Skip to content

Latest commit

 

History

History
1472 lines (1189 loc) · 56.5 KB

administrator.md

File metadata and controls

1472 lines (1189 loc) · 56.5 KB

Administrator Guide

Learn how to configure and manage the Postgres Operator in your Kubernetes (K8s) environment.

CRD registration and validation

On startup, the operator will try to register the necessary CustomResourceDefinitions Postgresql and OperatorConfiguration. The latter will only get created if the POSTGRES_OPERATOR_CONFIGURATION_OBJECT environment variable is set in the deployment yaml and is not empty. If the CRDs already exists they will only be patched. If you do not wish the operator to create or update the CRDs set enable_crd_registration config option to false.

CRDs are defined with a openAPIV3Schema structural schema against which new manifests of postgresql or OperatorConfiguration resources will be validated. On creation you can bypass the validation with kubectl create --validate=false.

By default, the operator will register the CRDs in the all category so that resources are listed on kubectl get all commands. The crd_categories config option allows for customization of categories.

Upgrading the operator

The Postgres Operator is upgraded by changing the docker image within the deployment. Before doing so, it is recommended to check the release notes for new configuration options or changed behavior you might want to reflect in the ConfigMap or config CRD. E.g. a new feature might get introduced which is enabled or disabled by default and you want to change it to the opposite with the corresponding flag option.

When using helm, be aware that installing the new chart will not update the Postgresql and OperatorConfiguration CRD. Make sure to update them before with the provided manifests in the crds folder. Otherwise, you might face errors about new Postgres manifest or configuration options being unknown to the CRD schema validation.

Minor and major version upgrade

Minor version upgrades for PostgreSQL are handled via updating the Spilo Docker image. The operator will carry out a rolling update of Pods which includes a switchover (planned failover) of the master to the Pod with new minor version. The switch should usually take less than 5 seconds, still clients have to reconnect.

Upgrade on cloning

With cloning, the new cluster manifest must have a higher version string than the source cluster and will be created from a basebackup. Depending of the cluster size, downtime in this case can be significant as writes to the database should be stopped and all WAL files should be archived first before cloning is started. Therefore, use cloning only to test major version upgrades and check for compatibility of your app with to Postgres server of a higher version.

In-place major version upgrade

Starting with Spilo 13, Postgres Operator can run an in-place major version upgrade which is much faster than cloning. First, you need to make sure, that the PGVERSION environment variable is set for the database pods. Since v1.6.0 the related option enable_pgversion_env_var is enabled by default.

In-place major version upgrades can be configured to be executed by the operator with the major_version_upgrade_mode option. By default it is set to off which means the cluster version will not change when increased in the manifest. Still, a rolling update would be triggered updating the PGVERSION variable. But Spilo's configure_spilo script will notice the version mismatch and start the old version again.

In this scenario the major version could then be run by a user from within the primary pod. Exec into the container and run:

python3 /scripts/inplace_upgrade.py N

where N is the number of members of your cluster (see numberOfInstances). The upgrade is usually fast, well under one minute for most DBs. Note, that changes become irrevertible once pg_upgrade is called. To understand the upgrade procedure, refer to the corresponding PR in Spilo.

When major_version_upgrade_mode is set to manual the operator will run the upgrade script for you after the manifest is updated and pods are rotated. It is also possible to define maintenanceWindows in the Postgres manifest to better control when such automated upgrades should take place after increasing the version.

Upgrade annotations

When an upgrade is executed, the operator sets an annotation in the PostgreSQL resource, either last-major-upgrade-success if the upgrade succeeds, or last-major-upgrade-failure if it fails. The value of the annotation is a timestamp indicating when the upgrade occurred.

If a PostgreSQL resource contains a failure annotation, the operator will not attempt to retry the upgrade during a sync event. To remove the failure annotation, you can revert the PostgreSQL version back to the current version. This action will trigger the removal of the failure annotation.

Non-default cluster domain

If your cluster uses a DNS domain other than the default cluster.local, this needs to be set in the operator configuration (cluster_domain variable). This is used by the operator to connect to the clusters after creation.

Namespaces

Select the namespace to deploy to

The operator can run in a namespace other than default. For example, to use the test namespace, run the following before deploying the operator's manifests:

kubectl create namespace test
kubectl config set-context $(kubectl config current-context) --namespace=test

All subsequent kubectl commands will work with the test namespace. The operator will run in this namespace and look up needed resources - such as its ConfigMap - there. Please note that the namespace for service accounts and cluster role bindings in operator RBAC rules needs to be adjusted to the non-default value.

Specify the namespace to watch

Watching a namespace for an operator means tracking requests to change Postgres clusters in the namespace such as "increase the number of Postgres replicas to 5" and reacting to the requests, in this example by actually scaling up.

By default, the operator watches the namespace it is deployed to. You can change this by setting the WATCHED_NAMESPACE var in the env section of the operator deployment manifest or by altering the watched_namespace field in the operator configuration. In the case both are set, the env var takes the precedence. To make the operator listen to all namespaces, explicitly set the field/env var to "*".

Note that for an operator to manage pods in the watched namespace, the operator's service account (as specified in the operator deployment manifest) has to have appropriate privileges to access the watched namespace. The operator may not be able to function in the case it watches all namespaces but lacks access rights to any of them (except K8s system namespaces like kube-system). The reason is that for multiple namespaces operations such as 'list pods' execute at the cluster scope and fail at the first violation of access rights.

Operators with defined ownership of certain Postgres clusters

By default, multiple operators can only run together in one K8s cluster when isolated into their own namespaces. But, it is also possible to define ownership between operator instances and Postgres clusters running all in the same namespace or K8s cluster without interfering.

