planner: variable `tidb_opt_enable_hash_join` to skip hash join #46575

coderplay · 2023-08-31T16:57:17Z

What problem does this PR solve?

This PR adds a global/session knob to disable hash join, in order to prevent query plan regression.
Use cases:

As an application developer, I’m very confident the queries don’t benefit from hash join. However, at times, bad plans were generated in which a hash join was selected. I will config the connection pool, and initialize the connections with set tidb_opt_enable_hash_join=no.
As a DBA, I’m aware the application is OLTP workload, which works properly in MySQL ~5.7. I’m very sure the SQLs don’t need hash join. I’ve seen cases in the past. bad plans were generated in which a hash join was selected. After confirming with application developers, I will disable hash join in cluster level with set global tidb_opt_enable_hash_join=off

Issue Number: close #46695

What is changed and how it works?

Check List

Tests

Unit test
Integration test
Manual test (add detailed scripts or steps below)
No code

Side effects

Performance regression: Consumes more CPU
Performance regression: Consumes more Memory
Breaking backward compatibility

Documentation

Release note

Please refer to Release Notes Language Style Guide to write a quality release note.

Add a new global/session variable `tidb_opt_enable_hash_join` to control whether hash join is enabled or not.

sre-bot · 2023-08-31T16:57:26Z

All committers have signed the CLA.

ti-chi-bot · 2023-08-31T16:57:27Z

Welcome @coderplay!

It looks like this is your first PR to pingcap/tidb 🎉.

I'm the bot to help you request reviewers, add labels and more, See available commands.

We want to make sure your contribution gets all the attention it needs!

Thank you, and welcome to pingcap/tidb. 😃

ti-chi-bot · 2023-08-31T16:57:27Z

Hi @coderplay. Thanks for your PR.

I'm waiting for a pingcap member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

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.

tiprow · 2023-08-31T16:57:41Z

Hi @coderplay. 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.

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.

coderplay · 2023-08-31T16:59:25Z

/review default

ti-chi-bot · 2023-08-31T16:59:47Z

@coderplay:

Potential Problems and Suggestions

Problem 1

In the PR description, there is a placeholder Issue Number: close #xxx. The issue number should be provided to link the relevant issue.

Suggestion:

Replace xxx with the actual issue number.

Problem 2

There are no tests to check if setting the global variable tidb_opt_enable_hash_join to no actually disables hash join at the cluster level.

Suggestion:

Add a test case to verify that setting the global variable tidb_opt_enable_hash_join to no actually disables hash join at the cluster level.

Problem 3

There is no documentation update for the newly added global/session variable tidb_opt_enable_hash_join.

Suggestion:

Update the documentation to include the new global/session variable tidb_opt_enable_hash_join, its purpose, and its usage.

Problem 4

In the PR description, the release note is set to None. However, this PR introduces a new global/session variable which might affect user behaviors.

Suggestion:

Update the release note to:

Add a new global/session variable `tidb_opt_enable_hash_join` to control whether hash join is enabled or not.

Overall, the PR looks good. Just a few minor issues and suggestions to make it even better.

In response to this:

/review default

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.

coderplay · 2023-09-06T15:39:53Z

/retest

ti-chi-bot · 2023-09-06T15:40:10Z

@coderplay: Cannot trigger testing until a trusted user reviews the PR and leaves an /ok-to-test message.

In response to this:

/retest

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.

tiprow · 2023-09-06T15:40:34Z

@coderplay: Cannot trigger testing until a trusted user reviews the PR and leaves an /ok-to-test message.

In response to this:

/retest

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.

coderplay · 2023-09-06T15:43:42Z

/cc @qw4990 @hawkingrei @chrysan @time-and-fate

coderplay · 2023-09-06T15:46:02Z

planner/core/casetest/rule/testdata/join_reorder_suite_out.json

+        "Warning": null
+      },
+      {
+        "SQL": "select /*+ leading(t4, t3, t2, t1) */ * from t1, t2, t3, t4",


Why hashjoin can't be disabled in this case?

