scheduler: support the hot-region-scheduler to generate multiple operators at the same time #4931

HunDunDM · 2022-05-12T06:33:50Z

What problem does this PR solve?

Issue Number: Ref #4949

What is changed and how does it work?

Allows a solution to contain multiple Operators to expand the search set, guaranteeing that a better solution can be found in some cases.

Check List

Tests

Unit test

Release note

Optimized the search strategy of the balance-hot-region-scheduler.

ti-chi-bot · 2022-05-12T06:33:51Z

[REVIEW NOTIFICATION]

This pull request has been approved by:

lhy1024
nolouch

To complete the pull request process, please ask the reviewers in the list to review by filling /cc @reviewer in the comment.
After your PR has acquired the required number of LGTMs, you can assign this pull request to the committer in the list by filling /assign @committer in the comment to help you merge this pull request.

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

Reviewer can indicate their review by submitting an approval review.
Reviewer can cancel approval by submitting a request changes review.

codecov · 2022-05-17T19:15:18Z

Codecov Report

Merging #4931 (09fd67f) into master (34a4cce) will decrease coverage by 0.24%.
The diff coverage is 88.46%.

@@            Coverage Diff             @@
##           master    #4931      +/-   ##
==========================================
- Coverage   75.87%   75.62%   -0.25%     
==========================================
  Files         311      311              
  Lines       30830    30909      +79     
==========================================
- Hits        23392    23375      -17     
- Misses       5450     5527      +77     
- Partials     1988     2007      +19

Flag	Coverage Δ
unittests	`75.62% <88.46%> (-0.25%)`	⬇️

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

Impacted Files	Coverage Δ
server/schedulers/hot_region.go	`84.17% <88.46%> (+0.65%)`	⬆️
pkg/dashboard/keyvisual/input/core.go	`0.00% <0.00%> (-33.34%)`	⬇️
server/tso/local_allocator.go	`64.86% <0.00%> (-13.52%)`	⬇️
server/schedulers/shuffle_hot_region.go	`55.55% <0.00%> (-10.11%)`	⬇️
pkg/dashboard/adapter/manager.go	`79.31% <0.00%> (-6.90%)`	⬇️
server/storage/endpoint/meta.go	`64.04% <0.00%> (-4.50%)`	⬇️
server/tso/allocator_manager.go	`61.39% <0.00%> (-4.33%)`	⬇️
pkg/etcdutil/etcdutil.go	`84.88% <0.00%> (-3.49%)`	⬇️
server/tso/tso.go	`67.79% <0.00%> (-3.39%)`	⬇️
server/schedule/hbstream/heartbeat_streams.go	`72.72% <0.00%> (-2.03%)`	⬇️
... and 13 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 34a4cce...09fd67f. Read the comment docs.

nolouch · 2022-06-08T04:35:04Z

ptal @lhy1024

lhy1024 · 2022-05-18T11:42:22Z

server/schedulers/hot_region.go

@@ -497,8 +537,43 @@ func (bs *balanceSolver) solve() []*operator.Operator {
 				bs.cur.dstStore = dstStore
 				bs.calcProgressiveRank()
 				tryUpdateBestSolution()
+
+				if searchRevertRegions && (bs.cur.progressiveRank >= -1 && bs.cur.progressiveRank <= 0) &&


Suggested change

if searchRevertRegions && (bs.cur.progressiveRank >= -1 && bs.cur.progressiveRank <= 0) &&

if searchRevertRegions && (bs.cur.progressiveRank == -1 || bs.cur.progressiveRank == 0) &&

I think we should support rank1 later.

server/schedulers/hot_region.go

lhy1024

When the resources are sufficient, the two branches of master and multi-op behave similarly, and there is almost no scheduling after stabilization which means the solution can be found (split.qps-threshold=500) , and there is still scheduling after the solution cannot be found (split.qps-threshold=3000)

When resources are insufficient, the number of peak scheduling of multi-op is more

nolouch

rest LGTM

nolouch · 2022-07-11T06:26:08Z

server/schedulers/hot_region.go

+				break
+			}
+			bs.decorateOperator(currentOp, true, targetLabel, sourceLabel, typ, dim)
+			ops = append(ops, currentOp)


I wonder if the RandBuckets inwaiting_operator will be influenced. The GetOperator may only get first one operator and miss the second one.

server/schedulers/hot_region.go

…ators at the same time Signed-off-by: HunDunDM <[email protected]>

Signed-off-by: HunDunDM <[email protected]>

nolouch · 2022-07-14T06:26:47Z

/merge

ti-chi-bot · 2022-07-14T06:26:48Z

@nolouch: It seems you want to merge this PR, I will help you trigger all the tests:

/run-all-tests

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 ti-community-infra/tichi repository.

ti-chi-bot · 2022-07-14T06:26:50Z

This pull request has been accepted and is ready to merge.

Commit hash: 28204c0

…ple operators at the same time (tikv#4931)" This reverts commit 8ab5c16. Signed-off-by: lhy1024 <[email protected]> Conflicts: server/schedulers/hot_region_test.go

ref #4912, ref #4931, ref #5384, ref tikv/tikv#13095 Signed-off-by: lhy1024 <[email protected]>

lhy1024 · 2022-09-09T06:04:42Z

/run-build-arm64 comment=true

ti-chi-bot added do-not-merge/needs-linked-issue do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. release-note Denotes a PR that will be considered when it comes time to generate release notes. labels May 12, 2022

ti-chi-bot requested a review from rleungx May 12, 2022 06:33

HunDunDM requested review from lhy1024 and removed request for rleungx May 12, 2022 06:34

HunDunDM force-pushed the multi-op branch 4 times, most recently from dd984d3 to d97bd4f Compare May 12, 2022 19:20

ti-chi-bot added the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label May 16, 2022

