Configuring additional devices in an IBM Z or IBM® LinuxONE environment

After installing OKD, you can configure additional devices for your cluster in an IBM Z or IBM® LinuxONE environment, which is installed with z/VM. The following devices can be configured:

  • Fibre Channel Protocol (FCP) host

  • FCP LUN

  • DASD

  • qeth

You can configure devices by adding udev rules using the Machine Config Operator (MCO) or you can configure devices manually.

The procedures described here apply only to z/VM installations. If you have installed your cluster with Fedora KVM on IBM Z or IBM® LinuxONE infrastructure, no additional configuration is needed inside the KVM guest after the devices were added to the KVM guests. However, both in z/VM and Fedora KVM environments the next steps to configure the Local Storage Operator and Kubernetes NMState Operator need to be applied.

Additional resources

Configuring additional devices using the Machine Config Operator (MCO)

Tasks in this section describe how to use features of the Machine Config Operator (MCO) to configure additional devices in an IBM Z or IBM® LinuxONE environment. Configuring devices with the MCO is persistent but only allows specific configurations for compute nodes. MCO does not allow control plane nodes to have different configurations.

Prerequisites

  • You are logged in to the cluster as a user with administrative privileges.

  • The device must be available to the z/VM guest.

  • The device is already attached.

  • The device is not included in the cio_ignore list, which can be set in the kernel parameters.

  • You have created a MachineConfig object file with the following YAML:

    1. apiVersion: machineconfiguration.openshift.io/v1
    2. kind: MachineConfigPool
    3. metadata:
    4.   name: worker0
    5. spec:
    6.   machineConfigSelector:
    7.     matchExpressions:
    8.       - {key: machineconfiguration.openshift.io/role, operator: In, values: [worker,worker0]}
    9.   nodeSelector:
    10.     matchLabels:
    11.       node-role.kubernetes.io/worker0: ""

Configuring a Fibre Channel Protocol (FCP) host

The following is an example of how to configure an FCP host adapter with N_Port Identifier Virtualization (NPIV) by adding a udev rule.

Procedure

  1. Take the following sample udev rule 441-zfcp-host-0.0.8000.rules:

    1. ACTION=="add", SUBSYSTEM=="ccw", KERNEL=="0.0.8000", DRIVER=="zfcp", GOTO="cfg_zfcp_host_0.0.8000"
    2. ACTION=="add", SUBSYSTEM=="drivers", KERNEL=="zfcp", TEST=="[ccw/0.0.8000]", GOTO="cfg_zfcp_host_0.0.8000"
    3. GOTO="end_zfcp_host_0.0.8000"
    5. LABEL="cfg_zfcp_host_0.0.8000"
    6. ATTR{[ccw/0.0.8000]online}="1"
    8. LABEL="end_zfcp_host_0.0.8000"

  2. Convert the rule to Base64 encoded by running the following command:

    1. $ base64 /path/to/file/

  3. Copy the following MCO sample profile into a YAML file:

    1. apiVersion: machineconfiguration.openshift.io/v1
    2. kind: MachineConfig
    3. metadata:
    4.    labels:
    5.      machineconfiguration.openshift.io/role: worker0 (1)
    6.    name: 99-worker0-devices
    7. spec:
    8.    config:
    9.      ignition:
    10.        version: 3.2.0
    11.      storage:
    12.        files:
    13.        - contents:
    14.            source: data:text/plain;base64,<encoded_base64_string> (2)
    15.          filesystem: root
    16.          mode: 420
    17.          path: /etc/udev/rules.d/41-zfcp-host-0.0.8000.rules (3)
    1The role you have defined in the machine config file.
    2The Base64 encoded string that you have generated in the previous step.
    3The path where the udev rule is located.

Configuring an FCP LUN

The following is an example of how to configure an FCP LUN by adding a udev rule. You can add new FCP LUNs or add additional paths to LUNs that are already configured with multipathing.