First, define the CONTROLLER_ID environment variable in the operator deployment manifest. Then specify the ID in every Postgres cluster manifest you want this operator to watch using the "acid.zalan.do/controller" annotation:

apiVersion: "acid.zalan.do/v1"
kind: postgresql
metadata:
  name: demo-cluster
  annotations:
    "acid.zalan.do/controller": "second-operator"
spec:
  ...

Every other Postgres cluster which lacks the annotation will be ignored by this operator. Conversely, operators without a defined CONTROLLER_ID will ignore clusters with defined ownership of another operator.

Understanding rolling update of Spilo pods

The operator logs reasons for a rolling update with the info level and a diff between the old and new StatefulSet specs with the debug level. To benefit from numerous escape characters in the latter log entry, view it in CLI with echo -e. Note that the resultant message will contain some noise because the PodTemplate used by the operator is yet to be updated with the default values used internally in K8s.

The StatefulSet is replaced if the following properties change:

  • annotations
  • volumeClaimTemplates
  • template volumes

The StatefulSet is replaced and a rolling updates is triggered if the following properties differ between the old and new state:

  • container name, ports, image, resources, env, envFrom, securityContext and volumeMounts
  • template labels, annotations, service account, securityContext, affinity, priority class and termination grace period

Note that, changes in SPILO_CONFIGURATION env variable under bootstrap.dcs path are ignored for the diff. They will be applied through Patroni's rest api interface, following a restart of all instances.

The operator also support lazy updates of the Spilo image. In this case the StatefulSet is only updated, but no rolling update follows. This feature saves you a switchover - and hence downtime - when you know pods are re-started later anyway, for instance due to the node rotation. To force a rolling update, disable this mode by setting the enable_lazy_spilo_upgrade to false in the operator configuration and restart the operator pod.

Delete protection via annotations

To avoid accidental deletes of Postgres clusters the operator can check the manifest for two existing annotations containing the cluster name and/or the current date (in YYYY-MM-DD format). The name of the annotation keys can be defined in the configuration. By default, they are not set which disables the delete protection. Thus, one could choose to only go with one annotation.

postgres-operator ConfigMap

apiVersion: v1
kind: ConfigMap
metadata:
  name: postgres-operator
data:
  delete_annotation_date_key: "delete-date"
  delete_annotation_name_key: "delete-clustername"

OperatorConfiguration

apiVersion: "acid.zalan.do/v1"
kind: OperatorConfiguration
metadata:
  name: postgresql-operator-configuration
configuration:
  kubernetes:
    delete_annotation_date_key: "delete-date"
    delete_annotation_name_key: "delete-clustername"

Now, every cluster manifest must contain the configured annotation keys to trigger the delete process when running kubectl delete pg. Note, that the Postgresql resource would still get deleted because the operator does not instruct K8s' API server to block it. Only the operator logs will tell, that the delete criteria was not met.

cluster manifest

apiVersion: "acid.zalan.do/v1"
kind: postgresql
metadata:
  name: demo-cluster
  annotations:
    delete-date: "2020-08-31"
    delete-clustername: "demo-cluster"
spec:
  ...

In case, the resource has been deleted accidentally or the annotations were simply forgotten, it's safe to recreate the cluster with kubectl create. Existing Postgres cluster are not replaced by the operator. But, when the original cluster still exists the status will be CreateFailed at first. On the next sync event it should change to Running. However, because it is in fact a new resource for K8s, the UID and therefore, the backup path to S3, will differ and trigger a rolling update of the pods.

Owner References and Finalizers

The Postgres Operator can set owner references to most of a cluster's child resources to improve monitoring with GitOps tools and enable cascading deletes. There are two exceptions:

The operator would clean these resources up with its regular delete loop unless they got synced correctly. If for some reason the initial cluster sync fails, e.g. after a cluster creation or operator restart, a deletion of the cluster manifest might leave orphaned resources behind which the user has to clean up manually.

Another option is to enable finalizers which first ensures the deletion of all child resources before the cluster manifest gets removed. There is a trade-off though: The deletion is only performed after the next two operator SYNC cycles with the first one setting a deletionTimestamp and the latter reacting to it. The final removal of the custom resource will add a DELETE event to the worker queue but the child resources are already gone at this point. If you do not desire this behavior consider enabling owner references instead.

postgres-operator ConfigMap

apiVersion: v1
kind: ConfigMap
metadata:
  name: postgres-operator
data:
  enable_finalizers: "false"
  enable_owner_references: "true"

OperatorConfiguration

apiVersion: "acid.zalan.do/v1"
kind: OperatorConfiguration
metadata:
  name: postgresql-operator-configuration
configuration:
  kubernetes:
    enable_finalizers: false
    enable_owner_references: true

⚠️ Please note, both options are disabled by default. When enabling owner references the operator cannot block cascading deletes, even when the delete protection annotations are in place. You would need an K8s admission controller that blocks the actual kubectl delete API call e.g. based on existing annotations.

Role-based access control for the operator

The manifest operator-service-account-rbac.yaml defines the service account, cluster roles and bindings needed for the operator to function under access control restrictions. The file also includes a cluster role postgres-pod with privileges for Patroni to watch and manage pods and endpoints. To deploy the operator with this RBAC policies use:

kubectl create -f manifests/configmap.yaml
kubectl create -f manifests/operator-service-account-rbac.yaml
kubectl create -f manifests/postgres-operator.yaml
kubectl create -f manifests/minimal-postgres-manifest.yaml

Namespaced service account and role binding

For each namespace the operator watches it creates (or reads) a service account and role binding to be used by the Postgres Pods. The service account is bound to the postgres-pod cluster role. The name and definitions of these resources can be configured. Note, that the operator performs no further syncing of namespaced service accounts and role bindings.

