Deploy on Kubernetes

In this guide, we will deploy SurrealDB to KIND (Kubernetes in Docker) using TiKV as the storage engine: TiKV is a cloud-native transactional key/value store that integrates well with Kubernetes thanks to their tidb-operator.

At the end, we will run a few experiments using SurrealQL to verify that we can interact with the new cluster and will destroy some Kubernetes pods to see that data is highly available.

Requirements

For this guide, we need to install:

• kubectl : To manage the Kubernetes cluster
• helm : To install SurrealDB server and TiKV
• KIND and Docker :To run a local Kubernetes cluster inside a Docker container
• Surreal CLI : To interact with the SurrealDB server

Create KIND Cluster

First, we need to create a KIND cluster. KIND is a tool for running local Kubernetes clusters using Docker container “nodes”. It’s a great tool for experimenting with Kubernetes without spending a lot of time creating a full-featured cluster.

Run the following command to create a cluster:

1. Let’s create a new cluster:

Crete new cluster
kind create cluster -n surreal-demo

2. Verify we can interact with the created cluster

Verify cluster
kubectl config current-context

The output of this command should be:

Output
kind-surreal-demo

3. Verify the nodes are running

Verify nodes
kubectl get ns

The output of this command should be:

Output
NAME                 STATUS   AGE
default              Active   79s
kube-node-lease      Active   79s
kube-public          Active   79s
kube-system          Active   79s
local-path-storage   Active   75s

Deploy TiDB operator

Now that we have a Kubernetes cluster, we can deploy the TiDB operator. TiDB operator is a Kubernetes operator that manages the lifecycle of TiDB clusters deployed to Kubernetes.

You can deploy it following these steps:

1. Install CRDS:

Install
kubectl create -f https://raw.githubusercontent.com/pingcap/tidb-operator/v1.4.5/manifests/crd.yaml

2. Install TiDB Operator Helm chart:

Update HELM repositories
helm repo add pingcap https://charts.pingcap.org
helm repo update
helm install \
	-n tidb-operator \
  --create-namespace \
	tidb-operator \
	pingcap/tidb-operator \
	--version v1.4.5

3. Verify that the Pods are running:

Verify Pods
kubectl get pods --namespace tidb-operator -l app.kubernetes.io/instance=tidb-operator

The output of this command should look like this:

Output
NAME                                       READY   STATUS    RESTARTS   AGE
tidb-controller-manager-56f49794d7-hnfz7   1/1     Running   0          20s
tidb-scheduler-8655bcbc86-66h2d            2/2     Running   0          20s

Create TiDB cluster

Now that we have the TiDB Operator running, it’s time to define a TiDB Cluster and let the Operator do the rest. One of the TiDB Cluster components is the TiKV, which we are interested in. Given this is a demo, we will use a basic example cluster, but there are several examples in the official GitHub repo in case you need a more production-grade deployment

Run the following commands to deploy the TiKV cluster:

1. Create a namespace for the TiDB cluster:

kubectl create ns tikv

2. Create the TiDB cluster:

Create TIDB Cluster
kubectl apply -n tikv -f https://raw.githubusercontent.com/pingcap/tidb-operator/v1.4.5/examples/basic/tidb-cluster.yaml

3. Check the cluster status and wait until it’s ready:

Verify TIDB Cluster
kubectl get -n tikv tidbcluster

The output of this command should look like this:

Output
NAME    READY   PD                  STORAGE   READY   DESIRE   TIKV   STORAGE   READY   DESIRE   TIDB   READY   DESIRE   AGE
basic   False   pingcap/pd:v6.5.0   1Gi       1       1               1Gi               1                       1        41s
$ kubectl get -n tikv tidbcluster
NAME    READY   PD                  STORAGE   READY   DESIRE   TIKV                  STORAGE   READY   DESIRE   TIDB                  READY   DESIRE   AGE
basic   True    pingcap/pd:v6.5.0   1Gi       1       1        pingcap/tikv:v6.5.0   1Gi       1       1        pingcap/tidb:v6.5.0   1       1        5m

Deploy SurrealDB

Now that we have a TiDB cluster running, we can deploy SurrealDB. For this guide, we will use the SurrealDB Helm chart. Run the following commands to deploy SurrealDB:

1. Add the SurrealDB Charts repository:

Add SurrealDB Helm repository
helm repo add surrealdb https://helm.surrealdb.com
helm repo update