I'll take a look at this.

no is not a valid value for a bool variable in TiDB, only on, off, 1, 0. After setting it to off, it can work:

@qw4990 My original meaning was that even in your PR, the hash join is not disabled when applying a leading hint along with a no_hash_join hint.

see my comment on #45538

In my case, the no_hash_join only works for t2 and t3. The top Hash Join is applied on t1 and result of (t2, t3, t4) instead of t2, so no_hash_join(t2) cannot take effect on it.

qw4990 · 2023-09-25T04:52:54Z

And do we need to pick this PR to v7.1? @coderplay @songrijie

codecov · 2023-09-25T05:11:19Z

Codecov Report

Merging #46575 (a6893ef) into master (ae11336) will decrease coverage by 0.1177%.
Report is 49 commits behind head on master.
The diff coverage is 100.0000%.

Additional details and impacted files

@@               Coverage Diff                @@
##             master     #46575        +/-   ##
================================================
- Coverage   72.9961%   72.8785%   -0.1177%     
================================================
  Files          1338       1367        +29     
  Lines        399462     409380      +9918     
================================================
+ Hits         291592     298350      +6758     
- Misses        89019      92174      +3155     
- Partials      18851      18856         +5

Flag	Coverage Δ
integration	`33.6852% <63.6363%> (?)`
unit	`73.1069% <100.0000%> (+0.1107%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Components	Coverage Δ
dumpling	`53.9913% <ø> (ø)`
parser	`84.9232% <ø> (-0.0388%)`	⬇️
br	`48.7605% <ø> (-4.3404%)`	⬇️

qw4990 · 2023-09-25T06:19:35Z

/retest-required

qw4990 · 2023-09-25T06:30:58Z

/retest-required

qw4990 · 2023-09-25T07:56:15Z

planner/core/optimizer.go

@@ -162,6 +162,7 @@ type logicalOptRule interface {
 func BuildLogicalPlanForTest(ctx context.Context, sctx sessionctx.Context, node ast.Node, infoSchema infoschema.InfoSchema) (Plan, types.NameSlice, error) {
 	sctx.GetSessionVars().PlanID.Store(0)
 	sctx.GetSessionVars().PlanColumnID.Store(0)
+	sctx.GetSessionVars().EnableHashJoin = true


Too many CI failures due to the default value of this variable:

Can we change the name of the variable from EnableHashJoin to DisableHashJoin and set it as false by default?

for most users, HashJoin should be allowed, so using false as the default value seems safer.

for some tests, the test framework won't use DefEnableHashJoin to initialize sctx.EnableHashJoin, instead it uses the default value of Bool in Golang, which is false; If we use EnableHashJoin, there will be plenty of test failures, changing it to DisableHashJoin can fix this.

After discussing with PM, we decided to keep it unchanged as EnableHashJoin.

Too many CI failures... For safety, how about using another solution, which is using DisableHashJoin internally, and exposing enable_hash_join to users.

Too many CI failures... For safety, how about using another solution, which is using DisableHashJoin internally, and exposing enable_hash_join to users.

@coderplay I've pushed some code to your PR directly (tomorrow is the code freeze deadline of v7.4 ...) by using this approach to solve CI failures.

Thanks for fixing those @qw4990 ! In the future, how can I detect those issues locally before submitting this PR?

qw4990 · 2023-09-25T11:52:23Z

/retest-required

qw4990 · 2023-09-25T13:04:56Z

@easonn7 PTAL

easonn7 · 2023-09-26T01:06:12Z

/approve

ti-chi-bot · 2023-09-26T01:06:19Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: easonn7, qw4990, time-and-fate

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

~~OWNERS~~ [qw4990,time-and-fate]
~~sessionctx/variable/OWNERS~~ [easonn7]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

ti-chi-bot · 2023-09-26T02:25:00Z

In response to a cherrypick label: new pull request created to branch release-6.5: #47276.

Signed-off-by: ti-chi-bot <ti-community-prow-bot@tidb.io>

ti-chi-bot · 2023-09-26T02:25:45Z

In response to a cherrypick label: new pull request created to branch release-7.1: #47277.

