Merge branch 'master' into coderanger-disruption-budget

CODEOWNERS

@@ -1,2 +1,2 @@
 # global owners
-*        @alexeyklyukin @erthalion @sdudoladov @Jan-M @CyberDem0n @avaczi @FxKu
+*        @alexeyklyukin @erthalion @sdudoladov @Jan-M @CyberDem0n @avaczi @FxKu @RafiaSabih

docker/Dockerfile

@@ -6,4 +6,9 @@ RUN apk --no-cache add ca-certificates
 
 COPY build/* /
 
+RUN addgroup -g 1000 pgo
+RUN adduser -D -u 1000 -G pgo -g 'Postgres operator' pgo
+
+USER 1000:1000
+
 ENTRYPOINT ["/postgres-operator"]

docs/reference/cluster_manifest.md

@@ -254,6 +254,19 @@ under the `clone` top-level key and do not affect the already running cluster.
   timestamp. When this parameter is set the operator will not consider cloning
   from the live cluster, even if it is running, and instead goes to S3. Optional.
 
+* **s3_endpoint**
+  the url of the S3-compatible service should be set when cloning from non AWS S3. Optional.
+
+* **s3_access_key_id**
+  the access key id, used for authentication on S3 service. Optional.
+
+* **s3_secret_access_key**
+  the secret access key, used for authentication on S3 service. Optional.
+
+* **s3_force_path_style**
+  to enable path-style addressing(i.e., http://s3.amazonaws.com/BUCKET/KEY) when connecting to an S3-compatible service
+  that lack of support for sub-domain style bucket URLs (i.e., http://BUCKET.s3.amazonaws.com/KEY). Optional.
+
 ### EBS volume resizing
 
 Those parameters are grouped under the `volume` top-level key and define the

docs/reference/operator_parameters.md

@@ -32,7 +32,7 @@ configuration.
   kubectl create -f manifests/postgresql-operator-default-configuration.yaml
   kubectl get operatorconfigurations postgresql-operator-default-configuration -o yaml
   ```
-  Note that the operator first registers the CRD of the `OperatorConfiguration`
+  Note that the operator first attempts to register the CRD of the `OperatorConfiguration`
   and then waits for an instance to be created. In between these two event the
   operator pod may be failing since it cannot fetch the not-yet-existing
   `OperatorConfiguration` instance.

docs/user.md

@@ -41,6 +41,15 @@ $ kubectl create -f manifests/minimal-postgres-manifest.yaml
 $ kubectl get pods -w --show-labels
 ```
 
+## Give K8S users access to create/list postgresqls
+
+```bash
+$ kubectl create -f manifests/user-facing-clusterroles.yaml
+```
+
+Creates zalando-postgres-operator:users:view, :edit and :admin clusterroles that are
+aggregated into the default roles.
+
 ## Connect to PostgreSQL
 
 With a `port-forward` on one of the database pods (e.g. the master) you can
@@ -254,6 +263,24 @@ metadata:
 Note that timezone is required for `timestamp`. Otherwise, offset is relative
 to UTC, see [RFC 3339 section 5.6) 3339 section 5.6](https://www.ietf.org/rfc/rfc3339.txt).
 
+For non AWS S3 following settings can be set to support cloning from other S3 implementations:
+
+```yaml
+apiVersion: "acid.zalan.do/v1"
+kind: postgresql
+metadata:
+  name: acid-test-cluster
+spec:
+  clone:
+    uid: "efd12e58-5786-11e8-b5a7-06148230260c"
+    cluster: "acid-batman"
+    timestamp: "2017-12-19T12:40:33+01:00"
+    s3_endpoint: https://s3.acme.org
+    s3_access_key_id: 0123456789abcdef0123456789abcdef
+    s3_secret_access_key: 0123456789abcdef0123456789abcdef
+    s3_force_path_style: true
+```
+
 ## Sidecar Support
 
