Adapting to Kubernetes

When using the MindIO ACP acceleration service in a container, you need to install the SDK in the container.

1. Modify the YAML file for creating a pod. The following uses the /home/testuser/mygpt.yaml file as an example to describe how to add a mapping volume.
   1. Open the mygpt.yaml file.

      vim /home/testuser/mygpt.yaml

   2. Press i to enter the insert mode and modify the mygpt.yaml file.
      

      • If volumeMounts and volumes do not exist, add all content to the file.
      • If volumeMounts and volumes exist, add the content following them.
      • (Optional) If DPC is used to access the storage in the environment, add the mapping path of the volume in the container. The content is as follows:

        volumeMounts:
            - mountPath: /opt/oceanstor/dataturbo/sdk/lib/libdpc_nds.so
              name: mindio-dpc-nds
              readOnly: false

        

        Do not change /opt/oceanstor/dataturbo/sdk/lib/libdpc_nds.so.

      • (Optional) If DPC is used to access storage in the environment, add the declaration of the volume to be mapped to the host as follows:

        volumes:
          - name: mindio-dpc-nds
            hostPath:
              path: /opt/oceanstor/dataturbo/sdk/lib/libdpc_nds.so
              type: File

   3. Press Esc, type :wq!, and press Enter to save the changes and exit.
2. Use the modified YAML file to create a pod.

   kubectl apply -f mygpt.yaml

3. Log in to the created pod. The following uses the pod named mygptdd in the namespace test-mindio as an example.

   kubectl exec -it mygptdd -n test-mindio /bin/bash

4. Upload the MindIO ACP SDK to the pod and install the SDK by referring to Installing the MindIO ACP SDK on Compute Nodes.