Procedure

  1. Take the following sample udev rule 41-zfcp-lun-0.0.8000:0x500507680d760026:0x00bc000000000000.rules:

    1. ACTION=="add", SUBSYSTEMS=="ccw", KERNELS=="0.0.8000", GOTO="start_zfcp_lun_0.0.8207"
    2. GOTO="end_zfcp_lun_0.0.8000"
    4. LABEL="start_zfcp_lun_0.0.8000"
    5. SUBSYSTEM=="fc_remote_ports", ATTR{port_name}=="0x500507680d760026", GOTO="cfg_fc_0.0.8000_0x500507680d760026"
    6. GOTO="end_zfcp_lun_0.0.8000"
    8. LABEL="cfg_fc_0.0.8000_0x500507680d760026"
    9. ATTR{[ccw/0.0.8000]0x500507680d760026/unit_add}="0x00bc000000000000"
    10. GOTO="end_zfcp_lun_0.0.8000"
    12. LABEL="end_zfcp_lun_0.0.8000"

  2. Convert the rule to Base64 encoded by running the following command:

    1. $ base64 /path/to/file/

  3. Copy the following MCO sample profile into a YAML file:

    1. apiVersion: machineconfiguration.openshift.io/v1
    2. kind: MachineConfig
    3. metadata:
    4.    labels:
    5.      machineconfiguration.openshift.io/role: worker0 (1)
    6.    name: 99-worker0-devices
    7. spec:
    8.    config:
    9.      ignition:
    10.        version: 3.2.0
    11.      storage:
    12.        files:
    13.        - contents:
    14.            source: data:text/plain;base64,<encoded_base64_string> (2)
    15.          filesystem: root
    16.          mode: 420
    17.          path: /etc/udev/rules.d/41-zfcp-lun-0.0.8000:0x500507680d760026:0x00bc000000000000.rules (3)
    1The role you have defined in the machine config file.
    2The Base64 encoded string that you have generated in the previous step.
    3The path where the udev rule is located.

Configuring DASD

The following is an example of how to configure a DASD device by adding a udev rule.

Procedure

  1. Take the following sample udev rule 41-dasd-eckd-0.0.4444.rules:

    1. ACTION=="add", SUBSYSTEM=="ccw", KERNEL=="0.0.4444", DRIVER=="dasd-eckd", GOTO="cfg_dasd_eckd_0.0.4444"
    2. ACTION=="add", SUBSYSTEM=="drivers", KERNEL=="dasd-eckd", TEST=="[ccw/0.0.4444]", GOTO="cfg_dasd_eckd_0.0.4444"
    3. GOTO="end_dasd_eckd_0.0.4444"
    5. LABEL="cfg_dasd_eckd_0.0.4444"
    6. ATTR{[ccw/0.0.4444]online}="1"
    8. LABEL="end_dasd_eckd_0.0.4444"

  2. Convert the rule to Base64 encoded by running the following command:

    1. $ base64 /path/to/file/

  3. Copy the following MCO sample profile into a YAML file:

    1. apiVersion: machineconfiguration.openshift.io/v1
    2. kind: MachineConfig
    3. metadata:
    4.    labels:
    5.      machineconfiguration.openshift.io/role: worker0 (1)
    6.    name: 99-worker0-devices
    7. spec:
    8.    config:
    9.      ignition:
    10.        version: 3.2.0
    11.      storage:
    12.        files:
    13.        - contents:
    14.            source: data:text/plain;base64,<encoded_base64_string> (2)
    15.          filesystem: root
    16.          mode: 420
    17.          path: /etc/udev/rules.d/41-dasd-eckd-0.0.4444.rules (3)
    1The role you have defined in the machine config file.
    2The Base64 encoded string that you have generated in the previous step.
    3The path where the udev rule is located.

Configuring qeth

The following is an example of how to configure a qeth device by adding a udev rule.

Procedure

  1. Take the following sample udev rule 41-qeth-0.0.1000.rules:

    1. ACTION=="add", SUBSYSTEM=="drivers", KERNEL=="qeth", GOTO="group_qeth_0.0.1000"
    2. ACTION=="add", SUBSYSTEM=="ccw", KERNEL=="0.0.1000", DRIVER=="qeth", GOTO="group_qeth_0.0.1000"
    3. ACTION=="add", SUBSYSTEM=="ccw", KERNEL=="0.0.1001", DRIVER=="qeth", GOTO="group_qeth_0.0.1000"
    4. ACTION=="add", SUBSYSTEM=="ccw", KERNEL=="0.0.1002", DRIVER=="qeth", GOTO="group_qeth_0.0.1000"
    5. ACTION=="add", SUBSYSTEM=="ccwgroup", KERNEL=="0.0.1000", DRIVER=="qeth", GOTO="cfg_qeth_0.0.1000"
    6. GOTO="end_qeth_0.0.1000"
    8. LABEL="group_qeth_0.0.1000"
    9. TEST=="[ccwgroup/0.0.1000]", GOTO="end_qeth_0.0.1000"
    10. TEST!="[ccw/0.0.1000]", GOTO="end_qeth_0.0.1000"
    11. TEST!="[ccw/0.0.1001]", GOTO="end_qeth_0.0.1000"
    12. TEST!="[ccw/0.0.1002]", GOTO="end_qeth_0.0.1000"
    13. ATTR{[drivers/ccwgroup:qeth]group}="0.0.1000,0.0.1001,0.0.1002"
    14. GOTO="end_qeth_0.0.1000"
    16. LABEL="cfg_qeth_0.0.1000"
    17. ATTR{[ccwgroup/0.0.1000]online}="1"
    19. LABEL="end_qeth_0.0.1000"

