This article provides an overview of the Data Semantic Layer (DSL) Bridge used when indexing an Explorer Information space in BI 4.0.
As Business Intelligence 4.0 provides a new semantic layer separating out the data processing layer from the report processing layer we will be confronted with new workflows to analyze. Explorer is no exception as we used the Webi Processing Server to pull data out of UNV files in Business Objects Enterprise XI3.1. Now in BI 4.0, Explorer supports only UNX and pulls data directly from the semantic layer.
The following is a sample of command line arguments of an Explorer Indexing Server:
-Djava.awt.headless=true -Dosgi.configuration.area=configuration.EXPL40.ExplorerIndexingServer -server -XX:MaxPermSize=256m -Dosgi.clean=true -Xmx1g -jar plugins/org.eclipse.equinox.launcher_1.0.101.R34x_v20081125.jar -loggingPath "C:/Program Files (x86)/SAP BusinessObjects/SAP BusinessObjects Enterprise XI 4.0/logging/" -connectivity_dir %CONNECTIONSERVER_DIR% -serverkind explorerIndexing
The Data Semantic Layer is an integral part of the Indexing Server and has an important parameter: -connectivity_dir %CONNECTIONSERVER_DIR% where it loads connectivity libraries from. A typical location is: "C:/Program Files (x86)/SAP BusinessObjects/SAP BusinessObjects Enterprise XI 4.0//dataAccess/connectionServer"
So we are able to use UNX without going via the DSL Bridge Service (part of the Adaptive Processing Server), which greatly reduces the list of servers required to the following:
* If you like to export a result of an Information Space to Webi for example, you will need to also start the Adaptive Processing Server (hosting the REBeanServer) and the WebIntelligence Processing Server.
What do we find in the dataAccess directory?
We find the same data access libraries as used by the DSL Bridge. These libraries will be used by any connection created in the Information Design Tool.
What do we not find here?
Connection libraries which are loaded via the CMC Advanced Parameters set in the Explorer Application Settings. For example a single HANA or BWA (TREX) Connection.
In addition we have the Excel Dataprovider which is also part of the native libraries and does not require the "-connectivity_dir". These drivers are loaded from the Explorer plugins directory: "C:\Program Files (x86)\SAP BusinessObjects\Explorer14.0\plugins"
"Out of memory"
The default setting on the command line parameters of the Java Heapsize is 1GB i.e. -Xmx1g. The 1GB value is a minimum and for most customers not enough. We have to keep in mind that this memory will be used for data retrieval and for indexing.
Even with 8GB set, it is unlikely that the indexing process will succeed for more than 3-4 Million rows (a better guideline is cells = rows x column), as after the indexing process it needs to distribute the finished index file across to the other Explorer servers. The copying process to the other servers fails if the generated index file is too big.
1662670 - "Failed to retrieve the data source details. The creation of the data source object tree failed." For any new Information Space with UNX universe
1577353 - SBOP Explorer 4.0 SP1 Sizing and Performance information