Signed-off-by: ti-chi-bot <ti-community-prow-bot@tidb.io>

#47359) close #46695

…) (#47277) close #46695

coderplay added 2 commits August 31, 2023 09:48

init commit

f2a5193

Merge branch 'master' of https://github.com/pingcap/tidb

a240907

coderplay requested a review from a team as a code owner August 31, 2023 16:57

ti-chi-bot bot added do-not-merge/needs-linked-issue release-note-none Denotes a PR that doesn't merit a release note. labels Aug 31, 2023

ti-chi-bot bot added size/L Denotes a PR that changes 100-499 lines, ignoring generated files. needs-ok-to-test Indicates a PR created by contributors and need ORG member send '/ok-to-test' to start testing. labels Aug 31, 2023

coderplay added 3 commits September 5, 2023 09:25

revert the library importing

e48b565

revert the library importing

ac6748e

merge the master

b0a6417

ti-chi-bot bot added do-not-merge/needs-tests-checked release-note Denotes a PR that will be considered when it comes time to generate release notes. and removed release-note-none Denotes a PR that doesn't merit a release note. do-not-merge/needs-linked-issue labels Sep 5, 2023

wuhuizuo removed the do-not-merge/needs-tests-checked label Sep 6, 2023

ti-chi-bot bot requested review from chrysan, hawkingrei, qw4990 and time-and-fate September 6, 2023 15:43

coderplay commented Sep 6, 2023

View reviewed changes

Make linter happy

cae12dc

fix

cef90b0

qw4990 added needs-cherry-pick-release-6.5 Should cherry pick this PR to release-6.5 branch. needs-cherry-pick-release-7.1 Should cherry pick this PR to release-7.1 branch. labels Sep 25, 2023

qw4990 added 2 commits September 25, 2023 15:01

fix

3bbd2df

fix ci

cdb93a7

qw4990 reviewed Sep 25, 2023

View reviewed changes

qw4990 added 3 commits September 25, 2023 17:16

fix ci

c6dc9b3

fix ci

e4dd5d7

fix ci

a6893ef

qw4990 approved these changes Sep 25, 2023

View reviewed changes

ti-chi-bot bot added the approved label Sep 26, 2023

ti-chi-bot bot merged commit 95fa30c into pingcap:master Sep 26, 2023
13 of 16 checks passed

ti-chi-bot mentioned this pull request Sep 26, 2023

planner: variable tidb_opt_enable_hash_join to skip hash join (#46575) #47276

Closed

12 tasks

ti-chi-bot pushed a commit to ti-chi-bot/tidb that referenced this pull request Sep 26, 2023

This is an automated cherry-pick of pingcap#46575

446fc04

Signed-off-by: ti-chi-bot <ti-community-prow-bot@tidb.io>

ti-chi-bot mentioned this pull request Sep 26, 2023

planner: variable tidb_opt_enable_hash_join to skip hash join (#46575) #47277

Merged

12 tasks

ti-chi-bot pushed a commit to ti-chi-bot/tidb that referenced this pull request Sep 26, 2023