  2. Convert the rule to Base64 encoded by running the following command:

    1. $ base64 /path/to/file/

  3. Copy the following MCO sample profile into a YAML file:

    1. apiVersion: machineconfiguration.openshift.io/v1
    2. kind: MachineConfig
    3. metadata:
    4.    labels:
    5.      machineconfiguration.openshift.io/role: worker0 (1)
    6.    name: 99-worker0-devices
    7. spec:
    8.    config:
    9.      ignition:
    10.        version: 3.2.0
    11.      storage:
    12.        files:
    13.        - contents:
    14.            source: data:text/plain;base64,<encoded_base64_string> (2)
    15.          filesystem: root
    16.          mode: 420
    17.          path: /etc/udev/rules.d/41-dasd-eckd-0.0.4444.rules (3)
    1The role you have defined in the machine config file.
    2The Base64 encoded string that you have generated in the previous step.
    3The path where the udev rule is located.

Next steps

Configuring additional devices manually

Tasks in this section describe how to manually configure additional devices in an IBM Z or IBM® LinuxONE environment. This configuration method is persistent over node restarts but not OKD native and you need to redo the steps if you replace the node.

Prerequisites

  • You are logged in to the cluster as a user with administrative privileges.

  • The device must be available to the node.

  • In a z/VM environment, the device must be attached to the z/VM guest.

Procedure

  1. Connect to the node via SSH by running the following command:

    1. $ ssh <user>@<node_ip_address>

    You can also start a debug session to the node by running the following command:

    1. $ oc debug node/<node_name>

  2. To enable the devices with the chzdev command, enter the following command:

    1. $ sudo chzdev -e 0.0.8000
    2.   sudo chzdev -e 1000-1002
    3.   sude chzdev -e 4444
    4.   sudo chzdev -e 0.0.8000:0x500507680d760026:0x00bc000000000000

Additional resources

See Persistent device configuration in IBM Documentation.

RoCE network Cards

RoCE (RDMA over Converged Ethernet) network cards do not need to be enabled and their interfaces can be configured with the Kubernetes NMState Operator whenever they are available in the node. For example, RoCE network cards are available if they are attached in a z/VM environment or passed through in a Fedora KVM environment.

Enabling multipathing for FCP LUNs

Tasks in this section describe how to manually configure additional devices in an IBM Z or IBM® LinuxONE environment. This configuration method is persistent over node restarts but not OKD native and you need to redo the steps if you replace the node.

On IBM Z and IBM® LinuxONE, you can enable multipathing only if you configured your cluster for it during installation. For more information, see “Installing FCOS and starting the OKD bootstrap process” in Installing a cluster with z/VM on IBM Z and IBM® LinuxONE.

Prerequisites

  • You are logged in to the cluster as a user with administrative privileges.

  • You have configured multiple paths to a LUN with either method explained above.

Procedure

  1. Connect to the node via SSH by running the following command:

    1. $ ssh <user>@<node_ip_address>

    You can also start a debug session to the node by running the following command:

    1. $ oc debug node/<node_name>

  2. To enable multipathing, run the following command:

    1. $ sudo /sbin/mpathconf --enable

  3. To start the multipathd daemon, run the following command:

    1. $ sudo multipath

  4. Optional: To format your multipath device with fdisk, run the following command:

    1. $ sudo fdisk /dev/mapper/mpatha

Verification

  • To verify that the devices have been grouped, run the following command:

    1. $ sudo multipath -II

    Example output

    1. mpatha (20017380030290197) dm-1 IBM,2810XIV
    2.    size=512G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
    3.     -+- policy='service-time 0' prio=50 status=enabled
    4.      |- 1:0:0:6  sde 68:16  active ready running
    5.      |- 1:0:1:6  sdf 69:24  active ready running
    6.      |- 0:0:0:6  sdg  8:80  active ready running
    7.      `- 0:0:1:6  sdh 66:48  active ready running

Next steps