DEV Community

Franck Pachot
Franck Pachot

Posted on

The cost of OKE for YugabyteDB - 1 Create the cluster

From a cost perspective, the Oracle Cloud could be an interesting platform for a distributed database like YugabyteDB. Because OKE, the managed kubernetes, works well and because the network cost between Availability Domains, within a region, is cheap. But saying it is one thing, let's try it. I have some cloud credits that will expire soon, let's use them to build and run a YugabyteDB database on OKE. This is a multi-part blog post. This first part is about the Kubernetes cluster creation

Compartment

In order to control the costs, I create a compartment for it:
Image description

Image description

Kubernetes

I'm using the "quick create" option

Image description

Image description

Image description

I'm creating a node pool with 3 nodes at 16 vCPU (8 OCPU)

Image description

and 1TB storage

Image description

With all other defaults

Image description

Image description

Image description

I didn't specify any custom tags but OKEclusterName: yugabytedb will be added automatically to the cluster, load balancers, block volumes, node pool, and nodes. With this and the compartment I'll be able to clearly identify the costs.

The Oracle Tags are CreatedOn: 2022-04-06T07:53:25.895Z for the cluster, CreatedOn: 2022-04-06T07:53:27.662Z for the cluster, and CreatedOn: 2022-04-06T07:53:27.662Z for the node pools.

It is important to understand that I have a worker in each Availability Zone:

Image description

This is important for High Availability, and also to check the cost of this configuration.

It is also important to verify that the block volumes are also in each Availability Domain:

Image description

Now, ensuring that I can run kubectlcommands from the cloud shell. The instructions are found in "Quick Start":

Image description

I'll continue mostly in command line from the cloud shell:

dev@cloudshell:~ (uk-london-1)$ 

 oci ce cluster create-kubeconfig --cluster-id ocid1.cluster.oc1.uk-london-1.aaaaaaaa6q2ybdrdvwcxd5ghjlexv3yw4vsm3nxuarzdq7wmcccu4lmj2e7a \
 --file $HOME/.kube/config --region uk-london-1 \
 --token-version 2.0.0  --kube-endpoint PUBLIC_ENDPOINT

Existing Kubeconfig file found at /home/dev/.kube/config and new config merged into it
Enter fullscreen mode Exit fullscreen mode

I can see my 3 nodes:

dev@cloudshell:~ (uk-london-1)$ 

 kubectl get nodes -o wide

NAME          STATUS   ROLES   AGE   VERSION   INTERNAL-IP   EXTERNAL-IP       OS-IMAGE                  KERNEL-VERSION                      CONTAINER-RUNTIME
10.0.10.103   Ready    node    11m   v1.22.5   10.0.10.103   141.147.103.245   Oracle Linux Server 7.9   5.4.17-2136.304.4.1.el7uek.x86_64   cri-o://1.22.3-1.ci.el7
10.0.10.152   Ready    node    10m   v1.22.5   10.0.10.152   141.147.113.126   Oracle Linux Server 7.9   5.4.17-2136.304.4.1.el7uek.x86_64   cri-o://1.22.3-1.ci.el7
10.0.10.51    Ready    node    11m   v1.22.5   10.0.10.51    130.162.191.79    Oracle Linux Server 7.9   5.4.17-2136.304.4.1.el7uek.x86_64   cri-o://1.22.3-1.ci.el7
Enter fullscreen mode Exit fullscreen mode

And the system pods:

dev@cloudshell:~ (uk-london-1)$

 kubectl get pods -A-o wide

NAMESPACE     NAME                                   READY   STATUS    RESTARTS      AGE   IP             NODE          NOMINATED NODE   READINESS GATES
kube-system   coredns-7bb8797d57-kw8v5               1/1     Running   0             15m   10.244.0.130   10.0.10.103   <none>           <none>
kube-system   coredns-7bb8797d57-smvpx               1/1     Running   0             19m   10.244.0.3     10.0.10.51    <none>           <none>
kube-system   coredns-7bb8797d57-zt86b               1/1     Running   0             15m   10.244.1.2     10.0.10.152   <none>           <none>
kube-system   csi-oci-node-84pnz                     1/1     Running   0             16m   10.0.10.51     10.0.10.51    <none>           <none>
kube-system   csi-oci-node-js6k7                     1/1     Running   1 (15m ago)   16m   10.0.10.152    10.0.10.152   <none>           <none>
kube-system   csi-oci-node-lqbg6                     1/1     Running   0             16m   10.0.10.103    10.0.10.103   <none>           <none>
kube-system   kube-dns-autoscaler-5cd75c9b4c-968vm   1/1     Running   0             19m   10.244.0.2     10.0.10.51    <none>           <none>
kube-system   kube-flannel-ds-2lbsf                  1/1     Running   1 (15m ago)   16m   10.0.10.103    10.0.10.103   <none>           <none>
kube-system   kube-flannel-ds-dtg8x                  1/1     Running   1 (15m ago)   16m   10.0.10.51     10.0.10.51    <none>           <none>
kube-system   kube-flannel-ds-rvvbc                  1/1     Running   1 (15m ago)   16m   10.0.10.152    10.0.10.152   <none>           <none>
kube-system   kube-proxy-h8qmf                       1/1     Running   0             16m   10.0.10.152    10.0.10.152   <none>           <none>
kube-system   kube-proxy-kq8gr                       1/1     Running   0             16m   10.0.10.51     10.0.10.51    <none>           <none>
kube-system   kube-proxy-wm9zr                       1/1     Running   0             16m   10.0.10.103    10.0.10.103   <none>           <none>
kube-system   proxymux-client-2p5qw                  1/1     Running   0             16m   10.0.10.51     10.0.10.51    <none>           <none>
kube-system   proxymux-client-46s88                  1/1     Running   0             16m   10.0.10.152    10.0.10.152   <none>           <none>
kube-system   proxymux-client-rlpmm                  1/1     Running   0             16m   10.0.10.103    10.0.10.103   <none>           <none>
Enter fullscreen mode Exit fullscreen mode

I have my Oracle Cloud Kubernetes cluster. In the next post, I'll install YugabyteDB there with a helm chart.

Discussion (0)