Fujitsu M10-1
 
Fujitsu M10-4S
 
Fujitsu M10-4
 
Fujitsu M10 PCI Expansion Unit
 

PLA-Support>Sun Systems>SPARC>Enterprise>SN-SPARC: Fujitsu M10
 

In this Document

Goal

Solution

Applies to:

Goal

The goal is to get the PCI Expansion Unit and the M10 it is connected to ready for maintenance.  Customers should perform the actions listed in the Fix section below and the Field Engineer should confirm these steps have been performed before starting maintenance.

Solution

 For additional information on the following information see the PCI Expansion Unit Service Manual 

Table 1 list the maintenance types for the various FRUs in a PCI Expansion Unit connected to a M10-1 and Table 2 list the same information for the M10-4 and M10-4s

Table 1 When the connection destination is a SPARC M10-1

1 - The link board can be mounted only on its dedicated slot.
2 - Needs to be released from the system.
3 - Available only with the redundant configuration

Table 2 When the connection destination is a SPARC M10-4 or SPARC M10-4s

1 - The link board can be mounted only on its dedicated slot.
2 - Needs to be released from the system.
3 - Available only with the redundant configuration
4 - The link card that is mounted on the connection destination server needs to be released from the system with the PCI hot plug (PHP) function.

An explanation of the five types of maintenance is as follows:

Active/hot maintenance

The maintenance is performed while the power to the physical partition to which the PCI expansion unit under maintenance belongs is on, and input power to the PCI expansion unit on which the target Field Replaceable Unit (FRU) is mounted is on (the power cord is connected). The active/hot maintenance is available only on the redundant configuration or on FRU that can use the PCI hot plug (PHP).

Active/cold maintenance

The maintenance is performed while the power to the physical partition to which the PCI expansion unit under maintenance belongs is on and input power to the the target PCI expansion unit is off (the power cord is removed). Release the link card that is mounted on the connection destination server from the system with the PCI hot plug (PHP) function.

Inactive/hot maintenance

The maintenance is performed while the power to the physical partition to which the PCI expansion unit under maintenance belongs is off and input power to the target PCI expansion unit is on (the power cord is connected).

Inactive/cold maintenance

The maintenance is performed while the power to the physical partition to which the PCI expansion unit under maintenance belongs is off and input power to the target PCI expansion unit is off (the power cord is removed).

System stopped maintenance

The maintenance is performed while the power to all the physical partitions is off and input power to the target PCI expansion unit is off (the power cord is removed).

Removing the PCI Expansion Unit from the server:

1. Log in to the XSCF shell.
2. Execute the ioxadm command to make the CHECK LED of the PCI expansion unit under maintenance blink.
   
   In the following example, the CHECK LED of PCIBOX#2008 is specified.
   
   XSCF>ioxadm locator on PCIBOX#2008
           Location Sensor Value Resolution Units
           PCIBOX#2008 LOCATE Blink - LED
     
3. Execute the ioxadm command to check the server to check the  PCI expansion it is connected to:
   
   XSCF> ioxadm list PCIBOX#2008
     
4. Execute the showboards command to check the physical partition that uses the link card
   
   XSCF> showboards -a
     
5. Stop the guest domain that virtually uses the PCI Express (PCIe) card of the PCI expansion unit
   Display the Oracle Solaris super-user prompt.
   Execute the svcadm(1M) command to enable the hotplug service.
   
   # svcadm enable hotplug
     
6. For the active/hot maintenance, execute the following command to check the mounting location by making the locator LED of the component under maintenance blink.
   The executed command depends on the component under maintenance.
   
   
   Note - No locator LED is mounted on link board, link cable, management cable, power supply unit, PCI tray, and fan backplane.  Go to the next Procedure
    
   
   1. For the the PCIe card execute the cfgadm(1M) command.
      
      To specify Ap_Id, execute the cfgadm -a command and check the Ap_Id of the PCIe slot.
      # cfgadm -x led=attn, mode=blink Ap_ID
       
   2. For the fan unit and the I/O board execute the ioxadm command of the XSCF firmware.
      To specify the component under maintenance, execute the ioxadm list command to check the component number, and input it.
      In the following example, the FAN#0 of PCIBOX#2008 is specified.
      
      XSCF> ioxadm locator on PCIBOX#2008/FAN#0
       
      
7. Execute the cfgadm(1M) command to release the link card from the logical domain
   To specify Ap_Id, execute the cfgadm -a command and check the Ap_Id of the link card.
   
   # cfgadm -c unconfigure Ap_ID
    
   
   
8. Execute the cfgadm(1M) command to stop the power supply to the link card
   To specify Ap_Id, execute the cfgadm -a command and check the Ap_Id of the link card.
   
   # cfgadm -c disconnect Ap_ID
    

Removing the Server which the PCI Expansion Unit is connected to:

The procedures for removing the server, which the PCI expansion unit under maintenance is connected to, from the system is:

1. Log in to the XSCF shell.
2. Execute the replacefru command to check the maintenance menu.
   
   XSCF> replacefru
    
   
3. Once the maintenance is complete restore the Server by completing the replacefru command

Turning off the power to PPAR:

Turning off the power to the physical partition (ppar) to which the PCI expansion unit under maintenance belongs. 

1. Execute the ioxadm command of the XSCF firmware to make the CHECK LED of the PCI expansion unit under maintenance blink.
   In the following example, the CHECK LED of PCIBOX#12B4 is specified
   
   XSCF> ioxadm locator on PCIBOX#12B4
    
   
   
2. Execute the ioxadm command to check the server to which the PCI expansion unit under maintenance is connected.
   
   XSCF> ioxadm list PCIBOX#12B4
    
   
   
3. Execute the showboards command to check the physical partition that uses the link card.
   
   XSCF> showboards -a
    
   
   
4. Execute the poweroff command to stop all logical domains on the physical partition under maintenance.
   
   XSCF> poweroff -p ppar_id
    

To restore power to the PPAR after the maintenance:

Stopping Entire System

The procedure to stop the entire system.

To restore power to the PPAR after the maintenance:

For cold maintenance

Remove the power cords from the PCI Expansion Unit

After the maintenance restore the power cords

This solution has no attachment