7.3 Servicing the Oracle Private Cloud Appliance Rack System

This section provides instructions to service replaceable components (CRUs/FRUs) in the appliance rack. Before starting any service procedure, read and follow the guidelines in Section 7.2, “Preparing Oracle Private Cloud Appliance for Service”.

7.3.1 Powering Down Oracle Private Cloud Appliance (When Required)

Some service procedures may require you to power down the Oracle PCA. Perform the following steps to manually power down the system.

Caution

Whenever a hardware system must be powered down, make sure that the virtual machines hosted by that system are shut down first. If you power down the appliance with running virtual machines, these will be in an error state when the system is returned to operation.

For details, consult the Oracle VM Manager User's Guide.

Shutting down the Oracle VM environment

  1. Log in to Oracle VM Manager and open the Servers and VMs tab.

  2. Using the navigation tree, select each virtual machine and click Stop to shut it down gracefully.

    If the applications hosted by your VMs require the services and machines to be shut down in a particular order, respect those requirements just like you would with physical machines.

    Once the VMs have been shut down, you can proceed to power off the compute nodes.

  3. Using the navigation tree, select each compute node and click Stop Server to shut it down gracefully.

  4. Using SSH and an account with superuser privileges, log into the active management node at the management virtual IP address. Stop Oracle VM Manager by entering the command service ovmm stop.

Powering down the system for service

  1. If, at this point, any compute nodes have not shut down properly, press the Power button on the running compute nodes in order to shut them down gracefully.

  2. Press the Power button on the management nodes in order to shut them down gracefully.

    Once the servers are powered off, you can proceed to power off the storage appliance.

  3. Press the Power button on the storage server heads attached to the chassis of the storage device.

  4. Press the Power button on both Fabric Interconnects.

  5. Toggle the rack Power switches to the Off position.

    Note

    The Ethernet and InfiniBand switches do not have power switches. They power off when power is removed, by way of the power distribution unit (PDU) or at the breaker in the data center.

Returning the system to operation after service or unplanned outage

  1. Toggle the power distribution unit (PDU) circuit breakers of both PDUs to the On position.

  2. Wait at least two minutes to allow the PDUs to complete their power-on sequence.

    The Ethernet and InfiniBand switches are powered on with the PDUs.

  3. Press the Power button on each Fabric Interconnect.

    Wait until the Status LED stops blinking and is solid green, indicating the system has finished booting.

  4. Press the Power button on the storage server heads.

    Wait approximately two minutes until the power-on self-test completes, and the Power/OK LED on the front panel lights and remains lit.

  5. Press the Power button on the management nodes.

    The management node that completes booting first assumes the master role.

    Note

    Compute nodes do not power on automatically like the internal ZFS Storage Appliance, switches and other components. Make sure that the management nodes and internal storage are up and running, then manually power on the compute nodes.

  6. When the management nodes are up, press the Power button on the compute nodes.

    Caution

    The compute node ILOM policy for automatic power-on is disabled, and must remain disabled, to prevent a server from booting prematurely and disrupting the correct boot order of the appliance components.

    When all compute nodes are up, verify the status of all system components in Oracle VM Manager.

    If no components are in error state, the appliance is ready to resume normal operation.

7.3.2 Service Procedures for Rack System Components

For parts that are not hot-swappable, power down the Oracle PCA before starting the service procedure. Generally speaking, hot-swappable components can be serviced without specific additional steps.

Table 7.13 Service Instructions for Rack System Components

Replaceable Part(s)

Hot-Swap

Instructions

Power cables

Ethernet cables

InfiniBand cables

SFP transceivers

(Oracle-qualified service technician only)

Cable management arms (CMAs)

(Oracle-qualified service technician only)

For removal and installation of a cable management arm, refer to the Oracle Server X5-2 Installation Guide (part no. E48314).

Slide rails

(Oracle-qualified service technician only)

To service the slide rails, the server must be removed from the rack. For instructions, refer to the Oracle Server X5-2 Service Manual (part no. E48320).

For slide rail installation instructions, refer to the section Attach Tool-less Slide-Rail Assemblies in the Oracle Server X5-2 Installation Guide (part no. E48314). To remove the slide rails, reverse the installation steps.