Give K8s users access to create/list postgresqls

By default postgresql custom resources can only be listed and changed by cluster admins. To allow read and/or write access to other human users apply the user-facing-clusterrole manifest:

kubectl create -f manifests/user-facing-clusterroles.yaml

It creates zalando-postgres-operator:user:view, :edit and :admin clusterroles that are aggregated into the K8s default roles.

For Helm deployments setting rbac.createAggregateClusterRoles: true adds these clusterroles to the deployment.

Password rotation in K8s secrets

The operator regularly updates credentials in the K8s secrets if the enable_password_rotation option is set to true in the configuration. It happens only for LOGIN roles with an associated secret (manifest roles, default users from preparedDatabases). Furthermore, there are the following exceptions:

  1. Infrastructure role secrets since rotation should happen by the infrastructure.
  2. Team API roles that connect via OAuth2 and JWT token (no secrets to these roles anyway).
  3. Database owners since ownership on database objects can not be inherited.
  4. System users such as postgres, standby and pooler user.

The interval of days can be set with password_rotation_interval (default 90 = 90 days, minimum 1). On each rotation the user name and password values are replaced in the K8s secret. They belong to a newly created user named after the original role plus rotation date in YYMMDD format. All priviliges are inherited meaning that migration scripts should still grant and revoke rights against the original role. The timestamp of the next rotation (in RFC 3339 format, UTC timezone) is written to the secret as well. Note, if the rotation interval is decreased it is reflected in the secrets only if the next rotation date is more days away than the new length of the interval.

Pods still using the previous secret values which they keep in memory continue to connect to the database since the password of the corresponding user is not replaced. However, a retention policy can be configured for users created by the password rotation feature with password_rotation_user_retention. The operator will ensure that this period is at least twice as long as the configured rotation interval, hence the default of 180 = 180 days. When the creation date of a rotated user is older than the retention period it might not get removed immediately. Only on the next user rotation it is checked if users can get removed. Therefore, you might want to configure the retention to be a multiple of the rotation interval.

Password rotation for single users

From the configuration, password rotation is enabled for all secrets with the mentioned exceptions. If you wish to first test rotation for a single user (or just have it enabled only for a few secrets) you can specify it in the cluster manifest. The rotation and retention intervals can only be configured globally.

spec:
  usersWithSecretRotation:
  - foo_user
  - bar_reader_user

Password replacement without extra users

For some use cases where the secret is only used rarely - think of a flyway user running a migration script on pod start - we do not need to create extra database users but can replace only the password in the K8s secret. This type of rotation cannot be configured globally but specified in the cluster manifest:

spec:
  usersWithInPlaceSecretRotation:
  - flyway
  - bar_owner_user

This would be the recommended option to enable rotation in secrets of database owners, but only if they are not used as application users for regular read and write operations.

Ignore rotation for certain users

If you wish to globally enable password rotation but need certain users to opt out from it there are two ways. First, you can remove the user from the manifest's users section. The corresponding secret to this user will no longer be synced by the operator then.

Secondly, if you want the operator to continue syncing the secret (e.g. to recreate if it got accidentally removed) but cannot allow it being rotated, add the user to the following list in your manifest:

spec:
  usersIgnoringSecretRotation:
  - bar_user

Turning off password rotation

When password rotation is turned off again the operator will check if the username value in the secret matches the original username and replace it with the latter. A new password is assigned and the nextRotation field is cleared. A final lookup for child (rotation) users to be removed is done but they will only be dropped if the retention policy allows for it. This is to avoid sudden connection issues in pods which still use credentials of these users in memory. You have to remove these child users manually or re-enable password rotation with smaller interval so they get cleaned up.

Use taints and tolerations for dedicated PostgreSQL nodes

To ensure Postgres pods are running on nodes without any other application pods, you can use taints and tolerations and configure the required toleration in the operator configuration.

As an example you can set following node taint:

kubectl taint nodes <nodeName> postgres=:NoSchedule

And configure the toleration for the Postgres pods by adding following line to the ConfigMap:

apiVersion: v1
kind: ConfigMap
metadata:
  name: postgres-operator
data:
  toleration: "key:postgres,operator:Exists,effect:NoSchedule"

For an OperatorConfiguration resource the toleration should be defined like this:

apiVersion: "acid.zalan.do/v1"
kind: OperatorConfiguration
metadata:
  name: postgresql-configuration
configuration:
  kubernetes:
    toleration:
      postgres: "key:postgres,operator:Exists,effect:NoSchedule"

Note that the K8s version 1.13 brings taint-based eviction to the beta stage and enables it by default. Postgres pods by default receive tolerations for unreachable and noExecute taints with the timeout of 5m. Depending on your setup, you may want to adjust these parameters to prevent master pods from being evicted by the K8s runtime. To prevent eviction completely, specify the toleration by leaving out the tolerationSeconds value (similar to how Kubernetes' own DaemonSets are configured)

Node readiness labels

The operator can watch on certain node labels to detect e.g. the start of a Kubernetes cluster upgrade procedure and move master pods off the nodes to be decommissioned. Key-value pairs for these node readiness labels can be specified in the configuration (option name is in singular form):

apiVersion: v1
kind: ConfigMap
metadata:
  name: postgres-operator
data:
  node_readiness_label: "status1:ready,status2:ready"
apiVersion: "acid.zalan.do/v1"
kind: OperatorConfiguration
metadata:
  name: postgresql-configuration
configuration:
  kubernetes:
    node_readiness_label:
      status1: ready
      status2: ready

The operator will create a nodeAffinity on the pods. This makes the node_readiness_label option the global configuration for defining node affinities for all Postgres clusters. You can have both, cluster-specific and global affinity, defined and they will get merged on the pods. If node_readiness_label_merge is configured to "AND" the node readiness affinity will end up under the same matchExpressions section(s) from the manifest affinity.

