[MCOL-3713] UM join mem overflow sometimes does not abort properly Created: 2020-01-06  Updated: 2020-03-13  Resolved: 2020-02-20

Status: Closed
Project: MariaDB ColumnStore
Component/s: None
Affects Version/s: None
Fix Version/s: 1.4.3

Type: Bug Priority: Major
Reporter: Patrick LeBlanc (Inactive) Assignee: Daniel Lee (Inactive)
Resolution: Fixed Votes: 0
Labels: None


 Description   

I found that the abort triggered by a mem overflow does not always work with the new multithreaded join code. The session will hang, and the memory used up to that point for that query will never get deallocated.



 Comments   
Comment by Patrick LeBlanc (Inactive) [ 2020-01-07 ]

marked it in testing, should be in review.

Comment by Patrick LeBlanc (Inactive) [ 2020-01-08 ]

Added a test to the regression suite. https://github.com/mariadb-corporation/mariadb-columnstore-regression-test/pull/174

PR for the engine mods: https://github.com/mariadb-corporation/mariadb-columnstore-engine/pull/1001

Comment by Daniel Lee (Inactive) [ 2020-02-20 ]

Build verified: 1.4.3-1 Azure build (20200220)

verified on a 10gb dbt3 databasse.

set infinidb_um_mem_limit=10;
set infinidb_ordered_only=ON;

repeated query 10 times

select count from orders, lineitem where l_orderkey = o_orderkey;

Generated at Thu Feb 08 02:44:53 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.