HunDunDM force-pushed the multi-op branch from 63acf49 to 564bc83 Compare May 17, 2022 19:07

ti-chi-bot removed the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label May 17, 2022

HunDunDM mentioned this pull request May 17, 2022

Improvement of hot scheduler when meets dimension conflict #4949

Closed

25 tasks

ti-chi-bot removed the do-not-merge/needs-linked-issue label May 17, 2022

HunDunDM marked this pull request as ready for review May 17, 2022 19:15

ti-chi-bot removed the do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. label May 17, 2022

ti-chi-bot added the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Jun 8, 2022

HunDunDM force-pushed the multi-op branch from c96f2a0 to cc9a0c4 Compare June 8, 2022 05:51

ti-chi-bot removed the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Jun 8, 2022

HunDunDM force-pushed the multi-op branch from 2538998 to a06eb61 Compare June 9, 2022 03:46

HunDunDM requested a review from nolouch June 9, 2022 04:09

lhy1024 reviewed Jun 9, 2022

View reviewed changes

lhy1024 reviewed Jun 14, 2022

View reviewed changes

server/schedulers/hot_region.go Show resolved Hide resolved

bufferflies reviewed Jun 22, 2022

View reviewed changes

server/schedulers/hot_region.go Outdated Show resolved Hide resolved

lhy1024 approved these changes Jun 30, 2022

View reviewed changes

ti-chi-bot added the status/LGT1 Indicates that a PR has LGTM 1. label Jun 30, 2022

HunDunDM force-pushed the multi-op branch from 88fbde6 to 5c2507e Compare July 5, 2022 18:49

HunDunDM requested a review from lhy1024 July 5, 2022 18:49

nolouch reviewed Jul 11, 2022

View reviewed changes

scheduler: support the hot-region-scheduler to generate multiple oper…

13d82a0

…ators at the same time Signed-off-by: HunDunDM <[email protected]>

ti-chi-bot added the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Jul 13, 2022

address comment

c4ae935

Signed-off-by: HunDunDM <[email protected]>

HunDunDM force-pushed the multi-op branch from 8465cc0 to c4ae935 Compare July 14, 2022 01:58

ti-chi-bot removed the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label Jul 14, 2022

address comment

28204c0

Signed-off-by: HunDunDM <[email protected]>

lhy1024 approved these changes Jul 14, 2022

View reviewed changes

nolouch approved these changes Jul 14, 2022

View reviewed changes

ti-chi-bot added status/LGT2 Indicates that a PR has LGTM 2. and removed status/LGT1 Indicates that a PR has LGTM 1. labels Jul 14, 2022

ti-chi-bot added the status/can-merge Indicates a PR has been approved by a committer. label Jul 14, 2022

Merge branch 'master' into multi-op

09fd67f

ti-chi-bot merged commit 8ab5c16 into tikv:master Jul 14, 2022

HunDunDM deleted the multi-op branch July 14, 2022 07:45

lhy1024 mentioned this pull request Aug 4, 2022

scheduler: revert two unstable optimization in hot scheduler #5392

Merged

ti-chi-bot pushed a commit that referenced this pull request Aug 4, 2022

scheduler: revert two unstable optimization in hot scheduler (#5392)

e3d48eb

ref #4912, ref #4931, ref #5384, ref tikv/tikv#13095 Signed-off-by: lhy1024 <[email protected]>

lhy1024 mentioned this pull request Sep 13, 2022

scheduler: unstable redundant schedule in hot scheduler #5489

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

scheduler: support the hot-region-scheduler to generate multiple operators at the same time #4931

scheduler: support the hot-region-scheduler to generate multiple operators at the same time #4931

HunDunDM commented May 12, 2022 •

edited

Loading

ti-chi-bot commented May 12, 2022 •

edited

Loading

codecov bot commented May 17, 2022 •

edited

Loading

nolouch commented Jun 8, 2022

lhy1024 May 18, 2022

lhy1024 Jun 9, 2022

lhy1024 left a comment •

edited

Loading

nolouch left a comment

nolouch Jul 11, 2022

nolouch commented Jul 14, 2022

ti-chi-bot commented Jul 14, 2022

ti-chi-bot commented Jul 14, 2022

lhy1024 commented Sep 9, 2022

	if searchRevertRegions && (bs.cur.progressiveRank >= -1 && bs.cur.progressiveRank <= 0) &&
	if searchRevertRegions && (bs.cur.progressiveRank == -1 \|\| bs.cur.progressiveRank == 0) &&

scheduler: support the hot-region-scheduler to generate multiple operators at the same time #4931

scheduler: support the hot-region-scheduler to generate multiple operators at the same time #4931

Conversation

HunDunDM commented May 12, 2022 • edited Loading

What problem does this PR solve?

What is changed and how does it work?

Check List

Release note

ti-chi-bot commented May 12, 2022 • edited Loading

codecov bot commented May 17, 2022 • edited Loading

Codecov Report

nolouch commented Jun 8, 2022

lhy1024 May 18, 2022

Choose a reason for hiding this comment

lhy1024 Jun 9, 2022

Choose a reason for hiding this comment

lhy1024 left a comment • edited Loading

Choose a reason for hiding this comment

nolouch left a comment

Choose a reason for hiding this comment

nolouch Jul 11, 2022

Choose a reason for hiding this comment

nolouch commented Jul 14, 2022

ti-chi-bot commented Jul 14, 2022

ti-chi-bot commented Jul 14, 2022

lhy1024 commented Sep 9, 2022

HunDunDM commented May 12, 2022 •

edited

Loading

ti-chi-bot commented May 12, 2022 •

edited

Loading

codecov bot commented May 17, 2022 •

edited

Loading

lhy1024 left a comment •

edited

Loading