MySQL Bugs: #106906: EnableRedoControl doesn't work with more LDM threads than Log parts

Bug #106906	EnableRedoControl doesn't work with more LDM threads than Log parts
Submitted:	4 Apr 2022 16:11	Modified:	5 Apr 2022 6:04
Reporter:	Mikael Ronström	Email Updates:
Status:	Verified	Impact on me:	None
Category:	MySQL Cluster: Cluster (NDB) storage engine	Severity:	S3 (Non-critical)
Version:	8.0.23	OS:	Any
Assigned to:		CPU Architecture:	Any

View
Add Comment
Files
Developer
Edit Submission
View Progress Log
Contributions

Description:
The get_redo_stats routine assumes that the LDM thread have at least one log part. If not it quickly exits upon return since redo_size == 0.

How to repeat:
Execute a benchmark against NDB with more LDM threads than log parts.
The size of the inserts should be around 4x bigger than the REDO log size.
One will run out of REDO log.

Suggested fix:
Ensure that get_redo_stats returns a proper REDO log write speed based on which REDO logs it uses.
Also ensure that it returns the max redo percentage in any REDO log it uses.

Hello Mikael,

Thank you for the report and feedback.

Sincerely,
Umesh