release v1.4.0

CHANGELOG/README.md

@@ -1,24 +1,42 @@
 # Changelog
 
+## 2023-03-23
+
+Release v1.4.0
+
++ Enhancement & New Features
+  + Support rebuild dn pod automatically
+  + Support consensus log backup
+  + Support point-in-time recovery(PITR)
+  + Support backup schedule
+  + Support restore without backup object
+  + Support debug runmode of PolarDB-X cluster
+  + Support default PrometheusRule for polardbx cluster.
+  + Use *Cluster IP* as the default communication address among dn pods
+  + Use container network as the default network type
++ Bug Fix
+  + Fix backup failure on follower when using non-host network
+  + Fix grafana cpu and mem usage issue for kube-state-metrics 2.3.0
+  + Fix restore issue when binlog checksum is none.
+
 ## 2023-03-17
 
 Release v1.4.0-beta
 
 + Enhancement & New Features
+  + Support rebuild dn pod automatically
   + Support consensus log backup
   + Support point-in-time recovery(PITR)
   + Support backup schedule
   + Support restore without backup object
-  + Support rebuild dn follower automatically when `Slave_SQL_Running` becomes `No` due to `Last_SQL_Error`
   + Support debug runmode of PolarDB-X cluster
-  + Support default PrometheusRule for polardbx cluster
+  + Support default PrometheusRule for polardbx cluster.
   + Use *Cluster IP* as the default communication address among dn pods
   + Use container network as the default network type
 + Bug Fix
   + Fix backup failure on follower when using non-host network
   + Fix grafana cpu and mem usage issue for kube-state-metrics 2.3.0
-  + Fix restore issue when binlog checksum is none
-  + Fix cdc crash during backup caused by sending heartbeat using XA transaction
+  + Fix restore issue when binlog checksum is none.
 
 ## 2022-11-17
 
@@ -81,7 +99,6 @@ Fix overriding my.cnf passing to the script without section.
 Release v1.2.1.
 
 This is a bugfix release.
-
 + Fix the CDC problem by setting the 'consensuslog_revise' variable of galaxyengine to ON.
 + Update docker registry mirror used in China mainland from ustc to sjtug.
 + Fix the scale in/out progress: use "schedule rebalance" command.
@@ -91,19 +108,11 @@ This is a bugfix release.
 Release v1.2.0.
 
 + Enhancement & New Features
-  + Provide a new CR `PolarDBXMonitor` for declaring the monitoring of some `PolarDBXCluster`. The controller will
-    create `ServiceMonitors` to make prometheus scraping the metrics.
-  + Provide a new chart `polardbx-monitor` which packages a
-    customized [kube-prometheus](https://github.com/prometheus-operator/kube-prometheus) release with predefined
-    dashboards for monitoring PolarDB-X clusters.
-  + Provide support for xpaxos version of [galaxyengine](https://github.com/ApsaraDB/galaxyengine). Now
-    the `PolarDBXCluster` controller will create a typical paxos cluster (leader + follower + logger) for each GMS and
-    DN by default.
-    + **Note** this is a breaking change. After the upgrade, the old `XStores` will be in an unmaintainable state.
-      Update of the GMS/DN is not possible due to incompatible data/log formats. You may have to delete all
-      the `PolarDBXCluster` in your Kubernetes before/after the upgrade.
-    + If you want to keep the compatibility, you can disable this feature by declaring feature gate with negative
-      symbol `EnableGalaxyCluster-`. After that, no multi-node `XStore` with galaxy engine can be created.
+  + Provide a new CR `PolarDBXMonitor` for declaring the monitoring of some `PolarDBXCluster`. The controller will create `ServiceMonitors` to make prometheus scraping the metrics.
+  + Provide a new chart `polardbx-monitor` which packages a customized [kube-prometheus](https://github.com/prometheus-operator/kube-prometheus) release with predefined dashboards for monitoring PolarDB-X clusters.
+  + Provide support for xpaxos version of [galaxyengine](https://github.com/ApsaraDB/galaxyengine). Now the `PolarDBXCluster` controller will create a typical paxos cluster (leader + follower + logger) for each GMS and DN by default. 
+    + **Note** this is a breaking change. After the upgrade, the old `XStores` will be in an unmaintainable state. Update of the GMS/DN is not possible due to incompatible data/log formats. You may have to delete all the `PolarDBXCluster` in your Kubernetes before/after the upgrade.
+    + If you want to keep the compatibility, you can disable this feature by declaring feature gate with negative symbol `EnableGalaxyCluster-`. After that, no multi-node `XStore` with galaxy engine can be created.
   + Support scaling up/down and self-healing of the `XStores`.
 
