fix(bugs): fixed a lot of bugs

Changelog.md

@@ -22,6 +22,7 @@ adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
  * Add `backupCompressCommand` and `backupDecompressCommand` to allow using
    different compressors/decompressors when backing up or restoring.
  * Add support for MySQL version 8.0
+ * Add `go modules` cache
 ### Changed
  * Only add `binlog-space-limit` for `percona` image
  * Make user-defined InitContainer take the precedence
@@ -36,8 +37,12 @@ adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
  * Set default MySQL server version to `5.7.31`
  * Generate CRDs with controller-gen `v0.5.0`
  * Update client-go to `v0.20.4`
+ * Update `getOrdinalFromHostname` and `IsFirstPodInSet`
 ### Removed
 ### Fixed
+ * Fix insufficient permissions during startup
+ * Fix the `xtrabackup` `--tables-exclude` cannot take effect
+ * Fix the pod unable to connect `Orchestrator`
  * Fix pod labels diff of map
  * Fixed backup cleanup job bug (#577)
  * Fix Kubebuilder path in Makefile.

Dockerfile

@@ -5,9 +5,12 @@ FROM golang:1.16.0 as builder
 
 # Copy in the go src
 WORKDIR /go/src/github.com/presslabs/mysql-operator
+
+COPY go.mod go.sum ./
+RUN go mod download
+
 COPY pkg/    pkg/
 COPY cmd/    cmd/
-COPY go.mod go.sum ./
 
 # Build
 RUN CGO_ENABLED=0 GOOS=linux GOARCH=amd64 go build -a -o mysql-operator github.com/presslabs/mysql-operator/cmd/mysql-operator

Dockerfile.orchestrator

@@ -6,9 +6,12 @@ FROM golang:1.16.0 as builder
 
 # Copy in the go src
 WORKDIR /go/src/github.com/presslabs/mysql-operator
+
+COPY go.mod go.sum ./
+RUN go mod download
+
 COPY pkg/    pkg/
 COPY cmd/    cmd/
-COPY go.mod go.sum ./
 
 # Build
 RUN CGO_ENABLED=0 GOOS=linux GOARCH=amd64 go build -a -o mysql-operator github.com/presslabs/mysql-operator/cmd/mysql-operator

Dockerfile.sidecar

@@ -6,9 +6,12 @@ FROM golang:1.16.0 as builder
 
 # Copy in the go src
 WORKDIR /go/src/github.com/presslabs/mysql-operator
+
+COPY go.mod go.sum ./
+RUN go mod download
+
 COPY pkg/    pkg/
 COPY cmd/    cmd/
-COPY go.mod go.sum ./
 
 # Build
 RUN CGO_ENABLED=0 GOOS=linux GOARCH=amd64 go build -a -o mysql-operator github.com/presslabs/mysql-operator/cmd/mysql-operator

charts/mysql-cluster/templates/cluster.yaml

@@ -46,20 +46,20 @@ spec:
 
   {{- if .Values.mysqlConf }}
   mysqlConf:
-    {{ toYaml .Values.mysqlConf | nindent 4 }}
+    {{- toYaml .Values.mysqlConf | nindent 4 }}
   {{- end }}
 
   {{- if .Values.podSpec }}
   podSpec:
-    {{ toYaml .Values.podSpec | nindent 4 }}
+    {{- toYaml .Values.podSpec | nindent 4 }}
   {{- end }}
 
   {{- if .Values.volumeSpec }}
   volumeSpec:
-    {{ toYaml .Values.volumeSpec | nindent 4 }}
+    {{- toYaml .Values.volumeSpec | nindent 4 }}
   {{- end }}
 
   {{- if .Values.initFileExtraSQL }}
   initFileExtraSQL:
-    {{ toYaml .Values.initFileExtraSQL | nindent 4 }}
+    {{- toYaml .Values.initFileExtraSQL | nindent 6 }}
   {{- end }}

charts/mysql-operator/templates/statefulset.yaml

@@ -49,7 +49,7 @@ spec:
           args:
             - --leader-election-namespace={{ .Release.Namespace }}
             # connect to orchestrator on localhost
-            - --orchestrator-uri=http://127.0.0.1:3000/api
+            - --orchestrator-uri=http://{{ template "mysql-operator.fullname" . }}.{{ .Release.Namespace }}/api
             {{- if .Values.sidecarImage }}
             - --sidecar-image={{ .Values.sidecarImage }}
             {{- end -}}

pkg/controller/mysqlcluster/internal/syncer/config_map.go

@@ -81,7 +81,7 @@ if [ ${read_only_status} -eq 0  ] && [ ${replica_status_count} -eq 0 ] && [ ${ha
 then
 		masterhostname=$( curl  -s "${ORCH_HTTP_API}/master/${ORCH_CLUSTER_ALIAS}" |  awk -F":" '{print $3}' | awk -F'"' '{print $2}' )
         echo "master from orchestrator: ${masterhostname}"
-        if [ "${FQDN}" == "${masterhostname}" ]
+        if [ "${MY_FQDN}" == "${masterhostname}" ]
         then
                 curl  -s "${ORCH_HTTP_API}/graceful-master-takeover-auto/${ORCH_CLUSTER_ALIAS}"
 				echo "graceful-master-takeover-auto is ongoing, sleep 5 seconds in order to make sure service can work well."

