Updating the operating system (OS) on a host, by either upgrading across major releases or updating the system software for a minor release, can impact the OKD software running on those machines. In particular, these updates can affect the iptables rules or ovs flows that OKD requires to operate.

Updating the operating system on a host

To safely upgrade the OS on a host:

  1. Drain the node in preparation for maintenance:

    $ oc adm drain <node_name> --force --delete-local-data --ignore-daemonsets
  2. In order to protect sensitive packages that do not need to be updated, apply the exclude rules to the host:

    # atomic-openshift-docker-excluder exclude
    # atomic-openshift-excluder exclude

    A reboot ensures that the host is running the newest versions and means that the container engine and OKD processes have been restarted, which forces them to check that all of the rules in other services are correct.

    # yum update
    # reboot

    However, instead of rebooting a node host, you can restart the services that are affected or preserve the iptables state. Both processes are described in the OKD iptables topic. The ovs flow rules do not need to be saved, but restarting the OKD node software fixes the flow rules.

  3. Configure the host to be schedulable again:

    $ oc adm uncordon <node_name>

Upgrading Nodes Running OpenShift Container Storage

If using OpenShift Container Storage, upgrade the OKD nodes running OpenShift Container Storage one at a time.

  1. To begin, recall the project in which OpenShift Container Storage was deployed.

  2. Confirm the node and pod selectors configured on the service’s daemonset.

    $ oc get daemonset -n <project_name> -o wide

    Use -o wide to include the pod selector in the output.

    These selectors are found under NODE-SELECTOR and SELECTOR, respectively. The example commands below will use glusterfs=storage-host and glusterfs=storage-pod, respectively.

  3. Given the daemonset’s node selector, confirm which hosts have the label, and hence are running pods from the daemonset:

    $ oc get nodes --selector=glusterfs=storage-host

    Chose a node which will have its operating system upgraded.

  4. Remove the daemonset label from the node:

    $ oc label node <node_name> glusterfs-

    This will cause the OpenShift Container Storage pod to terminate on that node.

The node can now have its OS upgraded as described above.

  1. To restart an OpenShift Container Storage pod on the node, relabel the node with the daemonset label:

    $ oc label node <node_name> glusterfs=storage-host
  2. Wait for the OpenShift Container Storage pod to respawn and appear.

  3. Given the daemonset’s pod selector, determine the name of the newly spawned pod by searching for a pod running on the node whose OS you upgraded:

    $ oc get pod -n <project_name> --selector=glusterfs=storage-pod -o wide

    Use -o wide to include which host the pod is running on in the output.

  4. oc rsh into the gluster pod to check the volume heal:

    $ oc rsh <pod_name>
    $ for vol in `gluster volume list`; do gluster volume heal $vol info; done
    $ exit

    Ensure all of the volumes are healed and there are no outstanding tasks. The heal info command lists all pending entries for a given volume’s heal process. A volume is considered healed when Number of entries for that volume is 0. Use gluster volume status <volume_name> for additional details about the volume. The Online state should be marked Y for all bricks.