 + Bug Fix
@@ -114,7 +123,7 @@ Release v1.2.0.
 Release v1.1.0.
 
 + Enhancement & New Features
-  + Support scaling in/out the PolarDBXCluster.
+  + Support scaling in/out the PolarDBXCluster. 
   + Support configure SSL on PolarDBXCluster.
   + Add a new CRD PolarDBXCluster for setting and reviewing the configs of PolarDB-X cluster.
     + `config.dynamic.CN` is not going to sync with cluster while phase is running.
@@ -123,22 +132,21 @@ Release v1.1.0.
   + Add label "polardbx/name" to Services、Secrets and ConfigMaps owned by PolarDBXCluster.
   + Support webhooks for PolarDBXCluster and PolarDBXClusterKnobs.
   + Support the complete spec of node selectors in `spec.topology.rules` of PolarDBXCluster.
-  + Create headless services for pods of xstore. Record DNS domains instead of raw IP for records of DN in GMS and
-    among galaxyengine xstores.
+  + Create headless services for pods of xstore. Record DNS domains instead of raw IP for records of DN in GMS and among galaxyengine xstores. 
   + Support overwrite image tag in values.yaml (helm).
   + Support collect metrics for hotspot JVM 11.
   + Add e2e test tests.
 
 + Bug Fix
-  + Fix the wrong call stack when logging with `flow.Error` in some cases.
+  + Fix the wrong call stack when logging with `flow.Error` in some cases. 
   + Fix the wrong timeout in polardbx-init.
   + Fix configuring host path of data volumes in values.yaml (helm).
   + Fix removing ini keys in galaxyengine's config.
   + Fix a `removeNull` in hsperfdata.
 
-## 2021-10-15
+## 2021-10-15 
 
-Release v1.0.0.
+Release v1.0.0. 
 
 + Provide v1 APIs:
   + PolarDBXCluster

README.md

@@ -11,12 +11,14 @@ Follow the [快速开始](https://doc.polardbx.com/quickstart/topics/quickstart-
 
 ## Documentation
 
-Please refer to the [CHANGELOG](./CHANGELOG/README.md) for the notable changes for each version.**
+Please refer to the [CHANGELOG](./CHANGELOG.md) for the notable changes for each version.**
 
 Refer to the documentations for more details, such as CRD definitions and operation guides.
 
 + [简体中文](https://doc.polardbx.com/quickstart/topics/quickstart-k8s.html)
-+ [English](docs/en/index.md)
++ English(TODO)
+
+如对文档有疑问，可提Issue或者Pull Request到[文档仓库](https://github.com/polardb/polardbx-operator-docs)
 
 ## Roadmap
 
@@ -42,9 +44,7 @@ Implemented:
 
 Working in progress:
 
-+ [T0] Backup Plan
-+ [T1] Point in Time Recovery
-+ [T2] Dashboard
++ [T0] Dashboard
 + ...
 
 ## License

api/v1/polardbxbackupbinlog_types.go

@@ -8,9 +8,9 @@ import (
 
 type PolarDBXBackupBinlogSpec struct {
 	// +kubebuilder:validation:Required
-	PxcName string `json:"pxcName,omitempty"`
-	// +kubebuilder:validation:Required
-	PxcUid string `json:"pxcUid,omitempty"`
+	PxcName string `json:"pxcName"`
+	// +optional
+	PxcUid string `json:"pxcUid"`
 	// +kubebuilder:default=168
 	RemoteExpireLogHours intstr.IntOrString `json:"remoteExpireLogHours,omitempty"`
 	// +kubebuilder:default=7

api/v1/system_task_types.go

@@ -27,6 +27,7 @@ type SystemTaskSpec struct {
 	CnReplicas  int                         `json:"cnReplicas,omitempty"`
 	CnResources corev1.ResourceRequirements `json:"cnResources,omitempty"`
 	DnResources corev1.ResourceRequirements `json:"dnResources,omitempty"`
+	Nodes       []string                    `json:"nodes,omitempty"`
 }
 