2. Get the TIKV PD service url:

Get TIKV PD service URL
kubectl get -n tikv svc/basic-pd

the output of this command should look like this:

Output
NAME       TYPE        CLUSTER-IP     EXTERNAL-IP   PORT(S)    AGE
basic-pd   ClusterIP   10.96.208.25   <none>        2379/TCP   10h

then set the TIKV_URL variable to the PD service URL:

Set TIKV_URL variable
export TIKV_URL=tikv://basic-pd.tikv:2379

3. Install the SurrealDB Helm chart with the TIKV_URL defined above and with auth disabled so we can create the initial credentials:

Install SurrealDB HELM chart
helm install --set surrealdb.path=$TIKV_URL --set surrealdb.unauthenticated=true --set image.tag=latest surrealdb-tikv surrealdb/surrealdb

4. Connect to the cluster and define the initial credentials (see in the section below how to connect):

surreal sql -e http://...
> DEFINE USER root ON ROOT PASSWORD 'StrongSecretPassword!' ROLES OWNER;

Verify you can connect to the database with the new credentials:
surreal sql -u root -p 'StrongSecretPassword!' -e http://...
> INFO FOR ROOT
[{ namespaces: { }, users: { root: "DEFINE USER root ON ROOT PASSHASH '...' ROLES OWNER" } }]

5. Now that the initial credentials have been created, enable authentication:

Update SurrealDB Helm Chart
helm upgrade --set surrealdb.path=$TIKV_URL --set image.tag=latest surrealdb-tikv surrealdb/surrealdb

Run SurrealDB experiments

Now that we have SurrealDB running, we can run some experiments to verify that everything is working as expected. For this guide, we will use the Surreal CLI. Run the following commands to run some experiments:

1. Start port-forwarding to the SurrealDB service:

kubectl port-forward svc/surrealdb-tikv 8000

Output
Forwarding from 127.0.0.1:8000 -> 8000
Forwarding from [::1]:8000 -> 8000

2. Connect to the SurrealDB server using the CLI from another shell:

surreal sql --conn 'http://localhost:8000' --user root --pass surrealdb

3. Create a SurrealDB database:

surreal sql --conn 'http://localhost:8000' --user root --pass surrealdb
> USE NS ns DB db;

ns/db> CREATE record;
{ id: record:wbd69kmc81l4fbee7pit }
ns/db> CREATE record;
{ id: record:vnyfsr22ovhmmtcm5y1t }
ns/db> CREATE record;
{ id: record:se49petzb7c4bc7yge0z }
ns/db> SELECT * FROM record;
[{ id: record:se49petzb7c4bc7yge0z }, { id: record:vnyfsr22ovhmmtcm5y1t }, { id: record:wbd69kmc81l4fbee7pit }]
ns/db>

The data created above has been persisted to the TiKV cluster. Let’s verify it by deleting the SurrealDB server and let Kubernetes recreate it.

Get pod
kubectl get pod

The output of this command should look like this:

Output
NAME                              READY   STATUS    RESTARTS   AGE
surrealdb-tikv-7488f6f654-lsrwp   1/1     Running   0          13m

Delete pod
kubectl delete pod surrealdb-tikv-7488f6f654-lsrwp

Get pod
kubectl get pod

Output
NAME                              READY   STATUS    RESTARTS   AGE
surrealdb-tikv-7488f6f654-bnkjz   1/1     Running   0          4s

Connect again and verify the data is still there (you may need to re-run the port-forwarding command):

surreal sql --conn 'http://localhost:8000' --user root --pass surrealdb
> USE NS ns DB db;

ns/db> SELECT * FROM record;
[{ id: record:se49petzb7c4bc7yge0z }, { id: record:vnyfsr22ovhmmtcm5y1t }, { id: record:wbd69kmc81l4fbee7pit }]
ns/db>

Note

Given we are using KIND, we use port-forwarding for demonstration purposes. In a full-featured Kubernetes cluster, you could set ingress.enabled=true when installing the SurrealDB Helm Chart and it would create a Load Balancer in front of the SurrealDB server pods.

Conclusion

This guide demonstrated how to deploy SurrealDB on Kubernetes using TiKV as a datastore. From here, you could try and deploy to EKS , GKE or AKS , and play with the different configurations for the TiKV cluster.