Thread: Access a Postgres storage with two independent instances

Access a Postgres storage with two independent instances

From
first last
Date:
tried to deploy Postgres deployment with Kubernetes, having three replicas that are accessing the same storage(PVC). Here is the configuration
```
apiVersion: v1
kind: PersistentVolume
metadata:
name: postgres-volume
labels:
type: local
spec:
persistentVolumeReclaimPolicy: Delete
capacity:
storage: 10Gi
accessModes:
- ReadWriteOnce
hostPath:
path: /data/postgresql
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: postgres-volume-claim
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 10Gi
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: postgres
spec:
replicas: 3
selector:
matchLabels:
app: postgres
template:
metadata:
labels:
app: postgres
spec:
containers:
- name: postgres
image: 'postgres:14'
env:
- name: POSTGRES_DB
value: db
- name: POSTGRES_USER
value: user
- name: POSTGRES_PASSWORD
value: pwd
volumeMounts:
- name: postgresdata
mountPath: /var/lib/postgresql/data
volumes:
- name: postgresdata
persistentVolumeClaim:
claimName: postgres-volume-claim
---
apiVersion: v1
kind: Service
metadata:
name: postgres
labels:
app: postgres
spec:
type: NodePort
ports:
- port: 5432
protocol: TCP
targetPort: 5432
nodePort: 32001
selector:
app: postgres

```

When you apply this configuration file `kubectl apply -f file_name.yaml`, you can create three pods. Here I have given the configuration to create the pods that use a single storage as a data store. But what actually happening is, that it uses separate storage. I have tested all the replicas manually by creating a db and table and see, if other database stores are affected or not. Anyway, it is not affected. And then I tried to apply the same config file change the image name and the env variables suit for MySQL and tried to deploy it. It works as expected. Anyway, it can not use single storage(https://stackoverflow.com/questions/78436945/access-a-mysql-storage-with-two-independent-instances).

Finally, my question is, why the scenario is different for MySQL and Postgres?  Could you please help me to figure out the issue? This is for my Final year project testing Kubernetes using fuzzing.

Re: Access a Postgres storage with two independent instances

From
Alan Hodgson
Date:
On Tue, 2024-05-07 at 00:47 +0530, first last wrote:
tried to deploy Postgres deployment with Kubernetes, having three replicas that are accessing the same storage(PVC). Here is the configuration

Finally, my question is, why the scenario is different for MySQL and Postgres?  Could you please help me to figure out the issue? This is for my Final year project testing Kubernetes using fuzzing.


Not sure what you're trying to accomplish, but PostgreSQL can definitely not share storage between active postmasters. Nor can MySQL, afaik.

Re: Access a Postgres storage with two independent instances

From
first last
Date:
Actually what I want is multiple Postgres instances to have read and write access to the same data storage directory. Anyway, that can not be possible as what you say. But, when I deploy MySQL/MongoDB deployment with multiple replicas to access the same storage, the all pods except one become CrashLoopBackOff state. This is happening for MySQL/Mongo. When it comes to Postgres, I deployed it with multiple pods, all the pods became healthy and there is no such CrashLoopBackOff state occurred. But those pods access different storages for every pod. Why this totally different from MySQL and Mongo?

On Tue, May 7, 2024 at 2:34 AM Alan Hodgson <ahodgson@lists.simkin.ca> wrote:
On Tue, 2024-05-07 at 00:47 +0530, first last wrote:
tried to deploy Postgres deployment with Kubernetes, having three replicas that are accessing the same storage(PVC). Here is the configuration

Finally, my question is, why the scenario is different for MySQL and Postgres?  Could you please help me to figure out the issue? This is for my Final year project testing Kubernetes using fuzzing.


Not sure what you're trying to accomplish, but PostgreSQL can definitely not share storage between active postmasters. Nor can MySQL, afaik.

On Tue, May 7, 2024 at 2:34 AM Alan Hodgson <ahodgson@lists.simkin.ca> wrote:
On Tue, 2024-05-07 at 00:47 +0530, first last wrote:
tried to deploy Postgres deployment with Kubernetes, having three replicas that are accessing the same storage(PVC). Here is the configuration

Finally, my question is, why the scenario is different for MySQL and Postgres?  Could you please help me to figure out the issue? This is for my Final year project testing Kubernetes using fuzzing.


Not sure what you're trying to accomplish, but PostgreSQL can definitely not share storage between active postmasters. Nor can MySQL, afaik.

Re: Access a Postgres storage with two independent instances

From
Laurenz Albe
Date:
On Tue, 2024-05-07 at 09:56 +0530, first last wrote:
> But, when I deploy MySQL/MongoDB deployment with multiple replicas to
> access the same storage, the all pods except one become CrashLoopBackOff
> state. This is happening for MySQL/Mongo. When it comes to Postgres, I
> deployed it with multiple pods, all the pods became healthy and there is
> no such CrashLoopBackOff state occurred. But those pods access different
> storages for every pod. Why this totally different from MySQL and Mongo?

Because it is different software?

"Why" questions are notoriously hard to answer (as anyone knows who has
ever had children).  Shared-storage architectures are not great, because
they have a single point of failure and only scale very moderately.  So
there are no efforts in PostgreSQL to make that work.

Yours,
Laurenz Albe



Re: Access a Postgres storage with two independent instances

From
Alan Hodgson
Date:
On Tue, 2024-05-07 at 09:56 +0530, first last wrote:
Actually what I want is multiple Postgres instances to have read and write access to the same data storage directory. Anyway, that can not be possible as what you say. But, when I deploy MySQL/MongoDB deployment with multiple replicas to access the same storage, the all pods except one become CrashLoopBackOff state. This is happening for MySQL/Mongo. When it comes to Postgres, I deployed it with multiple pods, all the pods became healthy and there is no such CrashLoopBackOff state occurred. But those pods access different storages for every pod. Why this totally different from MySQL and Mongo?

No idea, but this doesn't appear to have anything to do with PostgreSQL. You could load any image in that container and then see what the storage looks like. Probably better asked on a kubernetes list somewhere.

Random thought, though, if you actually have multiple nodes in your k8s cluster, you might just be seeing an artifact of the ReadWriteOnce access mode. For a real shared volume you probably want ReadWriteMany, which only a few storage backends support (and which, again, PostgreSQL would not work on).