 type SystemTaskStatus struct {

charts/polardbx-logcollector/Chart.yaml

@@ -2,8 +2,8 @@ apiVersion: v2
 name: polardbx-logcollector
 description: Helm chart of polardbx-operator logcollector plugin
 type: application
-version: 1.4.0-beta
-appVersion: v1.4.0-beta
+version: 1.4.0
+appVersion: v1.4.0
 keywords:
   - polardb-x
   - operator

charts/polardbx-monitor/Chart.yaml

@@ -2,8 +2,8 @@ apiVersion: v2
 name: polardbx-monitor
 description: Helm chart of polardbx-operator monitor plugin
 type: application
-version: 1.4.0-beta
-appVersion: v1.4.0-beta
+version: 1.4.0
+appVersion: v1.4.0
 keywords:
   - polardb-x
   - operator

charts/polardbx-operator/Chart.yaml

@@ -2,8 +2,8 @@ apiVersion: v2
 name: polardbx-operator
 description: Helm chart of polardbx-operator
 type: application
-version: 1.4.0-beta
-appVersion: v1.4.0-beta
+version: 1.4.0
+appVersion: v1.4.0
 keywords:
 - polardb-x
 - operator

charts/polardbx-operator/crds/polardbx.aliyun.com_polardbxbackupbinlogs.yaml

@@ -78,6 +78,8 @@ spec:
                       backup
                     type: string
                 type: object
+            required:
+            - pxcName
             type: object
           status:
             properties:

charts/polardbx-operator/crds/polardbx.aliyun.com_systemtasks.yaml

@@ -94,6 +94,10 @@ spec:
                       to an implementation-defined value. More info: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/'
                     type: object
                 type: object
+              nodes:
+                items:
+                  type: string
+                type: array
               taskType:
                 type: string
             type: object

charts/polardbx-operator/values.yaml

@@ -14,7 +14,7 @@ images:
   polardbxJob: polardbx-job
 
 # Default image tag. Use app version if not specified or 'latest' if useLatestImage is true.
-imageTag: v1.4.0-beta
+imageTag: v1.4.0
 
 # Uses the latest images for operator components.
 useLatestImage: false

pkg/hpfs/filestream/flowcontrol.go

@@ -113,6 +113,9 @@ func (f *FlowControlManger) LimitFlow(reader io.Reader, writer io.Writer, notify
 	var wg sync.WaitGroup
 	wg.Add(1)
 	go func() {
+		defer func() {
+			recover()
+		}()
 		defer wg.Done()
 		for {
 			bytes, ok := <-bufChan

pkg/operator/v1/systemtask/common/context.go

@@ -116,7 +116,7 @@ func (ctx *Context) GetXStoreByName(name string) (*v1.XStore, error) {
 
 func (ctx *Context) GetAllXStorePods() ([]corev1.Pod, error) {
 	pods := make([]corev1.Pod, 0)
-	for _, role := range []string{polardbxmeta.RoleDN, polardbxmeta.RoleGMS} {
+	for _, role := range []string{polardbxmeta.RoleDN} {
 		var podList corev1.PodList
 		err := ctx.Client().List(ctx.Context(), &podList, client.InNamespace(ctx.Namespace()), client.MatchingLabels(map[string]string{
 			polardbxmeta.LabelRole: role,
@@ -177,9 +177,23 @@ func (ctx *Context) GetNodeXStorePodMap(separateRole bool, logger bool) (map[str
 }
 
 func (ctx *Context) GetAllNodes() ([]corev1.Node, error) {
+	systemtask := ctx.MustGetSystemTask()
 	var nodeList corev1.NodeList
 	err := ctx.Client().List(ctx.Context(), &nodeList)
-	return nodeList.Items, err
+	result := make([]corev1.Node, 0)
+	if len(systemtask.Spec.Nodes) > 0 {
+		nodeMap := make(map[string]bool, len(systemtask.Spec.Nodes))
+		for _, nodeName := range systemtask.Spec.Nodes {
+			nodeMap[nodeName] = true
+		}
+		for _, item := range nodeList.Items {
+			if _, ok := nodeMap[item.Name]; ok {
+				result = append(result, item)
+			}
+		}
+	}
+
+	return result, err
 }
 
 func (rc *Context) SetControllerRef(obj client.Object) error {