version: prohibit operating TiCDC clusters across major and minor ver…

…sions (#2481) (#2599)
pingcap · Aug 21, 2021 · fea259b · fea259b
1 parent eeecc30
commit fea259b
Show file tree

Hide file tree

Showing 30 changed files with 570 additions and 261 deletions.
diff --git a/.github/workflows/integration.yml b/.github/workflows/integration.yml
@@ -4,9 +4,9 @@ name: integration-tests
 # events but only for the master branch
 on:
   push:
-    branches: [ master ]
+    branches: [master]
   pull_request:
-    branches: [ master ]
+    branches: [master]
 
 # See: https://docs.github.com/en/actions/reference/workflow-syntax-for-github-actions#concurrency.
 concurrency:
@@ -20,16 +20,16 @@ jobs:
     steps:
       - name: Free disk space
         run: |
-         sudo rm -rf /usr/local/lib/android
-         sudo rm -rf /usr/share/dotnet
-         docker volume prune -f
-         docker image prune -f
+          sudo rm -rf /usr/local/lib/android
+          sudo rm -rf /usr/share/dotnet
+          docker volume prune -f
+          docker image prune -f
       # Checks-out your repository under $GITHUB_WORKSPACE, so your job can access it
       - uses: actions/checkout@v2
 
       - uses: actions/setup-go@v2
         with:
-          go-version: '1.16.4'
+          go-version: "1.16.4"
 
       - name: Cache Vendor
         id: cache-vendor
@@ -44,8 +44,11 @@ jobs:
       - name: Pull images
         run: docker-compose -f docker-compose-avro.yml pull --ignore-pull-failures
 
-      - name: TiKV version
-        run: docker run pingcap/tikv:release-4.0-nightly -V
+      - name: TiDB version
+        run: |
+          docker run pingcap/tidb:v5.1.0-nightly -V
+          docker run pingcap/tikv:v5.1.0-nightly -V
+          docker run pingcap/pd:v5.1.0-nightly -V
 
       - name: Build Integration Framework
         run: |
@@ -58,6 +61,10 @@ jobs:
           cd $GITHUB_WORKSPACE/integration
           ./integration -protocol=mysql
 
+      - name: Try to dump container logs
+        if: ${{ failure() }}
+        run: docker-compose -f docker-compose-avro.yml logs -t
+
       - uses: actions/upload-artifact@v2
         if: ${{ always() }}
         with:
@@ -75,16 +82,16 @@ jobs:
     steps:
       - name: Free disk space
         run: |
-         sudo rm -rf /usr/local/lib/android
-         sudo rm -rf /usr/share/dotnet
-         docker volume prune -f
-         docker image prune -f
+          sudo rm -rf /usr/local/lib/android
+          sudo rm -rf /usr/share/dotnet
+          docker volume prune -f
+          docker image prune -f
       # Checks-out your repository under $GITHUB_WORKSPACE, so your job can access it
       - uses: actions/checkout@v2
 
       - uses: actions/setup-go@v2
         with:
-          go-version: '1.16.4'
+          go-version: "1.16.4"
 
       - name: Cache Vendor
         id: cache-vendor
@@ -99,8 +106,11 @@ jobs:
       - name: Pull images
         run: docker-compose -f docker-compose-avro.yml pull --ignore-pull-failures
 
-      - name: TiKV version
-        run: docker run pingcap/tikv:release-4.0-nightly -V
+      - name: TiDB version
+        run: |
+          docker run pingcap/tidb:v5.1.0-nightly -V
+          docker run pingcap/tikv:v5.1.0-nightly -V
+          docker run pingcap/pd:v5.1.0-nightly -V
 
       - name: Build Integration Framework
         run: |
@@ -113,6 +123,10 @@ jobs:
           cd $GITHUB_WORKSPACE/integration
           ./integration -protocol=simple-mysql-checking-old-value
 
+      - name: Try to dump container logs
+        if: ${{ failure() }}
+        run: docker-compose -f docker-compose-avro.yml logs -t
+
       - uses: actions/upload-artifact@v2
         if: ${{ always() }}
         with:
@@ -130,16 +144,16 @@ jobs:
     steps:
       - name: Free disk space
         run: |
-         sudo rm -rf /usr/local/lib/android
-         sudo rm -rf /usr/share/dotnet
-         docker volume prune -f
-         docker image prune -f
+          sudo rm -rf /usr/local/lib/android
+          sudo rm -rf /usr/share/dotnet
+          docker volume prune -f
+          docker image prune -f
       # Checks-out your repository under $GITHUB_WORKSPACE, so your job can access it
       - uses: actions/checkout@v2
 
       - uses: actions/setup-go@v2
         with:
-          go-version: '1.16.4'
+          go-version: "1.16.4"
 