pkg/orchestrator/orchestrator.go

@@ -91,7 +91,7 @@ func (o *orchestrator) Cluster(cluster string) ([]Instance, error) {
 }
 
 func (o *orchestrator) AuditRecovery(cluster string) ([]TopologyRecovery, error) {
-	path := fmt.Sprintf("audit-recovery/%s", cluster)
+	path := fmt.Sprintf("audit-recovery/cluster/%s", cluster)
 	var recoveries []TopologyRecovery
 	if err := o.makeGetRequest(path, &recoveries); err != nil {
 		return nil, err

pkg/sidecar/appconf.go

@@ -164,6 +164,7 @@ func initFileQuery(cfg *Config, gtidPurged string) []byte {
 	// configure operator utility user
 	queries = append(queries, createUserQuery(cfg.OperatorUser, cfg.OperatorPassword, "%",
 		[]string{"SUPER", "SHOW DATABASES", "PROCESS", "RELOAD", "CREATE", "SELECT"}, "*.*",
+		[]string{"REPLICATION SLAVE"}, "*.*",
 		[]string{"ALL"}, fmt.Sprintf("%s.*", toolsDbName))...)
 
 	// configure orchestrator user
@@ -195,9 +196,10 @@ func initFileQuery(cfg *Config, gtidPurged string) []byte {
 		[]string{"CREATE", "SELECT", "DELETE", "UPDATE", "INSERT"}, fmt.Sprintf("%s.%s", toolsDbName, toolsHeartbeatTableName),
 		[]string{"REPLICATION CLIENT"}, "*.*")...)
 
-	// the slave pod doesn't need to back up sys_operator.status, Xtrabackup might have some bugs that
-	// cause the table to be unclean. We can do this cleanup before the slave pod starts to avoid accidents.
-	if !cfg.IsFirstPodInSet() {
+	// when the status.ibd file does not exist
+	// need to delete the status table
+	_, err := os.Stat(path.Join(dataDir, constants.OperatorDbName, constants.OperatorStatusTableName+".ibd"))
+	if os.IsNotExist(err) {
 		queries = append(queries, fmt.Sprintf("DROP TABLE IF EXISTS %s.%s",
 			constants.OperatorDbName, constants.OperatorStatusTableName))
 	}

pkg/sidecar/configs.go

@@ -152,8 +152,7 @@ func (cfg *Config) MysqlDSN() string {
 
 // IsFirstPodInSet returns true if this pod has an ordinal of 0, meaning it is the first one in the set
 func (cfg *Config) IsFirstPodInSet() bool {
-	ordinal := getOrdinalFromHostname(cfg.Hostname)
-	return ordinal == 0
+	return getOrdinalFromHostname(cfg.Hostname) == 0
 }
 
 // ShouldCloneFromBucket returns true if it's time to initialize from a bucket URL provided
@@ -198,7 +197,7 @@ func (cfg *Config) XtrabackupArgs() []string {
 		"--backup",
 		"--slave-info",
 		"--stream=xbstream",
-		fmt.Sprintf("--tables-exclude=%s.%s", constants.OperatorDbName, constants.OperatorStatusTableName),
+		fmt.Sprintf("--tables-exclude=%s\\.%s", constants.OperatorDbName, constants.OperatorStatusTableName),
 		"--host=127.0.0.1",
 		fmt.Sprintf("--user=%s", cfg.ReplicationUser),
 		fmt.Sprintf("--password=%s", cfg.ReplicationPassword),
@@ -303,17 +302,17 @@ func getEnvValue(key string) string {
 }
 
 func getOrdinalFromHostname(hn string) int {
-	// mysql-master-1
-	// or
-	// stateful-ceva-3
-	l := strings.Split(hn, "-")
-	for i := len(l) - 1; i >= 0; i-- {
-		if o, err := strconv.ParseInt(l[i], 10, 8); err == nil {
-			return int(o)
-		}
+	begin := strings.LastIndex(hn, "-")
+	if begin < 0 {
+		return 0
+	}
+	index := hn[begin+1:]
+	o, err := strconv.Atoi(index)
+	if err != nil {
+		return -1
 	}
 
-	return 0
+	return o
 }
 
 // retryLookupHost tries to figure out a host IPs with retries