infoschema: Add ALL_SQL_DIGESTS field to the TIDB_TRX table #24863

MyonKeminta · 2021-05-24T11:22:53Z

What problem does this PR solve?

Problem Summary:
The TIDB_TRX table is part of Lock View , introduced in #22908 . However one of the important fields ALL_SQLS is not finished in that PR. This PR adds this field, and also avoids ALL_SQLS, DIGEST, State being inconsistent due to multiple atomic loads.

What is changed and how it works?

What's Changed: Makes the TxnInfo struct a hybrid thing of mutable fields (updated with atomic operations) and immutable fields (updated by replacing the whole struct). When updating the SQL Digest, the whole struct will be replaced. However, the other fields that are not so important (len and size) is still loaded by multiple atomic operations. Then the AllSQLs information is added to the struct as an immutable field.

This PR also added more tests.

Related changes

Check List

Tests

Unit test

Side effects

Performance regression
- Consumes more CPU
- Consumes more MEM

Release note

No release note.

Signed-off-by: MyonKeminta <MyonKeminta@users.noreply.github.com>

…trx-dev

MyonKeminta · 2021-05-24T11:33:27Z

/run-check_dev_2

session/session.go

session/txninfo/txn_info.go

session/txn.go

MyonKeminta · 2021-05-25T04:07:35Z

Is the bench command still available?
/bench

…trx-dev

Signed-off-by: MyonKeminta <MyonKeminta@users.noreply.github.com>

MyonKeminta · 2021-05-26T11:35:40Z

/run-all-tests

longfangsong · 2021-05-26T14:04:57Z

session/txn.go

+		return
+	}
+
+	info := txn.txnInfo.GetSnapInfo()


Though GetSnapInfo is thread safe, accessing txn.txnInfo is not, so here's a datarace with txn.txnInfo = info in storeTxnInfo.

Can we use atomicTxnInfo all the time?

onStmtStart is expected to run within the transaction's thread. The purpose I keeps the txnInfo field is to avoid atomic loading every time it's used, and I expect the txnInfo is only used within the transaction's thread and the atomicTxnInfo is only used for fetching information concurrently. Perhaps this is a bad idea. 😕

Ah sorry. I just notice the race reported by CI. It's weird to me that onStmtStart and onStmtEnd is invoked concurrently...

The test TestAmendForUniqueIndex in pessimistic_test.go is executing two queries concurrently in a single tk... @cfzjywxk Is this this kind of usage allowed..?

go func() { var err error err = tk2.ExecToErr("begin pessimistic") if err != nil { errCh <- err return } err = tk2.ExecToErr("insert into t values(5, 5)") if err != nil { errCh <- err return } err = tk2.ExecToErr("delete from t where id = 5") if err != nil { errCh <- err return } // let commit in tk start. errCh <- err time.Sleep(time.Millisecond * 100) err = tk2.ExecToErr("commit") errCh <- err }() err = <-errCh c.Assert(err, Equals, nil) tk.MustExec("commit") tk2.MustExec("admin check table t") err = <-errCh c.Assert(err, Equals, nil)

It's unexpected usage, need to fix.

cfzjywxk

/lgtm

longfangsong

/lgtm

ti-chi-bot · 2021-05-27T07:16:23Z

[REVIEW NOTIFICATION]

This pull request has been approved by:

cfzjywxk
longfangsong

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 writing /lgtm in a comment.
Reviewer can cancel approval by writing /lgtm cancel in a comment.

MyonKeminta · 2021-05-27T10:02:33Z

/merge

ti-chi-bot · 2021-05-27T10:02:37Z

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

Commit hash: 43cd7b4

ti-chi-bot · 2021-05-27T10:02:45Z

@MyonKeminta: 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.

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.

MyonKeminta · 2021-05-27T11:18:01Z

/run-unit-test

MyonKeminta · 2021-05-27T11:25:15Z

/merge

MyonKeminta · 2021-05-27T11:36:32Z

What's wrong with TestMemoryUsageAlarmVariable

[2021-05-27T11:29:47.765Z] ----------------------------------------------------------------------
[2021-05-27T11:29:47.765Z] FAIL: session_test.go:3812: testSessionSuite2.TestMemoryUsageAlarmVariable
[2021-05-27T11:29:47.765Z] 
[2021-05-27T11:29:47.765Z] session_test.go:3820:
[2021-05-27T11:29:47.765Z]     tk.MustQuery("select @@session.tidb_memory_usage_alarm_ratio").Check(testkit.Rows("0.7"))
[2021-05-27T11:29:47.765Z] /home/jenkins/agent/workspace/tidb_ghpr_unit_test/go/src/github.com/pingcap/tidb/util/testkit/testkit.go:63:
[2021-05-27T11:29:47.765Z]     res.c.Assert(resBuff.String(), check.Equals, needBuff.String(), res.comment)
[2021-05-27T11:29:47.765Z] ... obtained string = "[0.8]\n"
[2021-05-27T11:29:47.765Z] ... expected string = "[0.7]\n"
[2021-05-27T11:29:47.766Z] ... sql:select @@session.tidb_memory_usage_alarm_ratio, args:[]

MyonKeminta · 2021-05-27T11:41:47Z

/merge

tiancaiamao · 2021-07-02T14:37:03Z

session/txn.go

 	txn.initStmtBuf()
-	atomic.StoreUint64(&txn.EntriesCount, uint64(txn.Transaction.Len()))
-	atomic.StoreUint64(&txn.EntriesSize, uint64(txn.Transaction.Size()))
+	txn.recreateTxnInfo(


Why atomic? when will the changeInvalidToValid() function been called concurrently? @MyonKeminta

It should not be called concurrently, but the content of the txnInfo may be retrieved by another thread outside the session.

