How to deploy a PostgreSQL Cluster on Kubernetes + OpenEBS

Why Postgres on Kubernetes?

Well, the answer is in the question. If you are already running Kubernetes on some form of cloud, you understand the ease-of-use, scalability and monitoring benefits of Kubernetes that you can apply to your database at scale.

PostgreSQL is the preferred relational database for most developers around, although setting up a highly available Postgres cluster from scratch is always a challenge, being cloud-native adds a bit to the difficulty.

There are many ways to run high availability with PostgreSQL; for a list, see the PostgreSQL Documentation. To be honest, manually setting it up is quite painful, while there are better ways available. My favorite cloud-native Postgres cluster deployment projects are Crunchy Data‘s, Sorint.lab‘s Stolon and Zalando‘s Patroni/Spilo.

Since availability requires multi-node Kubernetes deployment instead of local Minikube setup, I’ll deploy crunchy-postgres on my existing K8s cluster on AWS with two worker nodes. If you don’t have a Kubernetes cluster yet, see the instructions to deploy one using StackPointCloud. Instructions after that are same in any cloud or on-premises deployment.

Prerequisites

Software
Cloud Provider

Deploy Crunchy PostgreSQL cluster using kubectl

Once you have OpenEBS storage classes created on your K8s cluster, you can use the following simple steps to launch a highly available PostgreSQL service with one master and one replica.

Download the files to your host, which has access to kubectl

cd $HOME
git clone https://github.com/openebs/openebs.git
cd openebs/k8s/demo/crunchy-postgres

Create the Stateful Set

The deployment will use the default images and credentials defined in the set.json file. To set custom users and passwords:

vi  ~/openebs/k8s/demo/crunchy-postgres/set.json

JSON file should look like below, feel free to edit the number of replicas, credentials and storage capacity. Default uses the openebs-standard storage class, and it is 400M.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
{
"apiVersion": "apps/v1beta1",
"kind": "StatefulSet",
"metadata": {
"name": "pgset"
},
"spec": {
"serviceName": "pgset",
"replicas": 2,
"template": {
"metadata": {
"labels": {
"app": "pgset"
}
},
"spec": {
"containers": [
{
"name": "pgset",
"image": "crunchydata/crunchy-postgres:centos7-9.6-1.4.0",
"ports": [
{
"containerPort": 5432,
"name": "postgres"
}
],
"env": [{
"name": "PG_MASTER_USER",
"value": "master"
}, {
"name": "PGHOST",
"value": "/tmp"
}, {
"name": "PG_MODE",
"value": "master"
}, {
"name": "PG_MASTER_PASSWORD",
"value": "password"
}, {
"name": "PG_USER",
"value": "testuser"
}, {
"name": "PG_PASSWORD",
"value": "password"
}, {
"name": "PG_DATABASE",
"value": "userdb"
}, {
"name": "PG_ROOT_PASSWORD",
"value": "password"
}],
"volumeMounts": [
{
"name": "pgdata",
"mountPath": "/pgdata",
"readOnly": false
}
]
}
]
}
},
"volumeClaimTemplates": [
{
"metadata": {
"name": "pgdata"
},
"spec": {
"accessModes": [
"ReadWriteOnce"
],
"storageClassName": "openebs-standard",
"resources": {
"requests": {
"storage": "400M"
}
}
}
}
]
}
}

Save the file and run the statefulset:

./run.sh

The above step will automatically create the OpenEBS volumes required for master and replica postgresql containers and few other Kubernetes objects:

  • Persistent Volumes (pvc-{UID1}, pvc-{UID2})
  • Persistent Volume Claim (pgdata-pgset-0, pgdata-pgset-1)
  • Replica Sets (pvc-{UID1}-ctrl-{random1},pvc-{UID1}-rep-{random2},pvc-{UID2}-ctrl-{random3},pvc-{UID4}-ctrl-{random4})
  • Service Account (pgset-sa)
  • Services (pgset, pgset-master, pgset-replica)
  • StatefulSet (pgset)
  • Pods (pgset-0, pgset-1)

The volume details can be inspected using the standard kubectl commands. To check persistent volume claims:

kubectl get pvc

Check persistent volumes:

kubectl get pv

List the services, and you will see pgset, master and replica created:

kubectl get service

List the statefulsets, and you will see pgset listed with two desired and current sets:

If you use the Kubernetes Dashboard, you can see the same under Workloads > Stateful Sets and quickly scale up as well.

Test your Database

If it is not installed previously, install psql client:

sudo apt-get install postgresql-client

Test the master as follows (default password is “password”, unless you changed it):

psql -h pgset-master -U testuser password -c 'table pg_stat_replication'

Above command should return output indicating that a single replica is connecting to the master.

Now, test the replica as follows:

psql -h pgset-replica -U testuser password -c 'create table foo (id int)'

This command should fail as the replica is read-only within a PostgreSQL cluster.

Please share:

Also published on Medium.

Leave a Reply

Your email address will not be published. Required fields are marked *