 Each cluster can specify arbitrary sidecars to run. These containers could be used for

manifests/complete-postgres-manifest.yaml

@@ -11,6 +11,7 @@ spec:
   teamId: "ACID"
   volume:
     size: 1Gi
+    #storageClass: my-sc
   numberOfInstances: 2
   users: #Application/Robot users
     zalando:
@@ -46,13 +47,13 @@ spec:
     pg_hba:
     - hostssl all all 0.0.0.0/0 md5
     - host    all all 0.0.0.0/0 md5
-    slots:
-      permanent_physical_1:
-        type: physical
-      permanent_logical_1:
-        type: logical
-        database: foo
-        plugin: pgoutput
+    #slots:
+    #  permanent_physical_1:
+    #    type: physical
+    #  permanent_logical_1:
+    #    type: logical
+    #    database: foo
+    #    plugin: pgoutput
     ttl: 30
     loop_wait: &loop_wait 10
     retry_timeout: 10

manifests/postgres-operator.yaml

@@ -21,6 +21,10 @@ spec:
           limits:
             cpu: 2000m
             memory: 500Mi
+        securityContext:
+          runAsUser: 1000
+          runAsNonRoot: true
+          readOnlyRootFilesystem: true
         env:
         # provided additional ENV vars can overwrite individual config map entries  
         - name: CONFIG_MAP_NAME

manifests/user-facing-clusterroles.yaml

@@ -0,0 +1,51 @@
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRole
+metadata:
+  labels:
+    rbac.authorization.k8s.io/aggregate-to-admin: "true"
+  name: zalando-postgres-operator:users:admin
+rules:
+- apiGroups:
+  - acid.zalan.do
+  resources:
+  - postgresqls
+  - postgresqls/status
+  verbs:
+  - "*"
+
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRole
+metadata:
+  labels:
+    rbac.authorization.k8s.io/aggregate-to-edit: "true"
+  name: zalando-postgres-operator:users:edit
+rules:
+- apiGroups:
+  - acid.zalan.do
+  resources:
+  - postgresqls
+  verbs:
+  - create
+  - update
+  - patch
+  - delete
+
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRole
+metadata:
+  labels:
+    rbac.authorization.k8s.io/aggregate-to-view: "true"
+  name: zalando-postgres-operator:users:view
+rules:
+- apiGroups:
+  - acid.zalan.do
+  resources:
+  - postgresqls
+  - postgresqls/status
+  verbs:
+  - get
+  - list
+  - watch
+

pkg/apis/acid.zalan.do/v1/postgresql_type.go

@@ -115,10 +115,14 @@ type Patroni struct {
 
 // CloneDescription describes which cluster the new should clone and up to which point in time
 type CloneDescription struct {
-	ClusterName  string `json:"cluster,omitempty"`
-	UID          string `json:"uid,omitempty"`
-	EndTimestamp string `json:"timestamp,omitempty"`
-	S3WalPath    string `json:"s3_wal_path,omitempty"`
+	ClusterName       string `json:"cluster,omitempty"`
+	UID               string `json:"uid,omitempty"`
+	EndTimestamp      string `json:"timestamp,omitempty"`
+	S3WalPath         string `json:"s3_wal_path,omitempty"`
+	S3Endpoint        string `json:"s3_endpoint,omitempty"`
+	S3AccessKeyId     string `json:"s3_access_key_id,omitempty"`
+	S3SecretAccessKey string `json:"s3_secret_access_key,omitempty"`
+	S3ForcePathStyle  *bool  `json:"s3_force_path_style,omitempty" defaults:"false"`
 }
 
 // Sidecar defines a container to be run in the same pod as the Postgres container.

pkg/apis/acid.zalan.do/v1/util_test.go

@@ -61,12 +61,12 @@ var cloneClusterDescriptions = []struct {
 	in  *CloneDescription
 	err error
 }{
-	{&CloneDescription{"foo+bar", "", "NotEmpty", ""}, nil},
-	{&CloneDescription{"foo+bar", "", "", ""},
+	{&CloneDescription{"foo+bar", "", "NotEmpty", "", "", "", "", nil}, nil},
+	{&CloneDescription{"foo+bar", "", "", "", "", "", "", nil},
 		errors.New(`clone cluster name must confirm to DNS-1035, regex used for validation is "^[a-z]([-a-z0-9]*[a-z0-9])?$"`)},
-	{&CloneDescription{"foobar123456789012345678901234567890123456789012345678901234567890", "", "", ""},
+	{&CloneDescription{"foobar123456789012345678901234567890123456789012345678901234567890", "", "", "", "", "", "", nil},
 		errors.New("clone cluster name must be no longer than 63 characters")},
-	{&CloneDescription{"foobar", "", "", ""}, nil},
+	{&CloneDescription{"foobar", "", "", "", "", "", "", nil}, nil},
 }
 
