Recovery of SAP HANA database from tape storage fails
Recovery of SAP HANA database from tape storage fails to complete successfully as the restore jobs hang indefinitely. This failure is specifically seen during the logs restore phase.
By default, SAP HANA does not read all the pipes that it has opened during recovery. As NetBackup processes the pipes randomly, it is possible that the pipe where NetBackup tries to write is not open. This situation creates a deadlock situation and hence the jobs are stalled in the Activity Monitor.
To streamline the read and write operations on the pipe, set the following parameter in the global.ini file:
In HANA Studio, double-click on instance tab.
-> -> ->-> enter 1 as the max_recovery_backint_channels value (by default, this value is 64).
This setting ensures that only a single request is sent to NetBackup. Hence only a single pipe is open for reading and NetBackup writes on that pipe only.
Note:
This issue is not seen in case of disk recovery. Hence no change is required for max_recovery_backint_channels in case of disk recovery.