考SAP notes说明：当所有服务同步状态不都是Active时，SAP不建议进行takeover操作，因为这意味着有数据丢失的风险。

   a、SAP note：2578019 - Service Crashes in DataAccess::PersistenceManagerImpl::endOfDataRecovery

   Do not perform a takeover if not for all services the REPLICATION_STATUS in M_SERVICE_REPLICATION shows ACTIVE. See SAP Note 2063657 for details.

   https://launchpad.support.sap.com/#/notes/2578019

   b、SAP note：2580302 - Emergency Shutdown of Indexserver Due to Log Position Inconsistency Upon Takeover

   With the fix the takeover in this case can succeed, but since one service wasn't in sync this implies data loss. You should follow the takeover decision guide of SAP Note 2063657 to assess if a takeover in this state is a feasible option

   https://launchpad.support.sap.com/#/notes/2580302

   c、SAP note：2063657 - SAP HANA System Replication Takeover Decision Guideline

   是否执行take over，推荐管理员根据实际情况判断。

   https://launchpad.support.sap.com/#/notes/2063657