Error when querying a SAC story that consumes data from an analytical model in SAP Datasphere

[miquelfornieles]

Hello everyone,

We are currently experiencing issues running SAC stories that consume SAP Datasphere analytical models. This issue has arisen during this summer. We have already created a SAP CASE, but we haven’t received any solution yet . For now, expanding the SAP Datasphere parameters is ruled out due to its extra cost.

We would appreciate any help to unblock this problem. The issue is that the execution time of SAC stories has increased significantly (there hasn’t been any notable increase in data), and in the worst cases, the report execution ends with a timeout (we are aware of the 3-minute limitation).

What concerns us is that the SAC report only displays a table with about 20,000 records. The analytical model consumes a view that is persisted. Furthermore, there are no other processes running in parallel, and the number of SAC executions is minimal.

The error message is:

"code":"inaTimeout","details":{"stack":"See correlation id ada620af-5346-439b-5037-5292cab57589","message":"execution aborted by timeout, Code: 613, SQL State: HY000","code":613,"sqlState":"HY000","name":"HanaError","originalMessage":"execution aborted by timeout"}}

The note related to SAC Timeout is:

0003584795 - Timeout Issue with Analytic Model in SAP Datasphere

Thanks in advance.

Regards,

Miquel

[Martin_Kuma]

Hi Miquel, 

since the story worked previously and you do not have more data, then the issue should not be in the data self, but one approach would be to suppress zeros, however SAC first loads all data and then applies zero suppresion, so a work-around is a dummy-dimension checking measures (not SAC dimension based on measures, but in DSP self) and suppressing zeros via the dummy-dimension. 

Also try a blank, empty story, where you re-create the table from the AM 1:1 to the original story self. Check if that works. Kind of base-line strip-down. If that works, the issue is in the original story self. Maybe there was an update of the SAC optimizer and you are using a feature which was de-optimized. If that does not work check the SAC Measurement Tool -> Client and Network tests, also do also check the ASSOCs you use, do persist those VIEWS as well (the ASSOC views). 

The SAC statistics can provide insight as well. I mean the background statistics, not the Analyze Story Performance in Runtime (I assume you already checked this one). MDS calls can be seen in the statistics including the SAC and browser time. 

Not sure what steps did you already check, maybe you can list out what / where was already done. 

Cheers, Martin 

[miquelfornieles]

Hello Martin, thank you for your comments. Unfortunately, I have applied your suggestions: taking a snapshot of all analytical model associations in DSP and disabling the zero suppression option in SAC, but I still get the error after the timeout occurs. It’s very strange because this behavior has affected all my SAC reports since summer. I will keep investigating. Regards.

[Martin_Kuma]

Hi, disabling not. Enabling in SAC, but SAC will first load the data and only after apply the suppression, so create a dummy dimension in the view if possible. Anyway my very first suggestion was to create a blank story and create the table there. You did not really respond to this suggestion. Regards

[miquelfornieles]

Hi MArtin, thanks again for your comment. I forgot to mention that we did test creating the SAC report from scratch; however, the issue persists. To display a table with 20k records, it takes almost 3 minutes, and sometimes a timeout occurs. Regards. Miquel

[pratik_doshi2]

Hi,

Had you tried to create a paginated table in SAC?
Maybe it can resolve your problem, please find a link for that:
https://community.sap.com/t5/technology-blog-posts-by-members/how-to-create-a-paginated-report-in-sa...

[miquelfornieles]

Hello, thank you for your suggestion to create table pages, but the same error still occurs. I will keep investigating. Regards. Miquel