explain analyze runtime info is zero when using tiflash

[guo-shaoge]

Bug Report

Please answer these questions before submitting your issue. Thanks!

1. Minimal reproduce step (Required)

1. generate 1000 rows for any table
2. add tiflash replica
3. run following sql:explain analyze select * from bc_bet_records_111 limit 1000 offset 1000;

2. What did you expect to see? (Required)

tiflash runtime info should not be zero

3. What did you see instead (Required)

4. What is your TiDB version? (Required)

MySQL [wetech]> select tidb_version()
    -> ;
+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| tidb_version()                                                                                                                                                                                                                                                                        |
+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| Release Version: v8.1.0-alpha-43-g681987317c-dirty
Edition: Community
Git Commit Hash: 681987317c7ae15946b6270e9e8fb1286edadd4f
Git Branch: fix_dml_ru_missing
UTC Build Time: 2024-04-15 08:23:29
GoVersion: go1.21.0
Race Enabled: false
Check Table Before Drop: false
Store: tikv |
+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.009 sec)

[guo-shaoge]

similar to #50455, but the root cause is different. For this, it's because MPPCoordinatorManager.serverAddr is empty, so tiflash will not report runtime info to tidb.

Because MPPCoordinatorManager.serverAddr is inited when createServer is called, but server.StatusAddr is still empty at that time. The server.StatusAddr is inited inServer.Run(), which is called after createServer

tidb/cmd/tidb-server/main.go

Lines 327 to 339 in 224ae15

	svr := createServer(storage, dom)
	exited := make(chan struct{})
	signal.SetupSignalHandler(func() {
	svr.Close()
	cleanup(svr, storage, dom)
	cpuprofile.StopCPUProfiler()
	resourcemanager.InstanceResourceManager.Stop()
	executor.Stop()
	close(exited)
	})
	topsql.SetupTopSQL()
	terror.MustNil(svr.Run(dom))

[yibin87]

/assign @yibin87

[yibin87]

Checked v7.5.0, the issue doesn't exist. There seems some changes during v7.5.0 and latest version. Will check it.

[yibin87]

In #51472, listenStatusHTTPServer is moved from NewServer to Server.run

[yibin87]

/label affects-8.1

[yibin87]

/label affects-8.0

[ti-chi-bot]

@yibin87: The label(s) affects-8.0 cannot be applied. These labels are supported: fuzz/sqlancer, challenge-program, compatibility-breaker, first-time-contributor, contribution, good first issue, correctness, duplicate, proposal, security, ok-to-test, needs-ok-to-test, needs-more-info, needs-cherry-pick-release-5.4, needs-cherry-pick-release-6.1, needs-cherry-pick-release-6.5, needs-cherry-pick-release-7.1, needs-cherry-pick-release-7.5, needs-cherry-pick-release-8.1, affects-5.4, affects-6.1, affects-6.5, affects-7.1, affects-7.5, affects-8.1, may-affects-5.4, may-affects-6.1, may-affects-6.5, may-affects-7.1, may-affects-7.5, may-affects-8.1.

In response to this:

/label affects-8.0

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.