What?

SQL Server 2005’s Analysis Services uses a process titled Msmdsrv.exe
The memory used by this process does not decrease as expected and results in error on deployment. The error usually is
“Server: The operation has been cancelled”

Why?

Memory utilization by Analysis Services increases when performing process operation, typically to the type of the object being processed.
a)    Dimension:
a.     Memory increase is approximately the size of the dimension plus Read-ahead and the process buffer settings for AS.
b.    Returns to the pre-processing level, for pre-processed dims. For the first time processing, memory usage increases by size of dim that is loaded in the server memory space.
b)     Cube:
a.    Memory increases by the size of the read-ahead and the process buffers.
b.    Memory usage returns to pre-processing levels after processing completes.

This problem occurs when the memory usage of the Analysis Services Server (Msmdsrv.exe) does not return back to the pre-processing levels.

How?

This issue can be resolved by deleting the unused databases and cubes that are saved. Also what helps is choosing  the processing option as “Full” instead of Default. Default processing is usually incremental update and when the file store of the attributes reach 4GB, error message are issued while processing cubes/dimensions in AS.

Possible HotFix :
http://support.microsoft.com/kb/912016/

So?

It is safer to retain just the databases and cubes that are currently being used and avoid unnecessary data dumps. Also, though incremental processing is fast, Full processing done saves memory usage

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s