  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: environment
            operator: In
            values:
            - pci
          - key: status1
            operator: In
            values:
            - ready
          - key: status2
            ...

If node_readiness_label_merge is set to "OR" (default) the readiness label affinty will be appended with its own expressions block:

  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: environment
            ...
        - matchExpressions:
          - key: storage
            ...
        - matchExpressions:
          - key: status1
            ...
          - key: status2
            ...

Enable pod anti affinity

To ensure Postgres pods are running on different topologies, you can use pod anti affinity and configure the required topology in the operator configuration.

Enable pod anti affinity by adding following line to the operator ConfigMap:

apiVersion: v1
kind: ConfigMap
metadata:
  name: postgres-operator
data:
  enable_pod_antiaffinity: "true"

Likewise, when using an OperatorConfiguration resource add:

apiVersion: "acid.zalan.do/v1"
kind: OperatorConfiguration
metadata:
  name: postgresql-configuration
configuration:
  kubernetes:
    enable_pod_antiaffinity: true

By default the type of pod anti affinity is requiredDuringSchedulingIgnoredDuringExecution, you can switch to preferredDuringSchedulingIgnoredDuringExecution by setting pod_antiaffinity_preferred_during_scheduling: true.

By default the topology key for the pod anti affinity is set to kubernetes.io/hostname, you can set another topology key e.g. failure-domain.beta.kubernetes.io/zone. See built-in node labels for available topology keys.

Pod Disruption Budget

By default the operator uses a PodDisruptionBudget (PDB) to protect the cluster from voluntarily disruptions and hence unwanted DB downtime. The MinAvailable parameter of the PDB is set to 1 which prevents killing masters in single-node clusters and/or the last remaining running instance in a multi-node cluster.

The PDB is only relaxed in two scenarios:

  • If a cluster is scaled down to 0 instances (e.g. for draining nodes)
  • If the PDB is disabled in the configuration (enable_pod_disruption_budget)

The PDB is still in place having MinAvailable set to 0. If enabled it will be automatically set to 1 on scale up. Disabling PDBs helps avoiding blocking Kubernetes upgrades in managed K8s environments at the cost of prolonged DB downtime. See PR #384 for the use case.

Add cluster-specific labels

In some cases, you might want to add labels that are specific to a given Postgres cluster, in order to identify its child objects. The typical use case is to add labels that identifies the Pods created by the operator, in order to implement fine-controlled NetworkPolicies.

postgres-operator ConfigMap

apiVersion: v1
kind: ConfigMap
metadata:
  name: postgres-operator
data:
  inherited_labels: application,environment

OperatorConfiguration

apiVersion: "acid.zalan.do/v1"
kind: OperatorConfiguration
metadata:
  name: postgresql-operator-configuration
configuration:
  kubernetes:
    inherited_labels:
    - application
    - environment

cluster manifest

apiVersion: "acid.zalan.do/v1"
kind: postgresql
metadata:
  name: demo-cluster
  labels:
    application: my-app
    environment: demo
spec:
  ...

network policy

kind: NetworkPolicy
apiVersion: networking.k8s.io/v1
metadata:
  name: netpol-example
spec:
  podSelector:
    matchLabels:
      application: my-app
      environment: demo

Custom Pod Environment Variables

The operator will assign a set of environment variables to the database pods that cannot be overridden to guarantee core functionality. Only variables with 'WAL_' and 'LOG_' prefixes can be customized to allow for backup and log shipping to be specified differently. There are three ways to specify extra environment variables (or override existing ones) for database pods:

The first two options must be referenced from the operator configuration making them global settings for all Postgres cluster the operator watches. One use case is a customized Spilo image that must be configured by extra environment variables. Another case could be to provide custom cloud provider or backup settings.

The last options allows for specifying environment variables individual to every cluster via the env section in the manifest. For example, if you use individual backup locations for each of your clusters. Or you want to disable WAL archiving for a certain cluster by setting WAL_S3_BUCKET, WAL_GS_BUCKET or AZURE_STORAGE_ACCOUNT to an empty string.

The operator will give precedence to environment variables in the following order (e.g. a variable defined in 4. overrides a variable with the same name in 5.):

  1. Assigned by the operator
  2. env section in cluster manifest
  3. Clone section (with WAL settings from operator config when s3_wal_path is empty)
  4. Standby section
  5. Pod environment secret via operator config
  6. Pod environment config map via operator config
  7. WAL and logical backup settings from operator config

Via ConfigMap

The ConfigMap with the additional settings is referenced in the operator's main configuration. A namespace can be specified along with the name. If left out, the configured default namespace of your K8s client will be used and if the ConfigMap is not found there, the Postgres cluster's namespace is taken when different:

postgres-operator ConfigMap

apiVersion: v1
kind: ConfigMap
metadata:
  name: postgres-operator
data:
  # referencing config map with custom settings
  pod_environment_configmap: default/postgres-pod-config

OperatorConfiguration

apiVersion: "acid.zalan.do/v1"
kind: OperatorConfiguration
metadata:
  name: postgresql-operator-configuration
configuration:
  kubernetes:
    # referencing config map with custom settings
    pod_environment_configmap: default/postgres-pod-config

referenced ConfigMap postgres-pod-config

apiVersion: v1
kind: ConfigMap
metadata:
  name: postgres-pod-config
  namespace: default
data:
  MY_CUSTOM_VAR: value

The key-value pairs of the ConfigMap are then added as environment variables to the Postgres StatefulSet/pods.

