fix dial-timeout not affected for client watch command

[chengjoey]

fix #18335

etcd/client/v3/client.go

Lines 327 to 330 in e7f5729

	var cancel context.CancelFunc
	dctx, cancel = context.WithTimeout(c.ctx, c.cfg.DialTimeout)
	defer cancel() // TODO: Is this right for cases where grpc.WithBlock() is not set on the dial options?
	}

As noted before, it is wrong to not have grpc.WithBlock()

in the test, we will also addoption grpc.WithBlock(

etcd/client/v3/client_test.go

Lines 104 to 118 in e7f5729

	// grpc.WithBlock to block until connection up or timeout
	testCfgs := []Config{
	{
	Endpoints: []string{"http://254.0.0.1:12345"},
	DialTimeout: 2 * time.Second,
	DialOptions: []grpc.DialOption{grpc.WithBlock()},
	},
	{
	Endpoints: []string{"http://254.0.0.1:12345"},
	DialTimeout: time.Second,
	DialOptions: []grpc.DialOption{grpc.WithBlock()},
	Username: "abc",
	Password: "def",
	},
	}

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: chengjoey
Once this PR has been reviewed and has the lgtm label, please assign wenjiaswe for approval. For more information see the Kubernetes Code Review Process.

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

Needs approval from an approver in each of these files:

• OWNERS

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

[k8s-ci-robot]

Hi @chengjoey. Thanks for your PR.

I'm waiting for a etcd-io 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-sigs/prow repository.

[jmhbnz]

/ok-to-test

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

Attention: Patch coverage is 0% with 14 lines in your changes missing coverage. Please review.

Project coverage is 68.77%. Comparing base (c86c93c) to head (edae2a1).
Report is 137 commits behind head on main.

❗ Current head edae2a1 differs from pull request most recent head 3d8fe91

Please upload reports for the commit 3d8fe91 to get more accurate results.

Files with missing lines	Patch %	Lines
etcdctl/ctlv3/command/global.go	0.00%	13 Missing ⚠️
etcdctl/ctlv3/command/watch_command.go	0.00%	1 Missing ⚠️

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

Files with missing lines	Coverage Δ
etcdctl/ctlv3/command/watch_command.go	44.08% <0.00%> (ø)
etcdctl/ctlv3/command/global.go	0.00% <0.00%> (ø)

... and 24 files with indirect coverage changes

@@           Coverage Diff           @@
##             main   #18336   +/-   ##
=======================================
  Coverage   68.77%   68.77%           
=======================================
  Files         420      420           
  Lines       35535    35548   +13     
=======================================
+ Hits        24438    24449   +11     
- Misses       9668     9673    +5     
+ Partials     1429     1426    -3     

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c86c93c...3d8fe91. Read the comment docs.

[chengjoey]

/test pull-etcd-unit-test-386
/test pull-etcd-unit-test-amd64
/test pull-etcd-e2e-amd64

[chengjoey]

/retest-required

[jberkus]

@chengjoey are you able to troubleshoot these test failures on your own? Or do you need help?

[chengjoey]

@chengjoey are you able to troubleshoot these test failures on your own? Or do you need help?

thanks @jberkus , i will troubleshoot the test failure in the next two days.

[ahrtr]

Thanks for raising this PR.

It's a valid fix, but it might break some of the existing user experience. I guess it might be the reason why some test cases failed. So I suggest to only fix this in main.

[chengjoey]

updates:
Modifying the original logic and adding Option grpc.WithBlock will indeed affect the old logic. For example, grpc_proxy must create a client before starting the server. At this time, setting Block will cause the server to never start.

etcd/server/etcdmain/grpc_proxy.go

Lines 239 to 269 in 010d462

	httpClient := mustNewHTTPClient()
	srvhttp, httpl := mustHTTPListener(lg, m, tlsInfo, client, proxyClient)
	if err := http2.ConfigureServer(srvhttp, &http2.Server{
	MaxConcurrentStreams: maxConcurrentStreams,
	}); err != nil {
	lg.Fatal("Failed to configure the http server", zap.Error(err))
	}
	errc := make(chan error, 3)
	go func() { errc <- newGRPCProxyServer(lg, client).Serve(grpcl) }()
	go func() { errc <- srvhttp.Serve(httpl) }()
	go func() { errc <- m.Serve() }()
	if len(grpcProxyMetricsListenAddr) > 0 {
	mhttpl := mustMetricsListener(lg, tlsInfo)
	go func() {
	mux := http.NewServeMux()
	grpcproxy.HandleMetrics(mux, httpClient, client.Endpoints())
	grpcproxy.HandleHealth(lg, mux, client)
	grpcproxy.HandleProxyMetrics(mux)
	grpcproxy.HandleProxyHealth(lg, mux, proxyClient)
	lg.Info("gRPC proxy server metrics URL serving")
	herr := http.Serve(mhttpl, mux)
	if herr != nil {
	lg.Fatal("gRPC proxy server metrics URL returned", zap.Error(herr))
	} else {
	lg.Info("gRPC proxy server metrics URL returned")
	}
	}()
	}

So I changed it to only watch_command to set the block when creating the client, and other commands remain the same.
This will only fix the watch dial-timeout, and will not affect other logic
@ahrtr @jberkus PTAL

[chengjoey]

Just like e2e test does，newClient will set grpc.WithBlock

etcd/tests/e2e/utils.go

Lines 50 to 54 in d6c0127

	ccfg := clientv3.Config{
	Endpoints: entpoints,
	DialTimeout: 5 * time.Second,
	DialOptions: []grpc.DialOption{grpc.WithBlock()},
	}

[jmhbnz]

Discussed during sig-etcd triage meeting, @chengjoey can you please rebase this to prepare it for review?

[chengjoey]

Discussed during sig-etcd triage meeting, @chengjoey can you please rebase this to prepare it for review?

done

[chengjoey]

/test pull-etcd-robustness-amd64

[jmhbnz]

Thanks for rebase, I'll review shortly, cc tech leads @ahrtr and @serathius to also review.