Error while runnibg reports with BIA on cube having HUGE DATA

[Former Member]

Hi friends,

I have indexed two invetory cube for BIA. out of which one contains low data, and another one contains enormously large data. While excuting the report on the small data cube iam able to get the query result.

But one excuting query on large cube, iam getting the error. like:

"""error executing physical plan: attribute enginee no: 6952.

Error executing physical plan: AttributeEngine: not enough memory.;in executor::Executor in cube: grq_zinv_cn2 .

The following error 6952 occurred in BIA server

Amount of data to be read from the BIA server is too large

Error reading the data of InfoProvider ZINV_CN2$X

Error while reading data; navigation is possible """".

As iam new to this BIA, i couldnt able to predict where the problem lies. Help me out to resolve this error.

Regrds,

Balaji.

[former_member184494 · Accepted solution]

Balaji,

What is your Revision number for BIA ..? and your BI sp levels..

also there are some OLAp parameters you can change to make the query execute but then ideally you would have to log an OSS for the same so that the information about changing your OLAp parameters can be given to you ...

Also check OSS notes - there are notes on how this can be handled...

[Former Member · Accepted solution]

Hi Arun varadha rajan, thanks for quick replay,

Can you please tell me how to check the Revision no?, the SP level is SP 16 and release 7 for basis, and sp 8 for BW.

Can you provide any oss notes for the refference.

Regards,

Balu.

[former_member184494 · Accepted solution]

Revision number is your TREX revision number .. you should be able to find it in RSDDBIAMON2 I guess .. some time since I have seen BIA screens...

[Former Member · Accepted solution]

Hi ,

I didnt find ant option in RSDDBIAMON2, to find the bia trex revision number,

Can any body help me out, how to find the trex revision number.

Regards,

Balaji

[jgleichmann · Accepted solution]

Hello,

to find the REV number got to transaction TREXADMIN -> overview: TREX-Version

Please open also the standalone tool TREXAdmin on OS, than go to Index: Usage -> Access Statistics Group Hosts, check the Memsize.

Check under "services -> load" the load at the time of the error.

Please tell us more details from your system landscape (blades, memory etc...).

Best Regards,

Jens

[Former Member · Accepted solution]

Hi friends,

iam using the trex ree 49, but even iam getting the same error.

in the SAP Note 1002839, it is mentioned that

"You can customize the maximum size of the overall resultset of a BIA query in the TREX configuration file TREXIndexServer.ini (see SAP_RETRIEVAL_PATH) for each TREX index server of your BIA blade landscape."

here my question is SAP_RETRIEVAL_PATH, and wr can i check this to set the maximum size of the result set.

One more question, the report is on multi provider, which contains 4 cubes. But only one cube is BIA Indexed. will it make this issue.

Regards,

Balaji.

[jgleichmann · Accepted solution]

Hi,

goto transaction trexadmin -> ini files, than select the TREXIndexServer.ini from the server which contains the index (backup and master or change all servers which is recommended )

the standard config looks like this:

[ olap ]

dim_fn_cache_size = 1000000

max_bits_without_hash = 22

max_buckets = 4500000

max_bucket_array_size = 500000000

max_result_size_one_index = 1300000

max_column_number_one_index = 3

max_result_size_merge = 3000000

max_column_number_merge = 3

arb_expand_multi_values = on

chunk_size = 0

here add the line:

max_cells_one_index = 40000000

save the files and restart the IndexServer service.

Best Regards,

Jens

Edited by: Jens Gleichmann on Mar 12, 2009 3:36 PM

[Vitaliy-R · Accepted solution]

You mentioned both SP 16 and SP 8. So, which one do you actually have? (I suppose SP 16, because it is difficult to find someone on so early service pack as 8, unles it is some test system).

If what Jens recommend does not work for you, then either your BIA not sized for this query, or data is not well distributed across blades (you need to reorg the data then), or query is returning too many records and cannot be really executed in the accelerator.

Regards,

-Vitaliy

[former_member418309 · Accepted solution]

Hi Balaji,

Yiu can change the settings as mentioned below.

The setting can be changed as per the below steps mentioned in note

1157582.

1. Start the TREX Administration Tool (transaction TREXADMIN).

(Or start the SAP NetWeaver BIA Accelerator Monitor (transaction

RSDDBIAMON) and navigate to the menue entry: "Goto -> TREX

Administration Tool".)

2. Navigate to the tab "Ini Files".

3. Open the TREX configuration file "TREXIndexServer.ini" by double

clicking it.

4. Change the value of the parameters max_cells_one_index and

max_result_size_merge under section (olap) as suggested in the

note 1002839.

5. Choose "Save to all hosts" to publish the changed TREX configuration

file TREXIndexServer.ini to all blades of your BIA landscape.

Regards,

Saravanan