[MCOL-4819] Request for improved recovery options Created: 2021-07-21  Updated: 2023-12-15

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

Type: New Feature Priority: Major
Reporter: Maria M Pflaum Assignee: Max Mether
Resolution: Unresolved Votes: 1
Labels: None

Issue Links:
Relates
relates to MCOL-5102 mcsRebuildEM - Triage - Failed to ope... Closed

 Description   

Other Ideas for different product enhancements/feature request based on customer experience.

  • Rebuild EM doesnt exist in CS 5- currently only works in CS 6+ and has no official documentation
  • Rebuild EM should work with S3 clusters - currently only works with single node deployments if it has access to all dbroots (meaning non-ha clusters is complicated to use or not possible)
  • make periodic copies of the BRM_saves_journal - Brm_saves should keep A & B copies of the BRM_saves_journal to reduce the timeframe of dataloss and/or load_brm should attempt more auto recovery options with its own saves A & B files
  • maxscale should work with cmapi- Putting a node in maintenance mode in maxscale should work with cmapi or cmapi should notice the maintenance flag and safely remove the node from the columnstore cluster
  • debug utility to cleanup extent map (BRM_saves_em) - a debug utility should exist to find orphaned objects/ extent files to either delete them

Current Rebuild EM Notes
Internal Link: https://mariadbcorp.atlassian.net/wiki/spaces/Support/pages/1576632337/Rebuild+the+extent+map


Generated at Thu Feb 08 02:53:14 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.