Using the cube merge functionality, intelligent cubes can be published on a separate Strategy Intelligence Server outside of a standalone or clustered Strategy production environment and then moved to the production environment.
The cube merge utility is available on Strategy Intelligence Server 9.4.1 and above on Windows and Linux operating systems.
Note: The utility is not available on Strategy 9.5.0 GA.
Important Notes:
- To ensure integrity of the intelligent cubes, make sure the metadata copies on the source and target environments are exactly same.
- Using the tool requires both the source and goal project(s) to be unloaded.
- Once the merge is complete, the cube index files of the target environment (for example, production) are to be replaced with the merged files.
The cube merge feature is available through the CacheParser tool. On Linux operating systems, it is called 'mstrcacheparser' and is located in the bin folder of the Strategy home folder (along with other Strategy tools and utilities) and on Windows, it is called 'CacheParser' and is located in the Program Files (x86)/Common Files/Strategy folder. See the example command at the end of the document for more information on how to run the tool.
Steps:
- Set up two Strategy Intelligence Servers: S1, S2 (S1 can be a production server and S2 a development server). S1 is the destination and S2 is the source where the cubes are republished.
- Publish intelligent cubes C1-C5 on server S1 (all cubes are initially published on S1).
- Publish intelligent cubes C4, C5 on server S2.
- Unload the required project on server S1(assuming intelligent cubes C4 and C5 belong to the same project P1).
- Run the cube merge tool to merge cube files on S1 and S2 and the output is created on some folder, say F1 (the merge tool output only contains the index files)
- From the intelligent cube folder on S1 delete all idx files for project P1 ONLY. Also delete .cube, .info files for cubes C4, C5 only. Keep the remaining cube files. Manually copy .cube, info.cube files of intelligent cubes C4, C5 from S2 to S1.
Note: If there is a case in which S2 has new intelligent cubes which previously did not exist on S1, then the .cube and the info.cube files of those intelligent cubes should be copied from S2 to S1 as well. The index files will be taken care of by the merge tool.
- Copy the contents of folder F1 (index files) to the S1 cube directory for that project P1.
- Load project P1 on Strategy Intelligence Server S1.
- Activate the intelligent cubes and load them.
- Run view reports on server S1. They will now use the new intelligent cubes which have been copied from server S2.
Cube merge tool command usage:
Usage is as follows:
Windows - Located in the <install drive>\Program Files (x86)\Common Files\Strategy folder. The utility needs to be run from command line.
CacheParser -m:merge -u:cube_directory -o:out_file
Linux - located in the
bin folder of the Strategy home folder. The utility needs to be run from a terminal window.
mstrcacheparser -m:merge -u:cube_directories -o:out_file_directory
The -m: merge option merges cube index files from multiple directories into one. -u must specify input directories separated by a semicolon, -o must specify a directory other than any of the input directories.
Windows
CacheParser -m:merge -u:"c:\cube1";"c:\cube2" -o:"c:\cube0"
Linux
./mstrcacheparser -m:merge -u:'/home/cube1;/home/cube2' -o:/home/cube0
The command takes cube index files from folders cube1 and cube2, merges them, and writes the merged index files to the folder cube0.
Note1:The 9s Strategy installation currently does not include the Cube Parser utlility.
Note2: You do not need to merge .idx files from multiple nodes if all of cubes (.cube files) exist in one node. In such case, just copy everything from the node, and put that in the corresponding folder of the destination server.
