Added configured cleanup function to remove the restored configmaps

Author: adshin21

docs/reference/operator_parameters.md

@@ -171,6 +171,9 @@ Those are top-level keys, containing both leaf keys and groups.
 * **repair_period**
   period between consecutive repair requests. The default is `5m`.
 
+* **pitr_backup_retention**
+  retention time for PITR backup config maps. The operator will clean up config maps older than the configured retention. The value is a duration string, e.g. "168h". The default is `168h`.
+
 * **set_memory_request_to_limit**
   Set `memory_request` to `memory_limit` for all Postgres clusters (the default
   value is also increased but configured `max_memory_request` can not be

manifests/operatorconfiguration.crd.yaml

@@ -113,6 +113,9 @@ spec:
               repair_period:
                 type: string
                 default: "5m"
+              pitr_backup_retention:
+                type: string
+                default: "168h"
               set_memory_request_to_limit:
                 type: boolean
                 default: false

manifests/postgresql-operator-default-configuration.yaml

@@ -19,6 +19,7 @@ configuration:
   min_instances: -1
   resync_period: 30m
   repair_period: 5m
+  pitr_backup_retention: 168h
   # set_memory_request_to_limit: false
   # sidecars:
   # - image: image:123

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

@@ -266,6 +266,7 @@ type OperatorConfigurationData struct {
 	Workers                       uint32                             `json:"workers,omitempty"`
 	ResyncPeriod                  Duration                           `json:"resync_period,omitempty"`
 	RepairPeriod                  Duration                           `json:"repair_period,omitempty"`
+	PitrBackupRetention           Duration                           `json:"pitr_backup_retention,omitempty"`
 	SetMemoryRequestToLimit       bool                               `json:"set_memory_request_to_limit,omitempty"`
 	ShmVolume                     *bool                              `json:"enable_shm_volume,omitempty"`
 	SidecarImages                 map[string]string                  `json:"sidecar_docker_images,omitempty"` // deprecated in favour of SidecarContainers

pkg/controller/postgresql.go

@@ -13,8 +13,8 @@ import (
 	"github.com/sirupsen/logrus"
 
 	v1 "k8s.io/api/core/v1"
-	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/api/errors"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/types"
 	"k8s.io/client-go/tools/cache"
 
@@ -27,8 +27,8 @@ import (
 )
 
 const (
-	restoreAnnotationKey      = "postgres-operator.zalando.org/action"
-	restoreAnnotationValue    = "restore-in-place"
+	restoreAnnotationKey   = "postgres-operator.zalando.org/action"
+	restoreAnnotationValue = "restore-in-place"
 )
 
 func (c *Controller) clusterResync(stopCh <-chan struct{}, wg *sync.WaitGroup) {
@@ -44,6 +44,9 @@ func (c *Controller) clusterResync(stopCh <-chan struct{}, wg *sync.WaitGroup) {
 			if err := c.processPendingRestores(); err != nil {
 				c.logger.Errorf("could not process pending restores: %v", err)
 			}
+			if err := c.cleanupRestores(); err != nil {
+				c.logger.Errorf("could not cleanup restores: %v", err)
+			}
 		case <-stopCh:
 			return
 		}
@@ -611,8 +614,6 @@ func (c *Controller) validateRestoreInPlace(pgOld, pgNew *acidv1.Postgresql) err
 	return nil
 }
 
-
-
 // handlerRestoreInPlace starts an asynchronous point-in-time-restore.
 // It creates a ConfigMap to store the state and then deletes the old Postgresql CR.
 func (c *Controller) handlerRestoreInPlace(pgOld, pgNew *acidv1.Postgresql) {
@@ -780,6 +781,44 @@ func (c *Controller) processSingleInProgressCm(cm v1.ConfigMap) error {
 	return nil
 }
 
+func (c *Controller) cleanupRestores() error {
+	c.logger.Debug("cleaning up old restore config maps")
+	namespace := c.opConfig.WatchedNamespace
+	if namespace == "" {
+		namespace = v1.NamespaceAll
+	}
+
+	cmList, err := c.KubeClient.ConfigMaps(namespace).List(context.TODO(), metav1.ListOptions{})
+	if err != nil {
+		return fmt.Errorf("could not list restore ConfigMaps: %v", err)
+	}
+
+	retention := c.opConfig.PitrBackupRetention
+	if retention <= 0 {
+		c.logger.Debugf("Pitr backup retention is not set, skipping cleanup")
+		return nil
+	}
+	c.logger.Debugf("Pitr backup retention is %s", retention.String())
+
+	for _, cm := range cmList.Items {
+		if !strings.HasPrefix(cm.Name, "pitr-") {
+			continue
+		}
+
+		age := time.Since(cm.CreationTimestamp.Time)
+		if age > retention {
+			c.logger.Infof("deleting old restore config map %q, age: %s", cm.Name, age.String())
+			err := c.KubeClient.ConfigMaps(cm.Namespace).Delete(context.TODO(), cm.Name, metav1.DeleteOptions{})
+			if err != nil {
+				c.logger.Errorf("could not delete config map %q: %v", cm.Name, err)
+				// continue with next cm
+			}
+		}
+	}
+
+	return nil
+}
+
 /*
 Ensures the pod service account and role bindings exists in a namespace
 before a PG cluster is created there so that a user does not have to deploy

pkg/util/config/config.go

@@ -18,6 +18,7 @@ type CRD struct {
 	ReadyWaitTimeout      time.Duration `name:"ready_wait_timeout" default:"30s"`
 	ResyncPeriod          time.Duration `name:"resync_period" default:"30m"`
 	RepairPeriod          time.Duration `name:"repair_period" default:"5m"`
+	PitrBackupRetention   time.Duration `name:"pitr_backup_retention" default:"168h"`
 	EnableCRDRegistration *bool         `name:"enable_crd_registration" default:"true"`
 	EnableCRDValidation   *bool         `name:"enable_crd_validation" default:"true"`
 	CRDCategories         []string      `name:"crd_categories" default:"all"`