This is an automated cherry-pick of pingcap#46575

772713f

Signed-off-by: ti-chi-bot <ti-community-prow-bot@tidb.io>

qw4990 mentioned this pull request Sep 28, 2023

planner: variable tidb_opt_enable_hash_join to skip hash join (#46575) #47359

Merged

12 tasks

ti-chi-bot bot pushed a commit that referenced this pull request Oct 8, 2023

planner: variable tidb_opt_enable_hash_join to skip hash join (#46575) (

c9dae10

#47359) close #46695

ti-chi-bot bot pushed a commit that referenced this pull request Oct 8, 2023

planner: variable tidb_opt_enable_hash_join to skip hash join (#46575…

d121f46

…) (#47277) close #46695

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

planner: variable `tidb_opt_enable_hash_join` to skip hash join #46575

planner: variable `tidb_opt_enable_hash_join` to skip hash join #46575

coderplay commented Aug 31, 2023 •

edited

Loading

sre-bot commented Aug 31, 2023 •

edited

Loading

ti-chi-bot bot commented Aug 31, 2023

ti-chi-bot bot commented Aug 31, 2023

tiprow bot commented Aug 31, 2023

coderplay commented Aug 31, 2023

ti-chi-bot bot commented Aug 31, 2023

coderplay commented Sep 6, 2023

ti-chi-bot bot commented Sep 6, 2023

tiprow bot commented Sep 6, 2023

coderplay commented Sep 6, 2023

coderplay Sep 6, 2023

qw4990 Sep 8, 2023

qw4990 Sep 8, 2023 •

edited

Loading

coderplay Sep 12, 2023 •

edited

Loading

qw4990 Sep 14, 2023 •

edited

Loading

qw4990 commented Sep 25, 2023

codecov bot commented Sep 25, 2023 •

edited

Loading

qw4990 commented Sep 25, 2023

qw4990 commented Sep 25, 2023

qw4990 Sep 25, 2023

qw4990 Sep 25, 2023

qw4990 Sep 25, 2023 •

edited

Loading

qw4990 Sep 25, 2023 •

edited

Loading

coderplay Sep 25, 2023 •

edited

Loading

qw4990 commented Sep 25, 2023

qw4990 commented Sep 25, 2023

easonn7 commented Sep 26, 2023

ti-chi-bot bot commented Sep 26, 2023

ti-chi-bot commented Sep 26, 2023

ti-chi-bot commented Sep 26, 2023

planner: variable tidb_opt_enable_hash_join to skip hash join #46575

planner: variable tidb_opt_enable_hash_join to skip hash join #46575

Conversation

coderplay commented Aug 31, 2023 • edited Loading

What problem does this PR solve?

What is changed and how it works?

Check List

Release note

sre-bot commented Aug 31, 2023 • edited Loading

ti-chi-bot bot commented Aug 31, 2023

ti-chi-bot bot commented Aug 31, 2023

tiprow bot commented Aug 31, 2023

coderplay commented Aug 31, 2023

ti-chi-bot bot commented Aug 31, 2023

Potential Problems and Suggestions

Problem 1

Problem 2

Problem 3

Problem 4

coderplay commented Sep 6, 2023

ti-chi-bot bot commented Sep 6, 2023

tiprow bot commented Sep 6, 2023

coderplay commented Sep 6, 2023

coderplay Sep 6, 2023

Choose a reason for hiding this comment

qw4990 Sep 8, 2023

Choose a reason for hiding this comment

qw4990 Sep 8, 2023 • edited Loading

Choose a reason for hiding this comment

coderplay Sep 12, 2023 • edited Loading

Choose a reason for hiding this comment

qw4990 Sep 14, 2023 • edited Loading

Choose a reason for hiding this comment

qw4990 commented Sep 25, 2023

codecov bot commented Sep 25, 2023 • edited Loading

Codecov Report

qw4990 commented Sep 25, 2023

qw4990 commented Sep 25, 2023

qw4990 Sep 25, 2023

Choose a reason for hiding this comment

qw4990 Sep 25, 2023

Choose a reason for hiding this comment

qw4990 Sep 25, 2023 • edited Loading

Choose a reason for hiding this comment

qw4990 Sep 25, 2023 • edited Loading

Choose a reason for hiding this comment

coderplay Sep 25, 2023 • edited Loading

Choose a reason for hiding this comment

qw4990 commented Sep 25, 2023

qw4990 commented Sep 25, 2023

easonn7 commented Sep 26, 2023

ti-chi-bot bot commented Sep 26, 2023

ti-chi-bot commented Sep 26, 2023

ti-chi-bot commented Sep 26, 2023

planner: variable `tidb_opt_enable_hash_join` to skip hash join #46575

planner: variable `tidb_opt_enable_hash_join` to skip hash join #46575

coderplay commented Aug 31, 2023 •

edited

Loading

sre-bot commented Aug 31, 2023 •

edited

Loading

qw4990 Sep 8, 2023 •

edited

Loading

coderplay Sep 12, 2023 •

edited

Loading

qw4990 Sep 14, 2023 •

edited

Loading

codecov bot commented Sep 25, 2023 •

edited

Loading

qw4990 Sep 25, 2023 •

edited

Loading

qw4990 Sep 25, 2023 •

edited

Loading

coderplay Sep 25, 2023 •

edited

Loading