想通过dep.yaml文件转储data.sql文件，不想通过bin/bash手动

Question

1. I Want to dump data file using YAML file

data.sql file in the volume

the YAML file is

apiVersion: v1
kind: Service
metadata:
  name: mysql
spec:
  ports:
  - port: 3306
  selector:
    app: mysql
  clusterIP: None
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: mysql
  labels:
    app: mysql
spec:
  replicas: 1
  selector:
    matchLabels:
      app: mysql
  strategy:
    type: Recreate
  template:
    metadata:
      labels:
        app: mysql
    spec:
        volumes:
        - name: mydir
          hostPath: 
            path: /var/local/test
            type: DirectoryOrCreate
        - name: myfile
          hostPath:
            path: /var/local/test/data.sql
            type: FileOrCreate 

        containers:
        - image: mysql:5.6
          name: mysql
          env:
            # Use secret in real usage
        
          - name: MYSQL_ROOT_PASSWORD
            value: root
          - name: MYSQL_USERNAME
            value: root
          - name: MYSQL_PASSWORD
            value: root
          - name: MYSQL_DATABASE
            value: sample  
          ports:
          - containerPort: 3306
            name: mysql
          volumeMounts:
          - name: mydir
            mountPath: /var/local/test/data.sql
          command: ["/bin/sh","-c"]
          args: ["mysql -u root -proot sample < data.sql"] 

using a command, args concept but not possible

If I access the pod it can be dump easily but I want dump action through YAML

**2. Restore the database file in local system **

Also want to restore the dumped aur new data in my system or any other location but it will be accessible

Don't share the bash option kubectl exec -it mysql-598cbfc789-2jlt7 -- mysql -u root -proot sample < data.sql

Answer 1

Mostly you can't do this in the way you describe. Kubernetes YAML isn't great for describing imperative actions like "restore a database dump". If you need to do it there, you need to do it in a Job separate from the database proper.

The most portable way to do this is to not attempt it in Kubernetes YAML at all (so remove, for example, the override command: and args: in your Deployment). Run kubectl apply -f ... as normal, and once the database comes up, run in two separate terminal windows

kubectl port-forward service/mysql 12345:3306
mysql -u root -proot -H127.0.0.1 -P12345 sample < data.sql

The first establishes a connection to the running database; the second uses the local mysql tool and your local filesystem, connecting to the forwarded port. 12345 can be any port number but needs to be the same in both commands.

Some other alternatives (focusing on kubectl exec ) are described in How do I restore a dump file from mysqldump using kubernetes? .

There are a couple of conceptual issues with the layout you show. The first is that a container runs only one command, and your sh -c 'mysql ...' runs instead of the MySQL daemon. A Kubernetes Pod also can't access your local filesystem, and if you use a hostPath volume like you show, you need to correctly guess which node the pod will run on and copy the data there; this isn't reliable and can lead to inconsistent results if you run it from different places.

---

If you can get the dump data into Kubernetes somehow, you can run a separate Job to do the restore:

apiVersion: batch/v1
kind: Job
metadata:
  name: mysql-restore
spec:
  template:
    spec:
      containers:
        - name: mysql-restore
          env: { same as in: the question }
          image: mysql:5.6
          volumeMounts: { ... }
          command: [sh, -c]
          args:
            - >-
                mysql sample < /var/local/test/data.sql

The question is what to put in the volumeMounts: . If the database dump is fairly small, you might be able to fit it into a ConfigMap and then mount the ConfigMap into the Pod , but creating the ConfigMap can be awkward and the size is limited to 1 MB. If you have a way to directly access the volume content (maybe you're mounting an NFS mount you can also access out-of-cluster) then you can copy the dump there. You also might be able to use network-accessible cloud storage like Amazon S3 or another HTTP-based service where you can curl the dump at startup time.