 var maintenanceWindows = []struct {

pkg/apis/acid.zalan.do/v1/zz_generated.deepcopy.go

@@ -50,6 +50,11 @@ func (in *AWSGCPConfiguration) DeepCopy() *AWSGCPConfiguration {
 // DeepCopyInto is an autogenerated deepcopy function, copying the receiver, writing into out. in must be non-nil.
 func (in *CloneDescription) DeepCopyInto(out *CloneDescription) {
 	*out = *in
+	if in.S3ForcePathStyle != nil {
+		in, out := &in.S3ForcePathStyle, &out.S3ForcePathStyle
+		*out = new(bool)
+		**out = **in
+	}
 	return
 }
 
@@ -464,7 +469,7 @@ func (in *PostgresSpec) DeepCopyInto(out *PostgresSpec) {
 			(*in)[i].DeepCopyInto(&(*out)[i])
 		}
 	}
-	out.Clone = in.Clone
+	in.Clone.DeepCopyInto(&out.Clone)
 	if in.Databases != nil {
 		in, out := &in.Databases, &out.Databases
 		*out = make(map[string]string, len(*in))

pkg/cluster/cluster.go

@@ -342,7 +342,7 @@ func (c *Cluster) compareStatefulSetWith(statefulSet *v1beta1.StatefulSet) *comp
 	if c.Statefulset.Spec.Template.Spec.ServiceAccountName != statefulSet.Spec.Template.Spec.ServiceAccountName {
 		needsReplace = true
 		needsRollUpdate = true
-		reasons = append(reasons, "new statefulset's serviceAccountName service asccount name doesn't match the current one")
+		reasons = append(reasons, "new statefulset's serviceAccountName service account name doesn't match the current one")
 	}
 	if *c.Statefulset.Spec.Template.Spec.TerminationGracePeriodSeconds != *statefulSet.Spec.Template.Spec.TerminationGracePeriodSeconds {
 		needsReplace = true
@@ -462,16 +462,16 @@ func (c *Cluster) compareContainers(description string, setA, setB []v1.Containe
 func compareResources(a *v1.ResourceRequirements, b *v1.ResourceRequirements) bool {
 	equal := true
 	if a != nil {
-		equal = compareResoucesAssumeFirstNotNil(a, b)
+		equal = compareResourcesAssumeFirstNotNil(a, b)
 	}
 	if equal && (b != nil) {
-		equal = compareResoucesAssumeFirstNotNil(b, a)
+		equal = compareResourcesAssumeFirstNotNil(b, a)
 	}
 
 	return equal
 }
 
-func compareResoucesAssumeFirstNotNil(a *v1.ResourceRequirements, b *v1.ResourceRequirements) bool {
+func compareResourcesAssumeFirstNotNil(a *v1.ResourceRequirements, b *v1.ResourceRequirements) bool {
 	if b == nil || (len(b.Requests) == 0) {
 		return len(a.Requests) == 0
 	}
@@ -884,7 +884,7 @@ func (c *Cluster) initInfrastructureRoles() error {
 	return nil
 }
 
