[MCOL-4917] Certain Extent Map operations doesn't scale well. Created: 2021-11-08  Updated: 2023-11-17  Resolved: 2022-05-16

Status: Closed
Project: MariaDB ColumnStore
Component/s: cpimport, ExeMgr, PrimProc
Affects Version/s: 5.6.2, 5.6.3, 6.1.1
Fix Version/s: 5.6.5, 6.4.1

Type: New Feature Priority: Critical
Reporter: Roman Assignee: Roman
Resolution: Done Votes: 0
Labels: None

Issue Links:
PartOf
includes MCOL-5089 Merge RBTree-based Extent Map with EM... Closed
includes MCOL-5090 Up-merge EMIndex + RBTree-based EM in... Closed
includes MCOL-5091 Up-merge RBTree-based EM into develop Closed
is part of MCOL-5313 Re-test new Exttent Map implementation Closed
Relates
relates to MCOL-4988 Table lock remained after DML failure... Closed
Sprint: 2021-14, 2021-15, 2021-16, 2021-17

 Description   

There is a major deficiency in parts of EM code that makes the operations slower. We need to profile and introduce lookup structures to speed up read/write ops in EM. ExtentMap::getDbRootHWMInfo() must be overloaded to serve multiple oids at a time. ExtentMap::setLocalHWM is called multiple times compacting journal into EM on MCS shutdown(save_brm presumably). setLocalHWM is utterly slow that forces systemd to kill save_brm on timeout.



 Comments   
Comment by Roman [ 2022-05-16 ]

All future work about merging MCOL-4912 and MCOL-4917 belongs to MCOL-5089.

Generated at Thu Feb 08 02:53:59 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.