Chapter 14. Changing resources for the OpenShift Data Foundation components

When you install OpenShift Data Foundation, it comes with pre-defined resources that the OpenShift Data Foundation pods can consume. In some situations with higher I/O load, it might be required to increase these limits.

14.1. Changing the CPU and memory resources on the rook-ceph pods

When you install OpenShift Data Foundation, it comes with pre-defined CPU and memory resources for the rook-ceph pods. You can manually increase these values according to the requirements.

You can change the CPU and memory resources on the following pods:

  • mgr
  • mds
  • rgw

The following example illustrates how to change the CPU and memory resources on the rook-ceph pods. In this example, the existing MDS pod values of cpu and memory are increased from 1 and 4Gi to 2 and 8Gi respectively.

  1. Edit the storage cluster: 

    # oc edit storagecluster -n openshift-storage <storagecluster_name>
    <storagecluster_name>

    Specify the name of the storage cluster.

    For example: 

    # oc edit storagecluster -n openshift-storage ocs-storagecluster

  2. Add the following lines to the storage cluster Custom Resource (CR): 

    spec:
  resources:
    mds:
      limits:
        cpu: 2
        memory: 8Gi
      requests:
        cpu: 2
        memory: 8Gi
  3. Save the changes and exit the editor.

  4. Alternatively, run the oc patch command to change the CPU and memory value of the mds pod: 

    # oc patch -n openshift-storage storagecluster <storagecluster_name>
    --type merge \
    --patch '{"spec": {"resources": {"mds": {"limits": {"cpu": "2","memory": "8Gi"},"requests": {"cpu": "2","memory": "8Gi"}}}}}'
    <storagecluster_name>

    Specify the name of the storage cluster.

    For example: 

    # oc patch -n openshift-storage storagecluster ocs-storagecluster \
    --type merge \
    --patch '{"spec": {"resources": {"mds": {"limits": {"cpu": "2","memory": "8Gi"},"requests": {"cpu": "2","memory": "8Gi"}}}}}'

14.2. Tuning the resources for the MCG

The default configuration for the Multicloud Object Gateway (MCG) is optimized for low resource consumption and not performance. For more information on how to tune the resources for the MCG, see the Red Hat Knowledgebase solution Performance tuning guide for Multicloud Object Gateway (NooBaa).