-// resolves naming conflicts between existing and new roles by chosing either of them.
+// resolves naming conflicts between existing and new roles by choosing either of them.
 func (c *Cluster) resolveNameConflict(currentRole, newRole *spec.PgUser) spec.PgUser {
 	var result spec.PgUser
 	if newRole.Origin >= currentRole.Origin {
@@ -978,7 +978,7 @@ func (c *Cluster) Switchover(curMaster *v1.Pod, candidate spec.NamespacedName) e
 	// signal the role label waiting goroutine to close the shop and go home
 	close(stopCh)
 	// wait until the goroutine terminates, since unregisterPodSubscriber
-	// must be called before the outer return; otherwsise we risk subscribing to the same pod twice.
+	// must be called before the outer return; otherwise we risk subscribing to the same pod twice.
 	wg.Wait()
 	// close the label waiting channel no sooner than the waiting goroutine terminates.
 	close(podLabelErr)

pkg/cluster/k8sres.go

@@ -329,7 +329,7 @@ func tolerations(tolerationsSpec *[]v1.Toleration, podToleration map[string]stri
 	return []v1.Toleration{}
 }
 
-// isBootstrapOnlyParameter checks asgainst special Patroni bootstrap parameters.
+// isBootstrapOnlyParameter checks against special Patroni bootstrap parameters.
 // Those parameters must go to the bootstrap/dcs/postgresql/parameters section.
 // See http://patroni.readthedocs.io/en/latest/dynamic_configuration.html.
 func isBootstrapOnlyParameter(param string) bool {
@@ -1266,6 +1266,29 @@ func (c *Cluster) generateCloneEnvironment(description *acidv1.CloneDescription)
 		result = append(result, v1.EnvVar{Name: "CLONE_METHOD", Value: "CLONE_WITH_WALE"})
 		result = append(result, v1.EnvVar{Name: "CLONE_TARGET_TIME", Value: description.EndTimestamp})
 		result = append(result, v1.EnvVar{Name: "CLONE_WAL_BUCKET_SCOPE_PREFIX", Value: ""})
+
+		if description.S3Endpoint != "" {
+		   result = append(result, v1.EnvVar{Name: "CLONE_AWS_ENDPOINT", Value: description.S3Endpoint})
+		   result = append(result, v1.EnvVar{Name: "CLONE_WALE_S3_ENDPOINT", Value: description.S3Endpoint})
+		}
+
+		if description.S3AccessKeyId != "" {
+		   result = append(result, v1.EnvVar{Name: "CLONE_AWS_ACCESS_KEY_ID", Value: description.S3AccessKeyId})
+		}
+
+		if description.S3SecretAccessKey != "" {
+		   result = append(result, v1.EnvVar{Name: "CLONE_AWS_SECRET_ACCESS_KEY", Value: description.S3SecretAccessKey})
+		}
+
+		if description.S3ForcePathStyle != nil {
+		   s3ForcePathStyle := "0"
+
+		   if *description.S3ForcePathStyle {
+			   s3ForcePathStyle = "1"
+		   }
+
+		   result = append(result, v1.EnvVar{Name: "CLONE_AWS_S3_FORCE_PATH_STYLE", Value: s3ForcePathStyle})
+		}
 	}
 
 	return result
@@ -1372,7 +1395,7 @@ func (c *Cluster) generateLogicalBackupJob() (*batchv1beta1.CronJob, error) {
 		return nil, fmt.Errorf("could not generate pod template for logical backup pod: %v", err)
 	}
 
-	// overwrite specifc params of logical backups pods
+	// overwrite specific params of logical backups pods
 	podTemplate.Spec.Affinity = &podAffinity
 	podTemplate.Spec.RestartPolicy = "Never" // affects containers within a pod
 

pkg/cluster/resources.go

@@ -361,7 +361,7 @@ func (c *Cluster) updateService(role PostgresRole, newService *v1.Service) error
 	// TODO: check if it possible to change the service type with a patch in future versions of Kubernetes
 	if newService.Spec.Type != c.Services[role].Spec.Type {
 		// service type has changed, need to replace the service completely.
-		// we cannot use just pach the current service, since it may contain attributes incompatible with the new type.
+		// we cannot use just patch the current service, since it may contain attributes incompatible with the new type.
 		var (
 			currentEndpoint *v1.Endpoints
 			err             error
@@ -369,7 +369,7 @@ func (c *Cluster) updateService(role PostgresRole, newService *v1.Service) error
 
