AS2005 Cube Processing

I have problems with processing cubes on AS2005. I get this error message while processing:

Memory error: Allocation failure: Not enough storage is available to process this command. Error in the OLAP storage engine.

This machine has 6GB of memory and once it hits 3.5GB the processing stops. AWE is turned on.

I would appreciate, if anyone can help on this.

Thanks,

Gopal

Analysis Services is not AWE aware. So it can only use 2GB on a 32bit box (3GB if you have /3GB switch in boot.ini). Looks like the processing operation is running out of memory.

What are the number and sizes of dimension and partition tables? Do you have aggregations? Look at the processing log (in the process dialog) and find out what is the operation that causes out of memory - dimension processing, partition processing, building aggregations, etc.

This could also happen due to excessive parallelism. Try processing the dimensions and partitions individually.

|||

Thanks for your reply.

I managed to process the partitions and dimensions individually. That worked.

Gopal

|||

Hi.

Just a short question:

Is it really true that SSAS can use only 2GB RAM on a 32bit system? The technical reference for SQL Server 2005 Standard and Enterprise states that SQL Server can use the amount of RAM supported by the operating system which is 4GB for a Server 2003 Standard edition and 32GB for a Server 2003 Enterprise edition, both on 32bit systems. Does this apply only to the SQL Server Database Engine then and not the Analysis Services Server?

Regards

Kjetil

T.K. Anand wrote:

Analysis Services is not AWE aware. So it can only use 2GB on a 32bit box (3GB if you have /3GB switch in boot.ini). Looks like the processing operation is running out of memory.

What are the number and sizes of dimension and partition tables? Do you have aggregations? Look at the processing log (in the process dialog) and find out what is the operation that causes out of memory - dimension processing, partition processing, building aggregations, etc.

This could also happen due to excessive parallelism. Try processing the dimensions and partitions individually.

|||

In the 64-bit versions of Analysis Services...has this memory use limit been removed?

...cordell...

|||Yes - 64bit version doesn't have these limitations.

Arithmetic operation failure. Is this a bug, or am I doing something wrong?

I used the following code in a function. The @.infloat is input parameter

For every other number than 16.31, 17.31, 18.31, 19.31 and 20.31 (at least those numbers I checked) the code runs ok. For those numbers the result is wrong.

What I want to do is return a ten digits number left filled with zeros, with the last two digits consitered as decimals.

What I do is multiply the given float with 100.00, and take the resulting integer

But when @.infloat is multiplied with 100.00 and @.infloat in (16.31, 17.31, 18.31, 19.31 and 20.31)

the result is (1630,1730,1830,1930,2030) instead of (1631,1731,1831,1931,2031)..

Can anyone try this ?

declare @.ingood int,

@.stuffme varchar(10),

@.infloat float, @.infl float

set @.infloat=18.31

set @.infloat=isnull(@.infloat,0)

set @.stuffme='0000000000'

set @.infl=@.infloat*(100.00)

select @.infl as infl

set @.ingood=cast(@.infl as int)

select @.ingood as ingood

set @.stuffme=stuff(@.stuffme,11-len(@.ingood),len(@.ingood),@.ingood)

select @.stuffme

Replace

set @.ingood=cast(@.infl as int)

WITH

set @.ingood=cast(@.infl as NUMeERIC (10,2))

|||Do you need high precision ? Floating point data is approximate; not all values in the data type range can be precisely represented.

See here :
select convert(float, 18.31)
Result : 18.309999999999999|||

Hi

I have tried MONEY and NUMERIC datatypes for @.infloat and @.infl from your example. They both work just fine.