🦭 mariadb-operator

Run and operate MariaDB in a cloud native way. Declaratively manage your MariaDB using Kubernetes CRDs rather than imperative commands.

Provisioning highly configurable MariaDB servers.
Multiple HA modes supported: SemiSync Replication and Galera. Automatic primary failover.
Take and restore backups. Scheduled backups. Backup rotation
PVCs and Kubernetes volumes (i.e. NFS) backup storage
Bootstrap new instances from backups and volumes (i.e NFS)
Manage users, grants and logical databases
Configure connections for your applications
Orchestrate and schedule sql scripts
Prometheus metrics
Validation webhooks to provide CRD inmutability
Additional printer columns to report the current CRD status
CRDs designed according to the Kubernetes API conventions
GitOps friendly
Multi-arch distroless based image
Install it using kubectl, helm or OLM

Bare minimum installation

This installation flavour provides the minimum resources required to run mariadb-operator in your cluster.

helm repo add mariadb-operator https://mariadb-operator.github.io/mariadb-operator
helm install mariadb-operator mariadb-operator/mariadb-operator

Recommended installation

The recommended installation includes the following features to provide a better user experience and reliability:

Metrics: Leverage prometheus operator to scrape metrics from both the mariadb-operator and the provisioned MariaDB instances.
Webhook certificate renewal: Automatic webhook certificate issuance and renewal using cert-manager. By default, a static self-signed certificate is generated.

helm repo add mariadb-operator https://mariadb-operator.github.io/mariadb-operator
helm install mariadb-operator mariadb-operator/mariadb-operator \
  --set metrics.enabled=true --set webhook.certificate.certManager=true

Openshift

The Openshift installation is managed separately in the mariadb-operator-helm repository, which contains a helm based operator that allows you to install mariadb-operator via OLM.

Quickstart

Let's see mariadb-operator🦭 in action! First of all, install the following configuration manifests that will be referenced by the CRDs further:

kubectl apply -f examples/manifests/config

Next, you can proceed with the installation of a MariaDB instance:

kubectl apply -f examples/manifests/mariadb_v1alpha1_mariadb.yaml

kubectl get mariadbs
NAME      READY   STATUS    PRIMARY POD     AGE
mariadb   True    Running   mariadb-0       3m57s

kubectl get statefulsets
NAME      READY   AGE
mariadb   1/1     2m12s

kubectl get services
NAME         TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)             AGE
mariadb      ClusterIP   10.96.235.145   <none>        3306/TCP,9104/TCP   2m17s

Up and running 🚀, we can now create our first logical database and grant access to users:

kubectl apply -f examples/manifests/mariadb_v1alpha1_database.yaml
kubectl apply -f examples/manifests/mariadb_v1alpha1_user.yaml
kubectl apply -f examples/manifests/mariadb_v1alpha1_grant.yaml

kubectl get databases
NAME        READY   STATUS    CHARSET   COLLATE           AGE
data-test   True    Created   utf8      utf8_general_ci   22s

kubectl get users
NAME              READY   STATUS    MAXCONNS   AGE
mariadb-metrics   True    Created   3          19m
user              True    Created   20         29s

kubectl get grants
NAME              READY   STATUS    DATABASE   TABLE   USERNAME          GRANTOPT   AGE
mariadb-metrics   True    Created   *          *       mariadb-metrics   false      19m
user              True    Created   *          *       user              true       36s

At this point, we can run our database initialization scripts:

kubectl apply -f examples/manifests/sqljobs

kubectl get sqljobs
NAME       COMPLETE   STATUS    MARIADB   AGE
01-users   True       Success   mariadb   2m47s
02-repos   True       Success   mariadb   2m47s
03-stars   True       Success   mariadb   2m47s

kubectl get jobs
NAME                  COMPLETIONS   DURATION   AGE
01-users              1/1           10s        3m23s
02-repos              1/1           11s        3m13s
03-stars-28067562     1/1           10s        106s

kubectl get cronjobs
NAME       SCHEDULE      SUSPEND   ACTIVE   LAST SCHEDULE   AGE
03-stars   */1 * * * *   False     0        57s             2m33s

Now that the database has been initialized, let's take a backup:

kubectl apply -f examples/manifests/mariadb_v1alpha1_backup_scheduled.yaml

After one minute, the backup should have completed:

kubectl get backups
NAME               COMPLETE   STATUS    MARIADB   AGE
backup-scheduled   True       Success   mariadb   15m

kubectl get cronjobs
NAME               SCHEDULE      SUSPEND   ACTIVE   LAST SCHEDULE   AGE
backup-scheduled   */1 * * * *   False     0        56s             15m

kubectl get jobs
NAME                                    COMPLETIONS   DURATION   AGE
backup-scheduled-27782894               1/1           4s         3m2s

Last but not least, let's provision a second MariaDB instance bootstrapping from the previous backup:

kubectl apply -f examples/manifests/mariadb_v1alpha1_mariadb_from_backup.yaml

kubectl get mariadbs
NAME                  READY   STATUS    PRIMARY POD             AGE
mariadb               True    Running   mariadb-0               7m47s
mariadb-from-backup   True    Running   mariadb-from-backup-0   53s

kubectl get restores
NAME                                         COMPLETE   STATUS    MARIADB               AGE
bootstrap-restore-mariadb-from-backup        True       Success   mariadb-from-backup   72s

kubectl get jobs
NAME                                         COMPLETIONS   DURATION   AGE
backup                                       1/1           9s         12m
bootstrap-restore-mariadb-from-backup        1/1           5s         84s

You can take a look at the whole suite of example CRDs available in examples/manifests.

High availability

This operator supports the following High Availability modes:

Single master HA via SemiSync Replication: The primary node allows both reads and writes, while secondary nodes only allow reads.
Multi master HA via Galera: All nodes support reads and writes, but it is recommended to perform writes in a single primary for preventing deadlocks.

In order to address nodes, mariadb-operator provides you with the following Kubernetes Services:

<mariadb-name>: To be used for read requests. It will point to all nodes.
<mariadb-name>-primary: To be used for write requests. It will point to a single node, the primary.
<mariadb-name>-secondary: To be used for read requests. It will point to all nodes, except the primary.

Whenever the primary changes, either by the user or by the operator, both the <mariadb-name>-primary and <mariadb-name>-secondary Services will be automatically updated by the operator to address the right nodes.

The primary may be manually changed by the user at any point by updating the spec.[replication|galera].primary.podIndex field. Alternatively, automatic primary failover can be enabled by setting spec.[replication|galera].primary.automaticFailover, which will make the operator to switch primary whenever the primary Pod goes down.

GitOps

You can configure mariadb-operator's CRDs in your git repo and reconcile them using your favorite GitOps tool, see an example with flux:

Run and operate MariaDB in a GitOps fashion using Flux

Roadmap

Take a look at our roadmap and feel free to open an issue to suggest new features.

Contributing

We welcome and encourage contributions to this project! Please check our contributing and development guides. PRs welcome!

Blogs

Run and operate MariaDB in Kubernetes with mariadb-operator - MariaDB Foundation

MartinWeise/mariadb-operator