Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
25.10.4
-
None
-
MXS-SPRINT-280
-
Can result in unexpected behaviour
-
Readwritesplit now respects sync_transaction_timeout more closely when values for it are very low.
Description
If a backend failure causes a replacement attempt to be made, the attempt is alwys done with a 1 second delay. If there's less than 1 second left until the timeout, the client would see a slowdown of up to 1 second.
Attachments
Issue Links
- blocks
-
MXS-7561 Automatically enable sync_transaction if throttling starts
-
- Closed
-