 		if role == Master {
 			// for the master service we need to re-create the endpoint as well. Get the up-to-date version of
-			// the addresses stored in it before the service is deleted (deletion of the service removes the endpooint)
+			// the addresses stored in it before the service is deleted (deletion of the service removes the endpoint)
 			currentEndpoint, err = c.KubeClient.Endpoints(c.Namespace).Get(c.endpointName(role), metav1.GetOptions{})
 			if err != nil {
 				return fmt.Errorf("could not get current cluster %s endpoints: %v", role, err)

pkg/controller/controller.go

@@ -197,25 +197,25 @@ func (c *Controller) initRoleBinding() {
 	// operator binds it to the cluster role with sufficient privileges
 	// we assume the role is created by the k8s administrator
 	if c.opConfig.PodServiceAccountRoleBindingDefinition == "" {
-		c.opConfig.PodServiceAccountRoleBindingDefinition = `
+		c.opConfig.PodServiceAccountRoleBindingDefinition = fmt.Sprintf(`
 		{
 			"apiVersion": "rbac.authorization.k8s.io/v1beta1",
 			"kind": "RoleBinding",
 			"metadata": {
-				   "name": "zalando-postgres-operator"
+				   "name": "%s"
 			},
 			"roleRef": {
 				"apiGroup": "rbac.authorization.k8s.io",
 				"kind": "ClusterRole",
-				"name": "zalando-postgres-operator"
+				"name": "%s"
 			},
 			"subjects": [
 				{
 					"kind": "ServiceAccount",
-					"name": "operator"
+					"name": "%s"
 				}
 			]
-		}`
+		}`, c.PodServiceAccount.Name, c.PodServiceAccount.Name, c.PodServiceAccount.Name)
 	}
 	c.logger.Info("Parse role bindings")
 	// re-uses k8s internal parsing. See k8s client-go issue #193 for explanation
@@ -230,9 +230,6 @@ func (c *Controller) initRoleBinding() {
 	default:
 		c.PodServiceAccountRoleBinding = obj.(*rbacv1beta1.RoleBinding)
 		c.PodServiceAccountRoleBinding.Namespace = ""
-		c.PodServiceAccountRoleBinding.ObjectMeta.Name = c.PodServiceAccount.Name
-		c.PodServiceAccountRoleBinding.RoleRef.Name = c.PodServiceAccount.Name
-		c.PodServiceAccountRoleBinding.Subjects[0].Name = c.PodServiceAccount.Name
 		c.logger.Info("successfully parsed")
 
 	}

pkg/controller/util.go

@@ -51,17 +51,18 @@ func (c *Controller) clusterWorkerID(clusterName spec.NamespacedName) uint32 {
 
 func (c *Controller) createOperatorCRD(crd *apiextv1beta1.CustomResourceDefinition) error {
 	if _, err := c.KubeClient.CustomResourceDefinitions().Create(crd); err != nil {
-		if !k8sutil.ResourceAlreadyExists(err) {
-			return fmt.Errorf("could not create customResourceDefinition: %v", err)
-		}
-		c.logger.Infof("customResourceDefinition %q is already registered and will only be updated", crd.Name)
+		if k8sutil.ResourceAlreadyExists(err) {
+			c.logger.Infof("customResourceDefinition %q is already registered and will only be updated", crd.Name)
 
-		patch, err := json.Marshal(crd)
-		if err != nil {
-			return fmt.Errorf("could not marshal new customResourceDefintion: %v", err)
-		}
-		if _, err := c.KubeClient.CustomResourceDefinitions().Patch(crd.Name, types.MergePatchType, patch); err != nil {
-			return fmt.Errorf("could not update customResourceDefinition: %v", err)
+			patch, err := json.Marshal(crd)
+			if err != nil {
+				return fmt.Errorf("could not marshal new customResourceDefintion: %v", err)
+			}
+			if _, err := c.KubeClient.CustomResourceDefinitions().Patch(crd.Name, types.MergePatchType, patch); err != nil {
+				return fmt.Errorf("could not update customResourceDefinition: %v", err)
+			}
+		} else {
+			c.logger.Errorf("could not create customResourceDefinition %q: %v", crd.Name, err)
 		}
 	} else {
 		c.logger.Infof("customResourceDefinition %q has been registered", crd.Name)