store/gcworker: reduce "full config reset" log on pd side (#44331) | tidb-test=pr/2145

[ti-chi-bot]

This is an automated cherry-pick of #44331

What problem does this PR solve?

Issue Number: close #33069

Problem Summary:

What is changed and how it works?

Similiar to #33082, reduce the duplicated request.

If there are N range to be deleted, there will be N logs like "full config reset" on the pd side.
It affects v6.1.1 and master both.

1. (*GCWorker).deleteRanges execute doGCPlacementRules function for each DelRangeTask

   tidb/store/gcworker/gc_worker.go

   Line 924 in b7d3c08

   if err := w.doGCPlacementRules(se, safePoint, r, gcPlacementRuleCache); err != nil {

2. doGCPlacementRules get the DDL history by the DelRangeTask.JobID
   

   tidb/store/gcworker/gc_worker.go

   Line 2165 in b7d3c08

   historyJob, err = ddl.GetHistoryJobByID(se, dr.JobID)

3. Retrive all the Physical Table IDs dropped by this DDL
   

   tidb/store/gcworker/gc_worker.go

   Line 2179 in b7d3c08

   if err = historyJob.DecodeArgs(&startKey, &physicalTableIDs); err != nil {

4. And delete all the Bundles for those Physical Table IDs

   tidb/store/gcworker/gc_worker.go

   Line 2220 in b7d3c08

   return infosync.PutRuleBundlesWithDefaultRetry(context.TODO(), bundles)

There are many duplicated physical IDs during step 1 and step 3

In #33069 it add a cache to handle the Delete rule
But note this line

tidb/store/gcworker/gc_worker.go

Line 2220 in b7d3c08

return infosync.PutRuleBundlesWithDefaultRetry(context.TODO(), bundles)

, there would still be duplicated request there, so pd print "full config reset" many times.

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)

update mysql.tidb set VARIABLE_VALUE = '3m' where VARIABLE_NAME = 'tikv_gc_run_interval';
update mysql.tidb set VARIABLE_VALUE = '3m' where VARIABLE_NAME = 'tikv_gc_life_time';

create table t (id int) partition by hash partitions 2000;
insert into t values (1),(2),(3), ....
insert into t select * from t;

drop table t;

wait 3min, check the pd log ...

After the fix, there would be only one line "full config reset" in the log for the drop table operation.

• No code

Side effects

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

Documentation

• Affects user behaviors
• Contains syntax changes
• Contains variable changes
• Contains experimental features
• Changes MySQL compatibility

Release note

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

None

[ti-chi-bot]

[REVIEW NOTIFICATION]

This pull request has been approved by:

• lcwangchao
• xhebox

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.

[hawkingrei]

/merge

[ti-chi-bot]

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

Commit hash: e239a5d

[tiancaiamao]

/test mysql-test

[tiancaiamao]

It's based on a previous hotfix #44306 and the mysql-test is a non-standard one (to pass CI).

[tiancaiamao]

/test mysql-test