Via Secret

The Secret with the additional variables is referenced in the operator's main configuration. To protect the values of the secret from being exposed in the pod spec they are each referenced as SecretKeyRef. This does not allow for the secret to be in a different namespace as the pods though

postgres-operator ConfigMap

apiVersion: v1
kind: ConfigMap
metadata:
  name: postgres-operator
data:
  # referencing secret with custom environment variables
  pod_environment_secret: postgres-pod-secrets

OperatorConfiguration

apiVersion: "acid.zalan.do/v1"
kind: OperatorConfiguration
metadata:
  name: postgresql-operator-configuration
configuration:
  kubernetes:
    # referencing secret with custom environment variables
    pod_environment_secret: postgres-pod-secrets

referenced Secret postgres-pod-secrets

apiVersion: v1
kind: Secret
metadata:
  name: postgres-pod-secrets
  namespace: default
data:
  MY_CUSTOM_VAR: dmFsdWU=

The key-value pairs of the Secret are all accessible as environment variables to the Postgres StatefulSet/pods.

Via Postgres Cluster Manifest

It is possible to define environment variables directly in the Postgres cluster manifest to configure it individually. The variables must be listed under the env section in the same way you would do for containers. Global parameters served from a custom config map or secret will be overridden.

apiVersion: "acid.zalan.do/v1"
kind: postgresql
metadata:
  name: acid-test-cluster
spec:
  env:
  - name: wal_s3_bucket
    value: my-custom-bucket
  - name: minio_secret_key
      valueFrom:
        secretKeyRef:
          name: my-custom-secret
          key: minio_secret_key

Limiting the number of min and max instances in clusters

As a preventive measure, one can restrict the minimum and the maximum number of instances permitted by each Postgres cluster managed by the operator. If either min_instances or max_instances is set to a non-zero value, the operator may adjust the number of instances specified in the cluster manifest to match either the min or the max boundary. For instance, of a cluster manifest has 1 instance and the min_instances is set to 3, the cluster will be created with 3 instances. By default, both parameters are set to -1.

Load balancers and allowed IP ranges

For any Postgres/Spilo cluster, the operator creates two separate K8s services: one for the master pod and one for replica pods. To expose these services to an outer network, one can attach load balancers to them by setting enableMasterLoadBalancer and/or enableReplicaLoadBalancer to true in the cluster manifest. In the case any of these variables are omitted from the manifest, the operator configuration settings enable_master_load_balancer and enable_replica_load_balancer apply. Note that the operator settings affect all Postgresql services running in all namespaces watched by the operator. If load balancing is enabled two default annotations will be applied to its services:

  • external-dns.alpha.kubernetes.io/hostname with the value defined by the operator configs master_dns_name_format and replica_dns_name_format. This value can't be overwritten. If any changing in its value is needed, it MUST be done changing the DNS format operator config parameters; and
  • service.beta.kubernetes.io/aws-load-balancer-connection-idle-timeout with a default value of "3600".

There are multiple options to specify service annotations that will be merged with each other and override in the following order (where latter take precedence):

  1. Default annotations if LoadBalancer is enabled
  2. Globally configured custom_service_annotations
  3. serviceAnnotations specified in the cluster manifest
  4. masterServiceAnnotations and replicaServiceAnnotations specified in the cluster manifest

To limit the range of IP addresses that can reach a load balancer, specify the desired ranges in the allowedSourceRanges field (applies to both master and replica load balancers). To prevent exposing load balancers to the entire Internet, this field is set at cluster creation time to 127.0.0.1/32 unless overwritten explicitly. If you want to revoke all IP ranges from an existing cluster, please set the allowedSourceRanges field to 127.0.0.1/32 or to an empty sequence []. Setting the field to null or omitting it entirely may lead to K8s removing this field from the manifest due to its handling of null fields. Then the resultant manifest will not contain the necessary change, and the operator will respectively do nothing with the existing source ranges.

Load balancer services can also be enabled for the connection pooler pods with manifest flags enableMasterPoolerLoadBalancer and/or enableReplicaPoolerLoadBalancer or in the operator configuration with enable_master_pooler_load_balancer and/or enable_replica_pooler_load_balancer. For the external-dns.alpha.kubernetes.io/hostname annotation the -pooler suffix will be appended to the cluster name used in the template which is defined in master|replica_dns_name_format.

Running periodic 'autorepair' scans of K8s objects

The Postgres Operator periodically scans all K8s objects belonging to each cluster and repairs all discrepancies between them and the definitions generated from the current cluster manifest. There are two types of scans:

  • sync scan, running every resync_period seconds for every cluster

  • repair scan, coming every repair_period only for those clusters that didn't report success as a result of the last operation applied to them.

Postgres roles supported by the operator

The operator is capable of maintaining roles of multiple kinds within a Postgres database cluster:

  • System roles are roles necessary for the proper work of Postgres itself such as a replication role or the initial superuser role. The operator delegates creating such roles to Patroni and only establishes relevant secrets.

  • Infrastructure roles are roles for processes originating from external systems, e.g. monitoring robots. The operator creates such roles in all Postgres clusters it manages, assuming that K8s secrets with the relevant credentials exist beforehand.

  • Per-cluster robot users are also roles for processes originating from external systems but defined for an individual Postgres cluster in its manifest. A typical example is a role for connections from an application that uses the database.

  • Human users originate from the Teams API that returns a list of the team members given a team id. The operator differentiates between (a) product teams that own a particular Postgres cluster and are granted admin rights to maintain it, (b) Postgres superuser teams that get superuser access to all Postgres databases running in a K8s cluster for the purposes of maintaining and troubleshooting, and (c) additional teams, superuser teams or members associated with the owning team. The latter is managed via the PostgresTeam CRD.