MyonKeminta added 8 commits May 20, 2021 18:01

Support replacing whole struct

eea350a

fix column def

b1b7422

Signed-off-by: MyonKeminta <MyonKeminta@users.noreply.github.com>

Fix tests

0126dd9

Add basic tests for all sqls and autocommit transactions

cee475e

Add more tests

4183553

Signed-off-by: MyonKeminta <MyonKeminta@users.noreply.github.com>

Merge branch 'master' of https://github.com/pingcap/tidb into m/tidb-…

cd5891c

…trx-dev

Add ALL_SQLS field to TIDB_TRX

b81d429

Update comments

318e1ce

MyonKeminta added the sig/transaction SIG:Transaction label May 24, 2021

MyonKeminta requested review from cfzjywxk and longfangsong May 24, 2021 11:22

ti-chi-bot added the size/XL Denotes a PR that changes 500-999 lines, ignoring generated files. label May 24, 2021

github-actions bot added the sig/sql-infra SIG: SQL Infra label May 24, 2021

cfzjywxk reviewed May 24, 2021

View reviewed changes

session/session.go Outdated Show resolved Hide resolved

session/txninfo/txn_info.go Outdated Show resolved Hide resolved

session/txn.go Show resolved Hide resolved

MyonKeminta mentioned this pull request May 25, 2021

Implement Lock View (Sprint 3) #24199

Closed

20 tasks

update txn digest info on PS protocol

9914ae6

ti-chi-bot added the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label May 26, 2021

Merge branch 'master' of https://github.com/pingcap/tidb into m/tidb-…

0522b42

…trx-dev

ti-chi-bot removed the needs-rebase Indicates a PR cannot be merged because it has merge conflicts with HEAD. label May 26, 2021

Renaming

bfe9733

Signed-off-by: MyonKeminta <MyonKeminta@users.noreply.github.com>

MyonKeminta requested a review from a team as a code owner May 26, 2021 08:37

MyonKeminta requested review from lzmhhh123 and removed request for a team May 26, 2021 08:37

MyonKeminta changed the title ~~infoschema: Add ALL_SQLS field to the TIDB_TRX table~~ infoschema: Add ALL_SQL_DIGESTS field to the TIDB_TRX table May 26, 2021

github-actions bot added the sig/execution SIG execution label May 26, 2021

Address comment

e5921e5

longfangsong reviewed May 26, 2021

View reviewed changes

Address comment

43cd7b4

cfzjywxk reviewed May 27, 2021

View reviewed changes

ti-chi-bot added the status/LGT1 Indicates that a PR has LGTM 1. label May 27, 2021

longfangsong approved these changes May 27, 2021

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 May 27, 2021

ti-chi-bot added the status/can-merge Indicates a PR has been approved by a committer. label May 27, 2021

Merge branch 'master' into m/tidb-trx-dev

96af737

ti-chi-bot merged commit c3a27c9 into pingcap:master May 27, 2021

MyonKeminta deleted the m/tidb-trx-dev branch May 27, 2021 11:50

tiancaiamao reviewed Jul 2, 2021

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

infoschema: Add ALL_SQL_DIGESTS field to the TIDB_TRX table #24863

infoschema: Add ALL_SQL_DIGESTS field to the TIDB_TRX table #24863

MyonKeminta commented May 24, 2021

MyonKeminta commented May 24, 2021

MyonKeminta commented May 25, 2021 •

edited

Loading

MyonKeminta commented May 26, 2021

longfangsong May 26, 2021 •

edited

Loading

MyonKeminta May 27, 2021

MyonKeminta May 27, 2021

MyonKeminta May 27, 2021

cfzjywxk May 27, 2021

cfzjywxk left a comment

longfangsong left a comment

ti-chi-bot commented May 27, 2021

MyonKeminta commented May 27, 2021

ti-chi-bot commented May 27, 2021

ti-chi-bot commented May 27, 2021

MyonKeminta commented May 27, 2021

MyonKeminta commented May 27, 2021

MyonKeminta commented May 27, 2021

MyonKeminta commented May 27, 2021

tiancaiamao Jul 2, 2021

MyonKeminta Jul 5, 2021

infoschema: Add ALL_SQL_DIGESTS field to the TIDB_TRX table #24863

infoschema: Add ALL_SQL_DIGESTS field to the TIDB_TRX table #24863

Conversation

MyonKeminta commented May 24, 2021

What problem does this PR solve?

What is changed and how it works?

Related changes

Check List

Release note

MyonKeminta commented May 24, 2021

MyonKeminta commented May 25, 2021 • edited Loading

MyonKeminta commented May 26, 2021

longfangsong May 26, 2021 • edited Loading

Choose a reason for hiding this comment

MyonKeminta May 27, 2021

Choose a reason for hiding this comment

MyonKeminta May 27, 2021

Choose a reason for hiding this comment

MyonKeminta May 27, 2021

Choose a reason for hiding this comment

cfzjywxk May 27, 2021

Choose a reason for hiding this comment

cfzjywxk left a comment

Choose a reason for hiding this comment

longfangsong left a comment

Choose a reason for hiding this comment

ti-chi-bot commented May 27, 2021

MyonKeminta commented May 27, 2021

ti-chi-bot commented May 27, 2021

ti-chi-bot commented May 27, 2021

MyonKeminta commented May 27, 2021

MyonKeminta commented May 27, 2021

MyonKeminta commented May 27, 2021

MyonKeminta commented May 27, 2021

tiancaiamao Jul 2, 2021

Choose a reason for hiding this comment

MyonKeminta Jul 5, 2021

Choose a reason for hiding this comment

MyonKeminta commented May 25, 2021 •

edited

Loading

longfangsong May 26, 2021 •

edited

Loading