Backing Up and Restoring BIA Data

Use

The BI Accelerator (BIA) is based on TREX technology. InfoCube data from SAP NetWeaver Business Intelligence (BI) is replicated by TREX, indexed, and made available as TREX indexes (OLAP and join indexes) for a high-performance analysis by BI. You need a TREX installation based on a 64-bit architecture in order to use the BI Accelerator. Our hardware partners deliver these TREX variants in preconfigured format as a BIA box.

Caution

Note that you cannot use a TREX installation configured for searching in metadata and documents in a 32-bit architecture for the BIA. A BIA box cannot be used for searching in metadata and documents.

You use Python scripts to save BIA data (TREX indexes and queues) during routine operation and then restore the data offline.

1. Backing Up the Data

Run the following scripts to set the correct environment variables:
- UNIX:
  Bourne shellsh, Bourne-again shellbash, Korn shellksh:
  
  . /usr/sap/<SAPSID>/TRX<instance_number>/TREXSettings.sh
  
  C shell csh:
  
  source usr/sap/<SAPSID>/TRX<instance_number>/TREXSettings.csh
- Windows: c:\usr\sap\<SAPSID>\TRX<instance_number>\TREXSettings.bat
Navigate to the TREX directory containing the Python script exportManager.py for exporting indexes: /usr/sap/<SAPSID>/TRX<instance_number>/exe/python_support.
Execute the Python script for exporting data with the following parameters:

python exportManager.py

Parameters for backing up BIA and TREX data
You use these parameters to back up both BIA data (OLAP and join indexes) and traditional TREX indexes. For more information about these generic export parameters, see Backing Up Data (Online).

--indexId=<my_index_id> (mandatory)

--excludeIndexId=<my_index_id> (optional)

--exportPath=<path_to_backup_directory> (optional)

--useQueueServer=<0|1>(optional)

--force=<0|1>( optional)

--verbose=<0|1>(optional)

--testOnly=<0|1>( optional)
Parameters for backing up BIA data only
You only use the following parameter to back up BIA data (OLAP and join indexes). This parameter only affects the export of OLAP and join indexes. It has no effect on other index types and is ignored if BIA data is not being backed up.

--schemaOnly=<0|1>( optional; only BIA)

Explanation of commands and the exportManager.py parameter for BIA data

Command	Explanation
--schemaOnly=<0\|1> (optional, only BIA)	0 (default): The import behaves as if the parameters --importParts=1 --importSharedParts=0 were set. 1: The import behaves as if the parameters --importParts=0 --sharedParts=0 were set. That is, only the metadata is imported. This parameter has no effect on logical indexes of the type splitted.
-- help	You can use the python exportManager.py --help command to display more information about the script.

Result

The specified index, <my_index_Id>, was backed up to the directory that you specified in --exportPath=<path_to_backup_directory>. If you did not specify a value for the backup directory, the TREX indexes were backed up in the default backup directory, /usr/sap/<SAPSID>/TRX<instance_number>/backup/index.

2. Restoring Data

Make sure that no old TREX files exist in the directory into which you want to import the BIA data. If necessary, delete old files. At file-system level, check the usr/sap/<SAPSID>/TRX<instance_number>/index and usr/sap/<SAPSID>/TRX<instance_number>/queue directories to make sure that all entries really were deleted.
Execute the following scripts to set the environment variables required by TREX and the Python scripts:
- UNIX:
  Bourne shellsh, Bourne-again shellbash, Korn shellksh:
  
  .\usr\sap\<SAPSID>\TRX<instance_number>\TREXSettings.sh
  
  C shell csh:
  
  source usr\sap\<SAPSID\TRX<instance_number>\TREXSettings.csh
- Windows: c:\usr\sap\<SAPSID>\TRX<instance_number>\TREXSettings.bat
Navigate to the TREX directory containing the Python script importManager.py for importing indexes: /usr/sap/<SAPSDI>/TRX<instance_number>/exe/python_support
Execute the Python script for importing data with the following parameters to restore the backed up BIA data (OLAP and join indexes).

python importManager.py

Parameters for restoring BIA and TREX data
You use these parameters to restore both BIA data (OLAP and join indexes) and traditional TREX indexes. For more information about these generic import parameters, see Restoring the Data.

--indexId=<my_index_id> (mandatory)

--excludeIndexId=<my_index_id> (optional)

--importPath=<path_to_backup_directory>(optional)

--force=<0|1>( optional)

--useQueueServer=<0|1>(optional)

--verbose=<0|1>(optional)

--testOnly=<0|1>( optional)

--updateExport=<0|1> (optional)

--directUse=<0|1>(optional)

--topologyPath=<path_to_the topology> (optional)

--repairSplittedParts=<0|1> (optional)

Parameters for restoring BIA data only

You only use the following parameter to restore BIA data (OLAP and join indexes). These parameters only affect the export of OLAP and join indexes. They have no effect on other index types and are ignored if BIA data is not being backed up.

--schemaOnly=<0|1>( optional, only BIA)

--importParts=<0|1>( optional, only BIA)

--importSharedParts=<0|1>( optional, only BIA)

Command	Explanation
--schemaOnly=<0\|1> (optional, only BIA)	0 (default): The import behaves as if the parameters --importParts=1 --importSharedParts=0 were set. 1: The import behaves as if the parameters --importParts=0 --sharedParts=0 were set. That is, only the metadata is imported. This parameter has no effect on logical indexes of the type splitted.
--importParts=<0\|1>`(` optional, only BIA`)`	0: Indexes belonging to a single InfoCube are not imported. 1 (default): Indexes belonging to a single InfoCube are not imported.
--importSharedParts=<0\|1>`(` optional, only BIA`)`	0 (default): Shared indexes (indexes that are accessed by multiple InfoCubes) are not imported. 1: Shared indexes are imported.
-- help	You can use the python exportManager.py --help command to display more information about the script.

Result

The backed-up data was restored.