[MXS-4911] Drop use of embedded server. Created: 2023-12-27  Updated: 2024-01-02

Status: In Progress
Project: MariaDB MaxScale
Component/s: QueryClassifier
Affects Version/s: 23.08
Fix Version/s: 23.08

Type: Task Priority: Major
Reporter: Johan Wikman Assignee: Johan Wikman
Resolution: Unresolved Votes: 0
Labels: None

Sprint: MXS-SPRINT-198

 Description   

The original embedded server based query classifier has been used for verifying the behaviour of the sqlite3 based query classifier. The assumption has been that they are sufficiently different, so if they classify something the same way, it is reasonable to expect that the classification is correct.
However, in recent server versions the embedded library is no longer built, which makes it more cumbersome to use. Further, since the server evolved there are frequently issues that need to be fixed. Thus, the time has come to retire the embedded library based query classifier.
If the current query classification result that the embedded server based and sqlite3 based query classifiers agrees upon, is saved, together with the statement itself, the embedded server based query classifier can be retired and the testing can henceforth be made using that saved result.


Generated at Thu Feb 08 04:32:02 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.