Memory allocation failed error

[rindia]

Hi,

I am running the SQL in SQL console (HANA revision 58). After 2 minutes, I am getting the error:

SAP DBTech JDBC: [2048]: column store error: search table error:  [9] Memory allocation failed.

The query is based on joining many HANA tables and views (non-materialized) with concatenation columns and some views has huge data volume.

I am looking for solution to execute the query with sufficient memory allocation. Is there any setting required to do so?

Regards

Raj

[Former Member]

Hi Raj,

Depending on your views will be necessary use variables / parameters to filter on lowest level of data.

If you are conducting development you have also an option to hard code these filters on biggest tables of your model.

Anyhow, this is a thing you must investigate as there's a possibility of your views are avoiding push down of selection filters.

No enough information to judge.

Regards, Fernando Da Rós

[rindia]

Hi Fernado,

I am on support and not in development.

Regards

Raj

[Ravi_Channe]

Hi Raj,

I agree with Fernando. Even my first impression of this post was, "not enough information".

I understand you are currently supporting the HANA application, but you may suggest the user to execute the reports with some additional filters.

I think many of the applications developed so far, were initially developed on the early revisions of HANA. Considering that HANA is relatively new development platform, the applications might not have been tuned properly for the performance. Now with high data volume you may face performance issues. But that doesn't mean, you cannot change the existing application with some Change Request.

As part of the support team, you can perform basic analysis with the query plan and identify the bottlenecks. You are among the very active member on this forum and quite experienced on HANA, . So I don't need to mention here how the basic analysis should be done.

As Fernando mentioned, I think the filters should be pushed down to reduce the result set generated. Unless the intermediate result set and data transfer between the engines is reduced, it would be difficult to get good performance or even query execution on HANA.

You already know quite many things on how to achieve it and I am sure you will figure that out.

Regards,

Ravi