PrimProc does not resume memory after query execution and is causing subsequent MCS restarts (MCOL-2130)

[MCOL-2148] apply the jemalloc patch to dev 1.1.7 Created: 2019-02-08  Updated: 2019-02-12  Resolved: 2019-02-12

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

Type: Sub-Task Priority: Major
Reporter: Zdravelina Sokolovska (Inactive) Assignee: Zdravelina Sokolovska (Inactive)
Resolution: Done Votes: 0
Labels: None


 Description   

apply the jemalloc patch to dev 1.1.7



 Comments   
Comment by Zdravelina Sokolovska (Inactive) [ 2019-02-12 ]

applied manually end check with intensive querying
ExeMgr is returning memory after operation but PrimProc is not resuming memory after operation. PrimProc is returning memory very slowly when other columnstore processes need more memory allocation.
As a result if swap memory is enabled are still possible occasional MCS system resets due to Columnstore predefined Columnstore.xml configuration in ModuleSwapCriticalThreshold(s)

Generated at Thu Feb 08 02:34:06 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.