HANA Express Out Of Memory Group By Queries

[Former Member]

I am using HANA Express 2.0 in a VirtualBox VM configured with 12GB memory and 8 Cpus. For my stress tests, I use:

• One column table with 30 TINYINT columns and 10M rows
• 8 concurrent threads that connect to HANA using the Python hdbcli.dbapi library
• Each thread executes 200000 queries sequentially, which all have the same structure but involving different columns. These queries involve GROUP BY 3 columns and 3 windows functions.
• Some of these threads end up with this error: (2048, 'column store error: search table error: [1000002] Allocation failed...

I have not been able to complete these tests without these types of memory errors. Could there be any obvious HANA Express configuration settings I missed? Also, I make sure to close every connection and cursor after the results (single row) are fetched. The memory problems remain. Is 12GB memory allocated to the VM too little for querying my less than 0.5 GB table? Does this sound like a memory leak in the database server?

To get to this point, I have learned to overcome several hurdles, but this memory problem does not go away easily. I hope someone has experienced similar problems and found a robust rolustion

[TammyPowlas]

I don't think 12GB is ever going to work in my experience with HANA Express - I think you need at least 16GB RAM, if not more.