Skip to main content

Power on and power off problems

Use this information to resolve issues when powering on or powering off the compute node.

Embedded hypervisor is not in the boot list​

Complete the following steps until the problem is solved.

  1. If the compute node has been installed, moved, or serviced recently, or if this is the first time the embedded hypervisor is being used, make sure that the device is connected properly and that there is no physical damage to the connectors.

  2. See the documentation that comes with the optional embedded hypervisor flash device for setup and configuration information.

  3. Check Lenovo ServerProven website to validate that the embedded hypervisor device is supported for the compute node.

  4. Make sure that the embedded hypervisor device is listed in the list of available boot options. From the management controller user interface, click Compute node Configuration > Boot Options.

    For information about accessing the management controller user interface, see the Opening and Using the XClarity Controller Web Interface section in the XCC documentation version compatible with your server at Lenovo XClarity Controller portal page.

  5. Check Lenovo Data Center Support for any tech tips (service bulletins) related to the embedded hypervisor and the compute node.

  6. Make sure that other software works on the compute node to ensure that it is working properly.

Single compute node does not power on​

Complete the following steps until the problem is resolved:
  1. If you recently installed, moved, or serviced the compute node, reseat the compute node in the bay. If the compute node was not recently installed, moved, or serviced, perform a virtual reseat through the CMM service command. For more information about the CMM service command, see the "Flex System Chassis Management Module: Command-Line Interface Reference Guide".
  2. Check the event log on the CMM 2 for any events that are related to the compute node and solve them.
  3. Make sure that the CMM 2 can recognize the compute node. Log in to the CMM 2 user interface and verify that the compute node appears in the chassis view. If the CMM 2 cannot recognize the compute node, remove it, and inspect the compute node and the back of the node bay to make sure that there is no physical damage to the connectors.
  4. Make sure that the power policy that is implemented on the CMM 2 is sufficient to enable the compute node to power on. You can view the power policy using the CMM 2 pmpolicy command or the CMM 2 web interface.
  5. Replace the system-board assembly (see System-board assembly replacement).
    Note
    Until you are able to replace the system-board assembly, you can attempt to power on the compute node from the CMM 2.

Multiple compute nodes do not power on​

Complete the following steps until the problem is resolved:
  1. If you recently installed, moved, or serviced the computer nodes, reseat the compute nodes in the bays. If the compute nodes were not recently installed, moved, or serviced, perform a virtual reseat through the CMM service command. For more information about the CMM service command, see the "Flex System Chassis Management Module: Command-Line Interface Reference Guide".
  2. Check the CMM 2 event log for any events that are related to the compute nodes, and solve them.

Compute node does not power off​

Complete the following steps until the problem is resolved:

  1. Attempt to power off the compute node through the CMM 2 interface.

  2. Attempt to restart the system-management processor for the compute node through the CMM 2 interface. Click the compute node in the chassis view and then click Restart System Mgmt Processor. After the system-management processor has been restarted, attempt to power off the node from the CMM 2.

  3. Attempt to power off the compute node by using the power button on the front of the compute node.

  4. Attempt to reset the compute node from the CMM 2 command-line interface (CLI) by using the reset command.

  5. Reseat the CMM 2. Then, perform steps 1-4 again.