I'm trying to deploy Postgres on my Kubernetes cluster and I have been successful to do this, but then I don't know how I can import my data which are in csv format. I already have the scripts which gets the path to data and create a database in a local instance of postgres, but when I deploy postgres on Kubernetes cluster then those scripts wont work because I can't see those script inside the pod.
I was looking for a solution to execute the scripts from host to inside the pod, or I can expose the directory of scripts and data to postgres pod. I've found the hostpath solution, but I don't know how to define multiple volumes for a deployment. (I'm using Rook cluster to provision the volume) Maybe a way to define a hostpath volume alongside a Rook volume so I can have access to the scripts and csv files inside the hostpath and then create the database inside the Rook volume. I don't know of this makes sense, but I would appreciate if someone help me with this.
If you're using the official docker image, or an image that is derived from it but didn't destroy its entrypoint, then they have documentation about /docker-entrypoint-initdb.d/*.sql, with the tl;dr as
kind: ConfigMap
spec:
import_csv.sql: |
COPY my_table FROM '/whatever/path/you/want.csv' FORMAT csv /* etc */
---
kind: Pod
spec:
containers:
- volumeMounts:
- name: my-initdb-configmap
mountPath: /docker-entrypoint-initdb.d
readOnly: true
# ...type deal