Disable Autostart for the HDB Instance
In order for LifeKeeper to successfully manage the SAP HANA resource during failover or system reboot, the Autostart feature which automatically starts the HDB instance on system boot must be disabled. In order to disable this feature, edit the instance profile for your HDB instance (typically located at /usr/sap/<SID>/SYS/profile/<SID>_HDB<##>_<HostName>) on all cluster nodes and ensure that the line
Autostart = 0
is present in the profile. Either add or modify this line, as necessary, and save the changes to the profile.
!Warning: If Autostart is not disabled for the HDB instance, then a machine failure of the primary SAP HANA System Replication site will result in a “System Replication split brain” scenario once the original primary node comes back online. In this scenario, the HDB instance is running and registered as primary master concurrently on multiple cluster nodes. As a result, LifeKeeper is unable to determine which site the user intends to be registered as the primary SAP HANA System Replication site. The SAP HANA resource is placed in the OSF (“Out of Service – Failed”) state on the original standby node and a warning message is broadcast to all open consoles until the situation is manually resolved by a database administrator. See Resolving Split Brain Scenarios for more details.
*Note: In order to ensure continued successful management of the SAP HANA resource, LifeKeeper will monitor the value of the Autostart parameter for the HDB instance on all cluster nodes each quickCheck interval (defined by the LKCHECKINTERVAL parameter, default 2 minutes) and will automatically disable Autostart if necessary.
Post your comment on this topic.