Thanks for the swift reply!
I changed the script accordingly, and now would you be able to point me to the right resources where i can learn setting up ‘proper error handling - save errors/warnings to dts log’ ?

ProcessFull option already processes all data and indexes. If you want to split processing off data and indexes, then insteead of doing ProcessFull you should do ProcessData and then ProcessIndexes.

Please note than in my script oCube.Process and oMeasureGroup.Process lines are commented out and left there just to show that you have an option to do processing at the higher level.
So your script should be:

For Each oCube In oDB.Cubes
For Each oMeasureGroup In oCube.MeasureGroups
For Each oPartition In oMeasureGroup.Partitions
oPartition.Process(Microsoft.AnalysisServices.ProcessType.ProcessData) ‘ Process all partitions
Next
Next
oCube.Process(Microsoft.AnalysisServices.ProcessType.ProcessIndexes)
Next

By the way, it is recommended to do process data and process indexes in the separate statements when you want to monitor if you have too many aggregates and because of that processing takes too long.

I am thinking about processing the indexes after processing the dimensions, measure groups and partitions in the cube. Is it required to do that?

If yes, does the following modified code work ?


For Each oCube In oDB.Cubes
oCube.Process(Microsoft.AnalysisServices.ProcessType.ProcessFull) ' Process all cubes

For Each oPartition In oMeasureGroup.Partitions
oPartition.Process(Microsoft.AnalysisServices.ProcessType.ProcessFull) ' Process all partitions
Next
Next
oCube.Process(Microsoft.AnalysisServices.ProcessType.ProcessIndexes)
Next


By default this package will already process everything in parallel. There is no way to limit how many parallel tasks will run, so it will use all processors available.
If you need to control parallelism, better way would be to build process XMLA script like I did in my post here: http://www.ssas-info.com/VidasMatelisBlog/7_ssis-package-to-process-all-dimensions-in-ssas-2005-db

Can you please give a bit more details why you are running 4 packages in parallel?

I have not implemented proper loging in this package. My goal was just to share code to process objects.

The code above helps a lot, I built the similar code in SSIS package for my project, it runs successfully, but I cannot see the detail processing log information, i.e which dimensions and partitions are processed, is there any log file generated by Analysis manager for this processing?

Thanks,
Emily

I could not find a way to control Parallelism during processing using ExecuteCaptureLog method. I’ll try to research a bit more, and if I’ll find answer, I’ll post it here.

You probably already know that you can control parallelism with XMLA. There is Parallel parameter for that like in this example:

< Batch xmlns="http://schemas.microsoft.com/analysisservices/2003/engine">
< Parallel MaxParallel="4">
< Process xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:ddl2="http://schemas.microsoft.com/analysisservices/2003/engine/2" xmlns:ddl2_2="http://schemas.microsoft.com/analysisservices/2003/engine/2/2">
< Object>
< DatabaseID>Adventure Works DW< /DatabaseID>
< /Object>
< Type>ProcessFull< /Type>
< WriteBackTableCreation>UseExisting< /WriteBackTableCreation>
< /Process>
< /Parallel>
< /Batch>

Vidas

Do you know whether there’s a way to control the parallelism of the partition reprocessing? I’m trying to write a package that will reprocess multiple partitions in one measure group, and I want to regulate the load on the server.

Thanks!

Small possible issue, I think the ExecutionSuccessfull should be commented out, unless failures are always the last item in the results collection. Otherwise a warning may set the bit back to true even if an error did occur but was processed earliar in the collection.

Regardless thanks for the code.

see http://forums.microsoft.com/MSDN/ShowPost.aspx?PostID=1117591&SiteID=1

It looks like in your blog you are talking about special case when during processing you also changing source table of measure group. Is that right?
Have you consider having SQL View that is used as a source for the measure group. Then you would change just view definition (example select from different table) instead of trying to solve this in AMO?

Vidas Matelis

Why do you compile AMO code as dll? Are there any benefits for that?