vector: A new load balance to improve the vector search performance when tiflash replica >1#9636
Conversation
…tiflash replica >1
|
@JinheLin @Lloyd-Pottiger PTAL |
|
@JaySon-Huang: GitHub didn't allow me to request PR reviews from the following users: kolafish. Note that only pingcap members and repo collaborators can review this PR, and authors cannot review their own PRs. DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: JinheLin, Lloyd-Pottiger The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
[LGTM Timeline notifier]Timeline:
|
|
@JaySon-Huang: Your PR was out of date, I have automatically updated it for you. At the same time I will also trigger all tests for you: /run-all-tests
If the CI test fails, you just re-trigger the test that failed and the bot will merge the PR for you after the CI passes. DetailsInstructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the ti-community-infra/tichi repository. |
|
/test pull-unit-test |
|
In response to a cherrypick label: new pull request created to branch |
…hen tiflash replica >1 (#9636) (#9651) close #9637 In this pr: [https://github.com/tikv/client-c/pull/193](https://github.com/tikv/client-c/pull/193) add a bool paramater centralized_schedule for Client::buildBatchCopTasks When set to true: Regions are concentrated into as few BatchCopTasks as possible This PR modifies the call to this function, setting the newly added parameter to true when the request is a ANN query. Signed-off-by: Calvin Neo <calvinneo1995@gmail.com> Co-authored-by: Yu Jin <yujin.yujin@gmail.com> Co-authored-by: JaySon-Huang <tshent@qq.com> Co-authored-by: Calvin Neo <CalvinNeo@users.noreply.github.com> Co-authored-by: JaySon <tshent@qq.com>
backport https://github.com/tidbcloud/tiflash-cse/pull/295
What problem does this PR solve?
Issue Number: close #9637
Problem Summary:
when tiflash replica>1, the vector search qps is only 60% of the case when the tiflash replica=1
What is changed and how it works?
Check List
Tests
Side effects
Documentation
Release note