How to Isolate and Replace Faulty DIMM Pairs

This procedure shows how to isolate and replace faulty DIMMs using the Oracle ILOM.

All DIMMs are configured, identified, and replaced in pairs. When there is a fault in a DIMM, you must replace both the faulty DIMM and the other DIMM in the pair. To identify DIMM pairings, see DIMM Hardware.

Because of the server's architecture, when the server detects a faulty DIMM, it disables other DIMM pairs as well. When the Oracle ILOM displays the faulty DIMMs, it also displays the disabled DIMMs. However, they are clearly identified as disabled, not faulty. When the faulty DIMMs are repaired, the server automatically places the disabled DIMMS back into service.

---

Caution - Before handling components, attach an antistatic wrist strap to a chassis ground (any unpainted metal surface). The system’s printed circuit boards and hard disk drives contain components that are extremely sensitive to static electricity.

---

Before You Begin

This task requires access to the Oracle ILOM and the ability to remove a CMOD from the server. It also requires that you have a pair of DIMMs to replace the faulty pair.

1. Use the Oracle ILOM CLI or web interface to identify the CMOD and the faulty DIMMs.

   • From the CLI:

     Enter the show faulty command.

     If the system has detected a faulty DIMM pair, the output contains one or more entries similar to the following:

     Target              | Property               | Value
     --------------------+------------------------+---------------------------------
     /SP/faultmgmt/0     | fru                    | /SYS/BL0/P0/D0
     /SP/faultmgmt/0/    | class                  | fault.memory.intel.dimm_ue
      faults/0           |                        | 
     
     ... (other information)
     
     /SP/faultmgmt/1     | fru                    | /SYS/BL0/P0/D4
     /SP/faultmgmt/1/    | class                  | fault.memory.intel.dimm_ue

     This identifies the faulty pair as DIMM 0 and DIMM 4 on CPU 0 of CMOD 0 (BL0).

     ---

     Note - This display might include other DIMM pairs that have been temporarily disabled. These are automatically returned to service when the faulty DIMM pair is repaired.

     ---

   • From the Oracle ILOM web interface:

     Click System Information > Fault Management.

     The following display identifies the faulty pair as DIMM 0 and DIMM 4 on CPU 0 of CMOD 0 (BL0).

     

   ---

   Note - This display might include other DIMM pairs that have been temporarily disabled. These are automatically returned to service when the faulty DIMM pair is repaired.

   ---

2. Remove the CMOD that has the faulty DIMMs, and then remove the CMOD cover.

   Refer to Adding, Removing, and Installing a CMOD (CRU) in Sun Server X2-8 (formerly Sun Fire X4800 M2) Service Manual.

3. Press the Fault Remind button on the CMOD to light the DIMM fault LEDs for any faulty DIMM pairs.

   The LEDs on the board adjacent to the DIMM slots light to identify faulty DIMM pairs.

   For the locations of DIMM fault LEDs, DIMM pairs, and the fault remind button, see DIMM Hardware.

   ---

   Note - If the DIMM fault LEDs do not identify the same DIMM pair as the Oracle ILOM, contact Oracle Service.

   ---

4. Visually inspect the DIMMs and the DIMM slots for damage or contamination.

   • If a DIMM is damaged, replace the DIMM pair.

   • If the DIMM slot is cracked or broken, contact Oracle Service.

   • If there is dirt or other contamination on the DIMM or on the slot, you might choose to clean it, reseat the DIMMs, and check the system again.

5. Replace the faulty DIMM pair.

   A pair is two of the following belonging to the same CPU:

   
   

   D0, D4 D8, D12 D1, D5 D9, D13 D2, D6 D10, D14 D3, D7 D11, D15

   Refer to Removing and Installing DIMMs (CRU) in Sun Server X2-8 (formerly Sun Fire X4800 M2) Service Manual for additional information.

6. Reinstall the CMOD in the chassis.
7. If the problem persists, run the following Oracle ILOM command sequence for each faulty DIMM:

   —> set /SYS/MB/Px/Dy clear_fault_action=true
   Are you sure you want to clear /SYS/MB/Px/Dx (y/n) y
   Set 'clear_fault_action' to true

   Where x is the node number and y is the DIMM number.

   This command clears the DIMM of random information that might be interpreted as faulty data.

8. If the problem persists, see How to Isolate and Correct Persistent DIMM Errors.

See Also

• How to Isolate and Correct Persistent DIMM Errors

• DIMM Hardware