Access to cloud resources from clusters in non-cloud environment

To access cloud resources like S3 from a cluster on bare metal you can use additional_secret_mount and additional_secret_mount_path configuration parameters. The cloud credentials will be provisioned in the Postgres containers by mounting an additional volume from the given secret to database pods. They can then be accessed over the configured mount path. Via Custom Pod Environment Variables you can point different cloud SDK's (AWS, GCP etc.) to this mounted secret, e.g. to access cloud resources for uploading logs etc.

A secret can be pre-provisioned in different ways:

  • Generic secret created via kubectl create secret generic some-cloud-creds --from-file=some-cloud-credentials-file.json
  • Automatically provisioned via a custom K8s controller like kube-aws-iam-controller

WAL archiving and physical basebackups

Spilo is shipped with WAL-E and its successor WAL-G to perform WAL archiving. By default, WAL-E is used for backups because it is more battle-tested. In addition to the continuous backup stream WAL-E/G pushes a physical base backup every night and 01:00 am UTC.

These are the pre-configured settings in the docker image:

BACKUP_NUM_TO_RETAIN: 5
BACKUP_SCHEDULE:      '00 01 * * *'
USE_WALG_BACKUP:      false (true for Azure and SSH)
USE_WALG_RESTORE:     false (true for S3, Azure and SSH)

Within Postgres you can check the pre-configured commands for archiving and restoring WAL files. You can find the log files to the respective commands under $HOME/pgdata/pgroot/pg_log/postgres-?.log.

archive_command:  `envdir "{WALE_ENV_DIR}" {WALE_BINARY} wal-push "%p"`
restore_command:  `envdir "{{WALE_ENV_DIR}}" /scripts/restore_command.sh "%f" "%p"`

You can produce a basebackup manually with the following command and check if it ends up in your specified WAL backup path:

envdir "/run/etc/wal-e.d/env" /scripts/postgres_backup.sh "/home/postgres/pgdata/pgroot/data"

You can also check if Spilo is able to find any backups:

envdir "/run/etc/wal-e.d/env" wal-g backup-list

Depending on the cloud storage provider different environment variables have to be set for Spilo. Not all of them are generated automatically by the operator by changing its configuration. In this case you have to use an extra configmap or secret.

Using AWS S3 or compliant services

When using AWS you have to reference the S3 backup path, the IAM role and the AWS region in the configuration.

postgres-operator ConfigMap

apiVersion: v1
kind: ConfigMap
metadata:
  name: postgres-operator
data:
  aws_region: eu-central-1
  kube_iam_role: postgres-pod-role
  wal_s3_bucket: your-backup-path

OperatorConfiguration

apiVersion: "acid.zalan.do/v1"
kind: OperatorConfiguration
metadata:
  name: postgresql-operator-configuration
configuration:
  aws_or_gcp:
    aws_region: eu-central-1
    kube_iam_role: postgres-pod-role
    wal_s3_bucket: your-backup-path

The referenced IAM role should contain the following privileges to make sure Postgres can send compressed WAL files to the given S3 bucket:

  PostgresPodRole:
    Type: "AWS::IAM::Role"
    Properties:
      RoleName: "postgres-pod-role"
      Path: "/"
      Policies:
        - PolicyName: "SpiloS3Access"
          PolicyDocument:
            Version: "2012-10-17"
            Statement:
              - Action: "s3:*"
                Effect: "Allow"
                Resource:
                  - "arn:aws:s3:::your-backup-path"
                  - "arn:aws:s3:::your-backup-path/*"

This should produce the following settings for the essential environment variables:

AWS_ENDPOINT='https://s3.eu-central-1.amazonaws.com:443'
WALE_S3_ENDPOINT='https+path://s3.eu-central-1.amazonaws.com:443'
WALE_S3_PREFIX=$WAL_S3_BUCKET/spilo/{WAL_BUCKET_SCOPE_PREFIX}{SCOPE}{WAL_BUCKET_SCOPE_SUFFIX}/wal/{PGVERSION}

The operator sets the prefix to an empty string so that spilo will generate it from the configured WAL_S3_BUCKET.

⚠️ When you overwrite the configuration by defining WAL_S3_BUCKET in the pod_environment_configmap you have to set WAL_BUCKET_SCOPE_PREFIX = "", too. Otherwise Spilo will not find the physical backups on restore (next chapter).

When the AWS_REGION is set, AWS_ENDPOINT and WALE_S3_ENDPOINT are generated automatically. WALG_S3_PREFIX is identical to WALE_S3_PREFIX. SCOPE is the Postgres cluster name.

⚠️ If both AWS_REGION and AWS_ENDPOINT or WALE_S3_ENDPOINT are defined backups with WAL-E will fail. You can fix it by switching to WAL-G with USE_WALG_BACKUP: "true".

Google Cloud Platform setup

When using GCP, there are two authentication methods to allow the postgres cluster to access buckets to write WAL-E logs: Workload Identity (recommended) or using a GCP Service Account Key (legacy).

Workload Identity setup

To configure the operator on GCP using Workload Identity these prerequisites are needed.