       - name: Cache Vendor
         id: cache-vendor
@@ -154,8 +168,11 @@ jobs:
       - name: Pull images
         run: docker-compose -f docker-compose-canal.yml pull --ignore-pull-failures
 
-      - name: TiKV version
-        run: docker run pingcap/tikv:release-4.0-nightly -V
+      - name: TiDB version
+        run: |
+          docker run pingcap/tidb:v5.1.0-nightly -V
+          docker run pingcap/tikv:v5.1.0-nightly -V
+          docker run pingcap/pd:v5.1.0-nightly -V
 
       - name: Build Integration Framework
         run: |
@@ -168,6 +185,10 @@ jobs:
           cd $GITHUB_WORKSPACE/integration
           ./integration -protocol=canalJson
 
+      - name: Try to dump container logs
+        if: ${{ failure() }}
+        run: docker-compose -f docker-compose-canal.yml logs -t
+
       - uses: actions/upload-artifact@v2
         if: ${{ always() }}
         with:
@@ -179,23 +200,22 @@ jobs:
         run: |
           $GITHUB_WORKSPACE/scripts/canal/canal-local-test.sh down
 
-
   Canal-integration:
     runs-on: ubuntu-latest
 
     steps:
       - name: Free disk space
         run: |
-         sudo rm -rf /usr/local/lib/android
-         sudo rm -rf /usr/share/dotnet
-         docker volume prune -f
-         docker image prune -f
+          sudo rm -rf /usr/local/lib/android
+          sudo rm -rf /usr/share/dotnet
+          docker volume prune -f
+          docker image prune -f
       # Checks-out your repository under $GITHUB_WORKSPACE, so your job can access it
       - uses: actions/checkout@v2
 
       - uses: actions/setup-go@v2
         with:
-          go-version: '1.16.4'
+          go-version: "1.16.4"
 
       - name: Cache Vendor
         id: cache-vendor
@@ -210,8 +230,11 @@ jobs:
       - name: Pull images
         run: docker-compose -f docker-compose-canal.yml pull --ignore-pull-failures
 
-      - name: TiKV version
-        run: docker run pingcap/tikv:release-4.0-nightly -V
+      - name: TiDB version
+        run: |
+          docker run pingcap/tidb:v5.1.0-nightly -V
+          docker run pingcap/tikv:v5.1.0-nightly -V
+          docker run pingcap/pd:v5.1.0-nightly -V
 
       - name: Build Integration Framework
         run: |
@@ -224,6 +247,10 @@ jobs:
           cd $GITHUB_WORKSPACE/integration
           ./integration -protocol=canal
 
+      - name: Try to dump container logs
+        if: ${{ failure() }}
+        run: docker-compose -f docker-compose-canal.yml logs -t
+
       - uses: actions/upload-artifact@v2
         if: ${{ always() }}
         with:
@@ -241,16 +268,16 @@ jobs:
     steps:
       - name: Free disk space
         run: |
-         sudo rm -rf /usr/local/lib/android
-         sudo rm -rf /usr/share/dotnet
-         docker volume prune -f
-         docker image prune -f
+          sudo rm -rf /usr/local/lib/android
+          sudo rm -rf /usr/share/dotnet
+          docker volume prune -f
+          docker image prune -f
       # Checks-out your repository under $GITHUB_WORKSPACE, so your job can access it
       - uses: actions/checkout@v2
 
       - uses: actions/setup-go@v2
         with:
-          go-version: '1.16.4'
+          go-version: "1.16.4"
 
       - name: Cache Vendor
         id: cache-vendor
@@ -265,8 +292,11 @@ jobs:
       - name: Pull images
         run: docker-compose -f docker-compose-avro.yml pull --ignore-pull-failures
 
-      - name: TiKV version
-        run: docker run pingcap/tikv:release-4.0-nightly -V
+      - name: TiDB version
+        run: |
+          docker run pingcap/tidb:v5.1.0-nightly -V
+          docker run pingcap/tikv:v5.1.0-nightly -V
+          docker run pingcap/pd:v5.1.0-nightly -V
 
       - name: Build Integration Framework
         run: |
@@ -279,6 +309,10 @@ jobs:
           cd $GITHUB_WORKSPACE/integration
           ./integration
 
+      - name: Try to dump container logs
+        if: ${{ failure() }}
+        run: docker-compose -f docker-compose-avro.yml logs -t
+
       - uses: actions/upload-artifact@v2
         if: ${{ always() }}
         with:

diff --git a/Dockerfile.development b/Dockerfile.development
@@ -1,4 +1,4 @@
-FROM golang:1.16-alpine as builder
+FROM golang:1.16-alpine3.12 as builder
 RUN apk add --no-cache git make bash
 WORKDIR /go/src/github.com/pingcap/ticdc
 COPY . .

