SSAS 2016 Question about whether if there is a partition level failure level (ETL) for whatever reason will it take down the cube in its entirety?
We are using SSAS 2016 Enterprise
We are building a mega cube with multiple companies data and we want to partition by Company.
We also will want to schedule separate ETL schedules for each company.
Question is in the event there is failure from a partition perspective will that take down the entire cube? In the event the data from one company goes wonky?
We are building a mega cube with multiple companies data and we want to partition by Company.
We also will want to schedule separate ETL schedules for each company.
Question is in the event there is failure from a partition perspective will that take down the entire cube? In the event the data from one company goes wonky?
ASKER
I researched.
As long as the processing command is wrapped in a transaction, the cube will remain in a good/query’able state should data from one company go wonky and cause a processing job to fail.
https://docs.microsoft.com/en-us/sql/analysis-services/multidimensional-models/processing-options-and-settings-analysis-services?view=sql-analysis-services-2017
https://docs.microsoft.com/en-us/sql/analysis-services/multidimensional-models/batch-processing-analysis-services?view=sql-analysis-services-2017
in terms of Proactive caching does this still hold true?
As long as the processing command is wrapped in a transaction, the cube will remain in a good/query’able state should data from one company go wonky and cause a processing job to fail.
https://docs.microsoft.com/en-us/sql/analysis-services/multidimensional-models/processing-options-and-settings-analysis-services?view=sql-analysis-services-2017
https://docs.microsoft.com/en-us/sql/analysis-services/multidimensional-models/batch-processing-analysis-services?view=sql-analysis-services-2017
in terms of Proactive caching does this still hold true?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
Thank you :)
It really depends upon the type of error you receive on your partition or cube level and the type of Error configuration you have configured at the Cube/partition or dimension level.
More info about Error configuration below..
https://docs.microsoft.com/en-us/sql/analysis-services/multidimensional-models/error-configuration-for-cube-partition-and-dimension-processing?view=sql-analysis-services-2017
>> In the event the data from one company goes wonky?
Yes, if ignore the error at partition level, then the data available for that error ed partition company might not be correct. It can be abnormal..