Limiting the Amount of Memory Available to a Query: Quick Tip

Posted January 28, 2019 by Jim Knicely, Big Data Solutions Architect, Vertica

Designers discussing project in office sample library
Every Vertica DBA has seen a user execute some crazy, memory intensive query that starves every other user request forcing them to queue. To avoid that situation, you can use the MAXQUERYMEMORYSIZE resource pool parameter, which sets the maximum amount of memory that the pool can allocate at runtime to process a query.


User JIM likes to run CROSS JOINS on tables with a lot of records which requires a significant amount of RAM. Let’s stop him from doing that! dbadmin=> CREATE RESOURCE POOL all_users MAXMEMORYSIZE '2G' MAXQUERYMEMORYSIZE '400000K'; CREATE RESOURCE POOL dbadmin=> ALTER USER jim RESOURCE POOL all_users; ALTER USER dbadmin=> \c - jim You are now connected as user "jim". dbadmin=> SELECT * FROM big_fact2 a CROSS JOIN big_fact2 b CROSS JOIN big_fact2 c; ERROR 8722: The minimal memory required by the query [497460 KB] exceeds the query cap size [400000 KB] HINT: Increase MAXQUERYMEMORYSIZE parameter of the 'all_users' resource pool Note that he’s still able to run queries that use less memory than the value of MAXQUERYMEMORYSIZE. dbadmin=> SELECT COUNT(*) FROM big_fact2; COUNT ----------- 200000000 (1 row) dbadmin=> SELECT COUNT(*) FROM big_fact2 WHERE c1 = 'AAAAC'; COUNT ------- 15 (1 row) Helpful Link:

Have fun!