New to KubeDB? Please start here.
FerretDB
What is FerretDB
FerretDB
is a Kubernetes Custom Resource Definitions
(CRD). It provides declarative configuration for FerretDB in a Kubernetes native way. You only need to describe the desired configuration in a FerretDB
object, and the KubeDB operator will create Kubernetes objects in the desired state for you.
FerretDB Spec
As with all other Kubernetes objects, a FerretDB needs apiVersion
, kind
, and metadata
fields. It also needs a .spec
section. Below is an example FerretDB object.
apiVersion: kubedb.com/v1alpha2
kind: FerretDB
metadata:
name: ferretdb
namespace: demo
spec:
version: "1.23.0"
replicas: 1
healthChecker:
failureThreshold: 3
periodSeconds: 20
timeoutSeconds: 10
authSecret:
name: ferretdb-auth
externallyManaged: false
backend:
postgresRef:
name: ha-postgres
namespace: demo
version: "13.13"
linkedDB: "ferretdb"
externallyManaged: false
sslMode: requireSSL
tls:
issuerRef:
apiGroup: cert-manager.io
name: ferretdb-ca-issuer
kind: Issuer
certificates:
- alias: server
subject:
organizations:
- kubedb:server
dnsNames:
- localhost
ipAddresses:
- "127.0.0.1"
monitor:
agent: prometheus.io/operator
prometheus:
serviceMonitor:
labels:
release: prometheus
interval: 10s
deletionPolicy: WipeOut
podTemplate:
spec:
containers:
- name: ferretdb
resources:
limits:
memory: 1Gi
requests:
cpu: 200m
memory: 256Mi
serviceTemplates:
- alias: primary
spec:
type: ClusterIP
ports:
- name: http
port: 9999
spec.version
spec.version
is a required field specifying the name of the FerretDBVersion crd where the docker images are specified. Currently, when you install KubeDB, it creates the following FerretDBVersion
resources,
1.18.0
,1.23.0
spec.replicas
spec.replicas
the number of members in ferretdb replicaset.
KubeDB uses PodDisruptionBudget
to ensure that majority of these replicas are available during voluntary disruptions so that quorum is maintained.
spec.healthChecker
It defines the attributes for the health checker.
spec.healthChecker.periodSeconds
specifies how often to perform the health check.spec.healthChecker.timeoutSeconds
specifies the number of seconds after which the probe times out.spec.healthChecker.failureThreshold
specifies minimum consecutive failures for the healthChecker to be considered failed.spec.healthChecker.disableWriteCheck
specifies whether to disable the writeCheck or not.
spec.authSecret
spec.authSecret
is an optional field that points to a Secret used to hold credentials for ferretdb
. If not set, KubeDB operator creates a new Secret {ferretdb-object-name}-auth
for storing the password for ferretdb
user for each FerretDB object.
As FerretDB use backend’s authentication mechanisms till now, this secret is basically a copy of backend postgres.
We can use this field in 3 mode.
- Using an external secret. In this case, You need to create an auth secret first with required fields, then specify the secret name when creating the FerretDB object using
spec.authSecret.name
& setspec.authSecret.externallyManaged
to true.
authSecret:
name: <your-created-auth-secret-name>
externallyManaged: true
- Specifying the secret name only. In this case, You need to specify the secret name when creating the FerretDB object using
spec.authSecret.name
.externallyManaged
is by default false.
authSecret:
name: <intended-auth-secret-name>
- Let KubeDB do everything for you. In this case, no work for you.
AuthSecret contains a user
key and a password
key which contains the username
and password
respectively for ferretdb
user.
Example:
$ kubectl create secret generic ferretdb-auth -n demo \
--from-literal=username=jhon \
--from-literal=password=O9xE1mZZDAdBTbrV
secret "ferretdb-auth" created
apiVersion: v1
data:
password: "O9xE1mZZDAdBTbrV"
username: "jhon"
kind: Secret
metadata:
name: ferretdb-auth
namespace: demo
type: Opaque
Secrets provided by users are not managed by KubeDB, and therefore, won’t be modified or garbage collected by the KubeDB operator (version 0.13.0 and higher).
spec.backend
spec.backend.externallyManaged
represents how the backend will be managed. If its false, KubeDB will automatically create a KubeDB managed postgres. Otherwise you need refer a AppBinding name and namespace which represents information for external Postgres.spec.backend.postgresRef
is a required field that points to theappbinding
associated with the backend postgres. If the postgres is KubeDB managed an AppBinding will be created automatically upon creating the postgres. If the postgres is not KubeDB managed then you need to create an appbinding yourself.spec.backend.postgresRef
takes the name (spec.backend.postgresRef.Name
) of the appbinding and the namespace (spec.backend.postgresRef.Namespace
) where the appbinding is created.spec.backend.version
represents the version of backend postgresspec.backend.linkedDB
represents in which database of backend postgres will be used by FerretDB to store data
spec.sslMode
Enables TLS/SSL or mixed TLS/SSL used for all network connections. The value of sslMode
field can be one of the following:
Value | Description |
---|---|
disabled | The server does not use TLS/SSL. |
requireSSL | The server uses and accepts only TLS/SSL encrypted connections. |
spec.tls
spec.tls
specifies the TLS/SSL configurations for the FerretDB. KubeDB uses cert-manager v1 api to provision and manage TLS certificates.