  • Workload Identity enabled on the GKE cluster where the operator will be deployed
  • A GCP service account with the proper IAM setup to access the GCS bucket for the WAL-E logs
  • An IAM policy granting the Kubernetes service account the roles/iam.workloadIdentityUser role on the GCP service account, e.g.:
gcloud iam service-accounts add-iam-policy-binding <GCP_SERVICE_ACCOUNT_NAME>@<GCP_PROJECT_ID>.iam.gserviceaccount.com \
    --role roles/iam.workloadIdentityUser \
    --member "serviceAccount:PROJECT_ID.svc.id.goog[<POSTGRES_OPERATOR_NS>/postgres-pod-custom]"

The configuration parameters that we will be using are:

  • wal_gs_bucket
  1. Create a custom Kubernetes service account to be used by Patroni running on the postgres cluster pods, this service account should include an annotation with the email address of the Google IAM service account used to communicate with the GCS bucket, e.g.
apiVersion: v1
kind: ServiceAccount
metadata:
  name: postgres-pod-custom
  namespace: <POSTGRES_OPERATOR_NS>
  annotations:
    iam.gke.io/gcp-service-account: <GCP_SERVICE_ACCOUNT_NAME>@<GCP_PROJECT_ID>.iam.gserviceaccount.com
  1. Specify the new custom service account in your operator paramaters

If using manual deployment or kustomize, this is done by setting pod_service_account_name in your configuration file specified in the postgres-operator deployment

If deploying the operator using Helm, this can be specified in the chart's values file, e.g.:

...
podServiceAccount:
  name: postgres-pod-custom
  1. Setup your operator configuration values. Ensure that the operator's configuration is set up like the following:
...
aws_or_gcp:
  # additional_secret_mount: ""
  # additional_secret_mount_path: ""
  # aws_region: eu-central-1
  # kube_iam_role: ""
  # log_s3_bucket: ""
  # wal_s3_bucket: ""
  wal_gs_bucket: "postgres-backups-bucket-28302F2"  # name of bucket on where to save the WAL-E logs
  # gcp_credentials: ""
...

Continue to shared steps below.

GCP Service Account Key setup

To configure the operator on GCP using a GCP service account key these prerequisites are needed.

  • A service account with the proper IAM setup to access the GCS bucket for the WAL-E logs
  • The credentials file for the service account.

The configuration parameters that we will be using are:

  • additional_secret_mount
  • additional_secret_mount_path
  • gcp_credentials
  • wal_gs_bucket
  1. Generate the K8s secret resource that will contain your service account's credentials. It's highly recommended to use a service account and limit its scope to just the WAL-E bucket.
apiVersion: v1
kind: Secret
metadata:
  name: psql-wale-creds
  namespace: default
type: Opaque
stringData:
  key.json: |-
    <GCP .json credentials>
  1. Setup your operator configuration values. With the psql-wale-creds resource applied to your cluster, ensure that the operator's configuration is set up like the following:
...
aws_or_gcp:
  additional_secret_mount: "psql-wale-creds"
  additional_secret_mount_path: "/var/secrets/google"  # or where ever you want to mount the file
  # aws_region: eu-central-1
  # kube_iam_role: ""
  # log_s3_bucket: ""
  # wal_s3_bucket: ""
  wal_gs_bucket: "postgres-backups-bucket-28302F2"  # name of bucket on where to save the WAL-E logs
  gcp_credentials: "/var/secrets/google/key.json"  # combination of the mount path & key in the K8s resource. (i.e. key.json)
...

Once you have set up authentication using one of the two methods above, continue with the remaining shared steps:

  1. Setup pod environment configmap that instructs the operator to use WAL-G, instead of WAL-E, for backup and restore.
apiVersion: v1
kind: ConfigMap
metadata:
  name: pod-env-overrides
  namespace: postgres-operator-system
data:
  # Any env variable used by spilo can be added
  USE_WALG_BACKUP: "true"
  USE_WALG_RESTORE: "true"
  CLONE_USE_WALG_RESTORE: "true"
  1. Then provide this configmap in postgres-operator settings:
...
# namespaced name of the ConfigMap with environment variables to populate on every pod
pod_environment_configmap: "postgres-operator-system/pod-env-overrides"
...

Azure setup

To configure the operator on Azure these prerequisites are needed:

  • A storage account in the same region as the Kubernetes cluster.

The configuration parameters that we will be using are:

  • pod_environment_secret
  • wal_az_storage_account
  1. Generate the K8s secret resource that will contain your storage account's access key. You will need a copy of this secret in every namespace you want to create postgresql clusters.

The latest version of WAL-G (v1.0) supports the use of a SASS token, but you'll have to make due with using the primary or secondary access token until the version of WAL-G is updated in the postgres-operator.

apiVersion: v1
kind: Secret
metadata:
  name: psql-backup-creds
  namespace: default
type: Opaque
stringData:
  AZURE_STORAGE_ACCESS_KEY: <primary or secondary access key>
  1. Setup pod environment configmap that instructs the operator to use WAL-G, instead of WAL-E, for backup and restore.
apiVersion: v1
kind: ConfigMap
metadata:
  name: pod-env-overrides
  namespace: postgres-operator-system
data:
  # Any env variable used by spilo can be added
  USE_WALG_BACKUP: "true"
  USE_WALG_RESTORE: "true"
  CLONE_USE_WALG_RESTORE: "true"
  WALG_AZ_PREFIX: "azure://container-name/$(SCOPE)/$(PGVERSION)" # Enables Azure Backups (SCOPE = Cluster name) (PGVERSION = Postgres version)
  1. Setup your operator configuration values. With the psql-backup-creds and pod-env-overrides resources applied to your cluster, ensure that the operator's configuration is set up like the following:
...
kubernetes:
  pod_environment_secret: "psql-backup-creds"
  pod_environment_configmap: "postgres-operator-system/pod-env-overrides"
aws_or_gcp:
  wal_az_storage_account: "postgresbackupsbucket28302F2"  # name of storage account to save the WAL-G logs
...

Restoring physical backups

If cluster members have to be (re)initialized restoring physical backups happens automatically either from the backup location or by running pg_basebackup on one of the other running instances (preferably replicas if they do not lag behind). You can test restoring backups by cloning clusters.