diff --git a/cmd/client.go b/cmd/client.go
@@ -78,6 +78,7 @@ var (
 	captureID               string
 	interval                uint
 	disableGCSafePointCheck bool
+	disableVersionCheck     bool
 
 	syncPointEnabled  bool
 	syncPointInterval time.Duration
@@ -126,6 +127,7 @@ type processorMeta struct {
 }
 
 func newCliCommand() *cobra.Command {
+	verifyCDCClusterVersionList := make([]string, 0)
 	command := &cobra.Command{
 		Use:   "cli",
 		Short: "Manage replication task and TiCDC cluster",
@@ -195,18 +197,25 @@ func newCliCommand() *cobra.Command {
 			if err != nil {
 				return errors.Annotatef(err, "fail to open PD client, pd=\"%s\"", cliPdAddr)
 			}
-			ctx := defaultContext
-			errorTiKVIncompatible := true // Error if TiKV is incompatible.
-			for _, pdEndpoint := range pdEndpoints {
-				err = version.CheckClusterVersion(ctx, pdCli, pdEndpoint, credential, errorTiKVIncompatible)
-				if err == nil {
-					break
+			checkVersion := needVerifyVersion(cmd, verifyCDCClusterVersionList)
+			if checkVersion && !disableVersionCheck {
+				ctx := defaultContext
+				_, err = verifyAndGetTiCDCClusterVersion(ctx, cdcEtcdCli)
+				if err != nil {
+					return err
+				}
+				errorTiKVIncompatible := true // Error if TiKV is incompatible.
+				for _, pdEndpoint := range pdEndpoints {
+					err = version.CheckClusterVersion(
+						ctx, pdCli, pdEndpoint, credential, errorTiKVIncompatible)
+					if err == nil {
+						break
+					}
+				}
+				if err != nil {
+					return err
 				}
 			}
-			if err != nil {
-				return err
-			}
-
 			return nil
 		},
 		Run: func(cmd *cobra.Command, args []string) {
@@ -215,14 +224,24 @@ func newCliCommand() *cobra.Command {
 			}
 		},
 	}
-	command.AddCommand(
-		newCaptureCommand(),
-		newChangefeedCommand(),
-		newProcessorCommand(),
-		newUnsafeCommand(),
-		newTsoCommand(),
-	)
 
+	capture := newCaptureCommand()
+	command.AddCommand(capture)
+	verifyCDCClusterVersionList = append(verifyCDCClusterVersionList, capture.Name())
+
+	changefeed := newChangefeedCommand()
+	command.AddCommand(changefeed)
+	verifyCDCClusterVersionList = append(verifyCDCClusterVersionList, changefeed.Name())
+
+	processor := newProcessorCommand()
+	command.AddCommand(processor)
+	verifyCDCClusterVersionList = append(verifyCDCClusterVersionList, processor.Name())
+
+	// Disable version check to ease internal testing.
+	command.PersistentFlags().BoolVar(&disableVersionCheck, "disable-version-check", false, "Disable version check")
+	_ = command.PersistentFlags().MarkHidden("disable-version-check")
+
+	command.AddCommand(newUnsafeCommand(), newTsoCommand())
 	return command
 }
 

diff --git a/cmd/client_changefeed.go b/cmd/client_changefeed.go
@@ -242,7 +242,10 @@ func newQueryChangefeedCommand() *cobra.Command {
 	return command
 }
 
-func verifyChangefeedParameters(ctx context.Context, cmd *cobra.Command, isCreate bool, credential *security.Credential, captureInfos []*model.CaptureInfo) (*model.ChangeFeedInfo, error) {
+func verifyChangefeedParameters(
+	ctx context.Context, cmd *cobra.Command, isCreate bool,
+	credential *security.Credential, cdcClusterVer version.TiCDCClusterVersion,
+) (*model.ChangeFeedInfo, error) {
 	if isCreate {
 		if sinkURI == "" {
 			return nil, errors.New("Creating changefeed without a sink-uri")
@@ -264,10 +267,6 @@ func verifyChangefeedParameters(ctx context.Context, cmd *cobra.Command, isCreat
 			return nil, err
 		}
 	}
-	cdcClusterVer, err := version.GetTiCDCClusterVersion(captureInfos)
-	if err != nil {
-		return nil, errors.Trace(err)
-	}
 	cfg := config.GetDefaultReplicaConfig()
 
 	if !cdcClusterVer.ShouldEnableOldValueByDefault() {
@@ -474,7 +473,11 @@ func newCreateChangefeedCommand() *cobra.Command {
 			if err != nil {
 				return err
 			}
-			info, err := verifyChangefeedParameters(ctx, cmd, true /* isCreate */, getCredential(), captureInfos)
+			cdcClusterVer, err := version.GetTiCDCClusterVersion(captureInfos)
+			if err != nil {
+				return err
+			}
+			info, err := verifyChangefeedParameters(ctx, cmd, true /* isCreate */, getCredential(), cdcClusterVer)
 			if err != nil {
 				return err
 			}