Closed tiancaiamao closed 1 month ago
Hi @tiancaiamao. Thanks for your PR.
PRs from untrusted users cannot be marked as trusted with /ok-to-test
in this repo meaning untrusted PR authors can never trigger tests themselves. Collaborators can still trigger tests on the PR using /test all
.
I understand the commands that are listed here.
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 74.5497%. Comparing base (
5d27b73
) to head (39eee29
). Report is 16 commits behind head on master.
[APPROVALNOTIFIER] This PR is APPROVED
This pull-request has been approved by: bb7133, Defined2014
The full list of commands accepted by this bot can be found here.
The pull request process is described here
Timeline:
2024-05-17 07:12:10.583660354 +0000 UTC m=+1810084.340795928
: :ballot_box_with_check: agreed by bb7133.2024-05-17 08:45:54.512209004 +0000 UTC m=+1815708.269344574
: :ballot_box_with_check: agreed by Defined2014.In response to a cherrypick label: new pull request created to branch release-8.1
: #53396.
What problem does this PR solve?
Issue Number: ref #38442 close #53393
Problem Summary:
What changed and how does it work?
The original design is that we try our best to avoid holes in AUTO_ID_CACHE=1, so when autoid leader exit, forceRebase is called to return the non-consumed in memory IDs back.
When there are a lot of tables, this operation could not be finish quickly. And tidb graceful shutdown can not done within 30s. This cause tidb can not restart quickly to provide service.
We think over it and deside to remove the force rebase on close behavior.
Check List
Tests
No unit test apply here (it requires tidb restart to cover the changes). Some integration test may notice the change.
Side effects
Documentation
Behavior change, not tidb restart can cause hole in auto id allocating using AUTO_ID_CACHE=1
Release note
Please refer to Release Notes Language Style Guide to write a quality release note.