If you need to provide a custom clone environment copy existing variables about your setup (backup location, prefix, access keys etc.) and prepend the CLONE_ prefix to get them copied to the correct directory within Spilo.

apiVersion: v1
kind: ConfigMap
metadata:
  name: postgres-pod-config
data:
  AWS_REGION: "eu-west-1"
  AWS_ACCESS_KEY_ID: "****"
  AWS_SECRET_ACCESS_KEY: "****"
  ...
  CLONE_AWS_REGION: "eu-west-1"
  CLONE_AWS_ACCESS_KEY_ID: "****"
  CLONE_AWS_SECRET_ACCESS_KEY: "****"
  ...

Standby clusters

The setup for standby clusters is similar to cloning when they stream changes from a WAL archive (S3 or GCS). If you are using additional environment variables to access your backup location you have to copy those variables and prepend the STANDBY_ prefix for Spilo to find the backups and WAL files to stream.

Alternatively, standby clusters can also stream from a remote primary cluster. You have to specify the host address. Port is optional and defaults to 5432. Note, that only one of the options (s3_wal_path, gs_wal_path, standby_host) can be present under the standby top-level key.

Logical backups

The operator can manage K8s cron jobs to run logical backups (SQL dumps) of Postgres clusters. The cron job periodically spawns a batch job that runs a single pod. The backup script within this pod's container can connect to a DB for a logical backup. The operator updates cron jobs during Sync if the job schedule changes; the job name acts as the job identifier. These jobs are to be enabled for each individual Postgres cluster by updating the manifest:

apiVersion: "acid.zalan.do/v1"
kind: postgresql
metadata:
  name: demo-cluster
spec:
  enableLogicalBackup: true

There a few things to consider when using logical backups:

  1. Logical backups should not be seen as a proper alternative to basebackups and WAL archiving which are described above. At the moment, the operator cannot restore logical backups automatically and you do not get point-in-time recovery but only snapshots of your data. In its current state, see logical backups as a way to quickly create SQL dumps that you can easily restore in an empty test cluster.

  2. The example image implements the backup via pg_dumpall and upload of compressed and encrypted results to an S3 bucket. pg_dumpall requires a superuser access to a DB and runs on the replica when possible.

  3. Due to the limitation of K8s cron jobs it is highly advisable to set up additional monitoring for this feature; such monitoring is outside of the scope of operator responsibilities.

  4. The operator does not remove old backups.

  5. You may use your own image by overwriting the relevant field in the operator configuration. Any such image must ensure the logical backup is able to finish in presence of pod restarts and simultaneous invocations of the backup cron job.

  6. For that feature to work, your RBAC policy must enable operations on the cronjobs resource from the batch API group for the operator service account. See example RBAC

  7. Resources of the pod template in the cron job can be configured. When left empty default values of spilo pods will be used.

Sidecars for Postgres clusters

A list of sidecars is added to each cluster created by the operator. The default is empty.

kind: OperatorConfiguration
configuration:
  sidecars:
  - image: image:123
    name: global-sidecar
    ports:
    - containerPort: 80
      protocol: TCP
    volumeMounts:
    - mountPath: /custom-pgdata-mountpoint
      name: pgdata
  - ...

In addition to any environment variables you specify, the following environment variables are always passed to sidecars:

  • POD_NAME - field reference to metadata.name
  • POD_NAMESPACE - field reference to metadata.namespace
  • POSTGRES_USER - the superuser that can be used to connect to the database
  • POSTGRES_PASSWORD - the password for the superuser

Setting up the Postgres Operator UI

Since the v1.2 release the Postgres Operator is shipped with a browser-based configuration user interface (UI) that simplifies managing Postgres clusters with the operator.

Building the UI image

The UI runs with Node.js and comes with it's own Docker image. However, installing Node.js to build the operator UI is not required. It is handled via Docker containers when running:

make docker

Configure endpoints and options

The UI talks to the K8s API server as well as the Postgres Operator REST API. K8s API server URLs are loaded from the machine's kubeconfig environment by default. Alternatively, a list can also be passed when starting the Python application with the --cluster option.

The Operator API endpoint can be configured via the OPERATOR_API_URL environment variables in the deployment manifest. You can also expose the operator API through a service. Some displayed options can be disabled from UI using simple flags under the OPERATOR_UI_CONFIG field in the deployment.

The viewing and creation of clusters within the UI is limited to the namespace specified by the TARGET_NAMESPACE option. To allow the creation and viewing of clusters in all namespaces, set TARGET_NAMESPACE to *.

Deploy the UI on K8s

Now, apply all manifests from the ui/manifests folder to deploy the Postgres Operator UI on K8s. Replace the image tag in the deployment manifest if you want to test the image you've built with make docker. Make sure the pods for the operator and the UI are both running.

sed -e "s/\(image\:.*\:\).*$/\1$TAG/" manifests/deployment.yaml | kubectl apply -f manifests/
kubectl get all -l application=postgres-operator-ui

Local testing

For local testing you need to apply K8s proxying and operator pod port forwarding so that the UI can talk to the K8s and Postgres Operator REST API. The Ingress resource is not needed. You can use the provided run_local.sh script for this. Make sure that:

  • Python dependencies are installed on your machine
  • the K8s API server URL is set for kubectl commands, e.g. for minikube it would usually be https://192.168.99.100:8443.
  • the pod label selectors for port forwarding are correct

When testing with minikube you have to build the image in its docker environment (running make docker doesn't do it for you). From the ui directory execute:

# compile and build operator UI
make docker

# build in image in minikube docker env
eval $(minikube docker-env)
docker build -t ghcr.io/zalando/postgres-operator-ui:v1.13.0 .

# apply UI manifests next to a running Postgres Operator
kubectl apply -f manifests/

# install python dependencies to run UI locally
pip3 install -r requirements
./run_local.sh