The following fields are configurable in the spec.tls
section:
issuerRef
is a reference to theIssuer
orClusterIssuer
CR of cert-manager that will be used byKubeDB
to generate necessary certificates.apiGroup
is the group name of the resource that is being referenced. Currently, the only supported value iscert-manager.io
.kind
is the type of resource that is being referenced. KubeDB supports bothIssuer
andClusterIssuer
as values for this field.name
is the name of the resource (Issuer
orClusterIssuer
) being referenced.
certificates
(optional) are a list of certificates used to configure the server and/or client certificate. It has the following fields:alias
represents the identifier of the certificate. It has the following possible value:server
is used for server certificate identification.client
is used for client certificate identification.metrics-exporter
is used for metrics exporter certificate identification.
secretName
(optional) specifies the k8s secret name that holds the certificates.This field is optional. If the user does not specify this field, the default secret name will be created in the following format:
<database-name>-<cert-alias>-cert
.subject
(optional) specifies anX.509
distinguished name. It has the following possible field,organizations
(optional) are the list of different organization names to be used on the Certificate.organizationalUnits
(optional) are the list of different organization unit name to be used on the Certificate.countries
(optional) are the list of country names to be used on the Certificate.localities
(optional) are the list of locality names to be used on the Certificate.provinces
(optional) are the list of province names to be used on the Certificate.streetAddresses
(optional) are the list of a street address to be used on the Certificate.postalCodes
(optional) are the list of postal code to be used on the Certificate.serialNumber
(optional) is a serial number to be used on the Certificate. You can find more details from Here
duration
(optional) is the period during which the certificate is valid.renewBefore
(optional) is a specifiable time before expiration duration.dnsNames
(optional) is a list of subject alt names to be used in the Certificate.ipAddresses
(optional) is a list of IP addresses to be used in the Certificate.uris
(optional) is a list of URI Subject Alternative Names to be set in the Certificate.emailAddresses
(optional) is a list of email Subject Alternative Names to be set in the Certificate.privateKey
(optional) specifies options to control private keys used for the Certificate.encoding
(optional) is the private key cryptography standards (PKCS) encoding for this certificate’s private key to be encoded in. If provided, allowed values are “pkcs1” and “pkcs8” standing for PKCS#1 and PKCS#8, respectively. It defaults to PKCS#1 if not specified.
spec.monitor
FerretDB managed by KubeDB can be monitored with builtin-Prometheus and Prometheus operator out-of-the-box. To learn more,
spec.deletionPolicy
deletionPolicy
gives flexibility whether to nullify
(reject) the delete operation of FerretDB
CR or which resources KubeDB should keep or delete when you delete FerretDB
CR. KubeDB provides following four deletion policies:
- DoNotTerminate
- Delete
- WipeOut (
Default
)
When deletionPolicy
is DoNotTerminate
, KubeDB takes advantage of ValidationWebhook
feature in Kubernetes 1.9.0 or later clusters to implement DoNotTerminate
feature. If admission webhook is enabled, DoNotTerminate
prevents users from deleting the database as long as the spec.deletionPolicy
is set to DoNotTerminate
.
Following table show what KubeDB does when you delete FerretDB CR for different deletion policies,
Behavior | DoNotTerminate | Delete | WipeOut |
---|---|---|---|
1. Block Delete operation | ✓ | ✗ | ✗ |
2. Delete PetSet | ✗ | ✓ | ✓ |
3. Delete Services | ✗ | ✓ | ✓ |
4. Delete Secrets | ✗ | ✗ | ✓ |
If you don’t specify spec.deletionPolicy
KubeDB uses Delete
deletion policy by default.
spec.podTemplate
KubeDB allows providing a template for pod through spec.podTemplate
. KubeDB operator will pass the information provided in spec.podTemplate
to the PetSet created for FerretDB.
KubeDB accept following fields to set in spec.podTemplate:
- metadata:
- annotations (pod’s annotation)
- labels (pod’s labels)
- controller:
- annotations (statefulset’s annotation)
- labels (statefulset’s labels)
- spec:
- volumes
- initContainers
- containers
- imagePullSecrets
- nodeSelector
- affinity
- serviceAccountName
- schedulerName
- tolerations
- priorityClassName
- priority
- securityContext
- livenessProbe
- readinessProbe
- lifecycle
You can check out the full list here. Uses of some field of spec.podTemplate
is described below,
spec.podTemplate.spec.nodeSelector
spec.podTemplate.spec.nodeSelector
is an optional field that specifies a map of key-value pairs. For the pod to be eligible to run on a node, the node must have each of the indicated key-value pairs as labels (it can have additional labels as well). To learn more, see here .
spec.podTemplate.spec.resources
spec.podTemplate.spec.resources
is an optional field. This can be used to request compute resources required by the database pods. To learn more, visit here.
spec.serviceTemplates
You can also provide template for the services created by KubeDB operator for Kafka cluster through spec.serviceTemplates
. This will allow you to set the type and other properties of the services.
KubeDB allows following fields to set in spec.serviceTemplates
:
alias
represents the identifier of the service. It has the following possible value:stats
is used for the exporter service identification.
- metadata:
- labels
- annotations
- spec:
- type
- ports
- clusterIP
- externalIPs
- loadBalancerIP
- loadBalancerSourceRanges
- externalTrafficPolicy
- healthCheckNodePort
- sessionAffinityConfig
See here to understand these fields in detail.