M8000/M9000 How to Replace a CPU:ATR:630:4

Asset ID:	1-71-1303263.1
Update Date:	2016-01-20
Keywords:

Solution Type Technical Instruction Sure

Solution 1303263.1 : M8000/M9000 How to Replace a CPU:ATR:630:4

Applies to:

Sun SPARC Enterprise M8000 Server - Version Not Applicable and later
Sun SPARC Enterprise M9000-32 Server - Version Not Applicable and later
Sun SPARC Enterprise M9000-64 Server - Version Not Applicable and later
Information in this document applies to any platform.

Goal

CAP PROBLEM OVERVIEW: CPU Replacement
********************************************************************************
To report errors or request improvements on this procedure,
please go to http://support.us.oracle.com and put a comment on Doc ID: 1303263.1
********************************************************************************

Solution

DISPATCH INSTRUCTIONS

WHAT SKILLS DOES THE ENGINEER NEED:(IS A SITE ENGINEER AVAILABLE?) Mseries training and basic field engineer skills

DELIVERY REQUIREMENT: Platform outage may be required. This should be negotiated with the customer during contracted support hours.

TASK COMPLEXITY: 4

Time Estimate: 90 minutes

FIELD ENGINEER INSTRUCTIONS

WHAT STATE SHOULD THE SYSTEM BE IN TO BE READY TO PERFORM THE RESOLUTION ACTIVITY? :

A data backup is not a pre-requisite but is a wise precaution
For an active replacement the customer is not required to perform a system shutdown
For a hot replacement the customer is required to perform a domain(s) shutdown
For a cold replacement the customer should conduct an orderly software system shutdown. Then power down the entire platform.

DISPATCH INSTRUCTIONS

WHAT SKILLS DOES THE ENGINEER NEED:(IS A SITE ENGINEER AVAILABLE?)
The engineer should be M8000/M9000 trained

FIELD ENGINEER INSTRUCTIONS

Review and follow <Document 1021062.1> while working with removed CMU.

WHAT ACTION DOES THE ENGINEER NEED TO TAKE:

Active Replacement

If domain consists of all Sparc64 VII or Sparc64 VII+ processors the domain will have had to have been powered on with cpumode set to compatible for DR to work with a Sparc64 VI CMU.
If the XSB is in quad mode the sb will need to be removed from all domains it is associated with.

1. Connect to serial port of the active XSCF to ensure that communication will not be lost
2. Execute showfru, showdcl, showboards to identify the affected domains

Note: IO to the XSB must be quiesced to perform the next step. See the DR Users Guide chapter 2 for more information.

3. If the xsb contains an IOUA card the card will need to be deconfigured.

XSCF> cfgdevice -l
XSCF> cfgdevice -c detach -p port_no

   4. Execute the deleteboard command to dynamically disconnect the CMU to be removed from all the affected domains.
   5. Execute the replacefru command.
      - On the maintenance menu, select 1, "CMU/IOU."
      - From here, perform the operation according to the guidance displayed in the maintenance menu
   6. Remove the CMU
   7. Replace the CPU
   8. Reinstall the CMU
   9. Continue to follow the instructions from the maintenance menu. When the display of the maintenance menu returns to the initial page exit the maintenance menu
   10. Execute the addboard command to dynamically connect the IOU to all the affected domains

Hot Replacement

Connect to serial port of the active XSCF to ensure that communication will not be lost
Execute showfru, showdcl, showboards to identify the affected domains
Execute the poweroff -d command to shut down the affected domains
Execute the replacefru command

On the maintenance menu, select 1, "CMU/IOU."
From here, perform the operation according to the guidance displayed in the maintenance menu

Remove the CMU
Replace the CPU
Reinstall the CMU
Continue to follow the instructions from the maintenance menu. When the display of the maintenance menu returns to the initial page exit the maintenance menu
Execute the poweron -d command to power on the affected domains

Cold Replacement

Customer should perform an orderly shutdown of the domains.
Change the mode switch setting on the operator panel from Locked to Service.
From a console that is connected to XSCF, log in to XSCF, and execute the poweroff -a command.
Confirm that the power-off processing is completed by confirming that the POWER LED (green) on the operator panel is off.

Note: On a M9000 with an expansion cabinet the base cabinet should be switched off first then the expansion cabinet.

Switch off all main line switches for the AC section (ACS) in the power supply system.
Remove the CMU
Replace the CPU.
Reinstall the CMU

Note: On a M9000 with an expansion cabinet the expansion cabinet should be switched on first then the base cabinet.

Switch on all main line switches for the AC section (ACS) in the power supply system.
Confirm that the XSCF STANDBY LED (green) on the operator panel remains lit. If it is blinking, wait until it remains lit.
Start (turn on power) all domains together

When the maintenance work is completed, be sure to confirm the mode switch setting and return the switch to Locked.

OBTAIN CUSTOMER ACCEPTANCE

WHAT ACTION DOES THE CUSTOMER NEED TO TAKE TO RETURN THE SYSTEM TO AN OPERATIONAL STATE:

Check full functionality has been retained

Confirm hardware by running the following commands:
XSCF> showhardconf -M
XSCF> showstatus

PARTS NOTE:

None

REFERENCE INFORMATION:
For more information on DR please use the Dynamic Reconfiguration (DR) Users Guide http://download.oracle.com/docs/cd/E19415-01/821-2796-11
Service Manual: http://download.oracle.com/docs/cd/E19415-01/index.html

Attachments

This solution has no attachment