Sun Fire 4810 Server
 
Sun Fire 3800 Server
 
Sun Netra 1290 Server
 
Sun Fire 6800 Server
 
Sun Fire E6900 Server
 
Sun Fire E2900 Server
 
Sun Fire 4800 Server
 
Sun Fire V1280 Server
 
Sun Fire E4900 Server
 
Sun Netra 1280 Server
 

PLA-Support>Sun Systems>SPARC>Enterprise>SN-SPARC: SF-x8x0/Ex900
 
_Old GCS Categories>Sun Microsystems>Servers>Entry-Level Servers
 
_Old GCS Categories>Sun Microsystems>Servers>Midrange Servers
 
_Old GCS Categories>Sun Microsystems>Servers>Midrange V and Netra Servers
 

Applies to:

Purpose

Description

This document provides the basic troubleshooting steps to follow when needing to diagnose the cause of Interconnect Errors on Sun Fire[TM] Midrange Servers

Symptoms:

• A System Board, I/O Board, or Repeater may have been recently serviced, replaced, or reseated.
• A domain may not be able to boot.
• A domain could be described as down, can't be setkeyswitched on, can't be powered on, or as having failed POST.
• Error messages displayed in the System Controller (SC) log files (showlogs -v) or on the console could include messages like:

Failed AR interconnect test.

CPU Board V3 at /N0/SB1 has been removed from domain C due to a failure in interconnection test.
Service action required.

AR Interconnect test: System board SB1/ar0 address repeater connections to system board RP3/ar0 failed

DX Interconnect test: System board /N0/SB1 data line connections to system board RP0 failed

NOTE:  The example errors can be associated to any domain, RP, or any System Board (SB) or I/O Board (IB), and the examples above are not exclusive to these faults.

System Type:

• Sun Fire[TM] v1280, 3800, 4800, 4810, 6800, E2900, E4900, E6900
• Netra[TM] 1280, 1290

Troubleshooting Steps

Steps to Follow
Collect the appropriate troubleshooting data and contact Oracle Support Services.
The error you have encountered is a board interconnection issue.  Essentially, this is a board connectivity issue.  It is likely a hardware defect, a board or slot issue, or a board "seating" issue.  The event requires that a Sun Support Engineer is engaged to diagnose and resolve this event.

Please contact Oracle Support Services in order to diagnose this issue.  Being prepared with the following troubleshooting data will allow that engineer to immediately begin diagnosis of the issue, and decrease the time to resolution.

Please provide:

• Explorer with scextended or 1280extended option (depending on platform type);  See Document 1019066.1 for details
•  When Explorer data can not be captured, please obtain the list of System Controller (SC) commands from Document 1003529.1.
  
  

Please validate that each troubleshooting step below is true for your environment.

The steps will provide instructions or a link to a document, for validating the step and taking corrective action  as necessary.

The steps are ordered in the most appropriate sequence to isolate the issue and identify the proper resolution. 

Please do not skip a step.

1.  Verify the components implicated in the interconnect errors were not recently replaced, reseated, or "handled".

   - Recently "handled" hardware would include any board that has been removed or inserted to replace it or hardware components contained on it.

  - Since the error is an interconnection problem, the physical act of servicing or handling the board could be the cause of the problem.

   Reference: 

Document 1019218.1 Sun Fire[TM] Midrange Servers:  How to identify pin or socket damage.

2.  Verify that the errors persist after executing System Controller Failover (dual SC config) or an SC Reset (single SC config).

   - Failover (scfailover) is only available on systems with Dual SCs.

     Reference:

Document 1003245.1

Sun Fire[TM] 3800-6900: System Controller failover functionality

   - On Sun Fire[TM] v1280/E2900 and Netra[TM} 1280/1290 (single SC configurations) you will need to utilize the resetsc command to

     reset the SC and confirm it's sanity.

     Reference:

Document 1012388.1

Sun Fire[TM] V1280/2900 LOM Quick Command Reference

    - If errors persist on both SCs or after the resetsc is issued, proceed to Step 3.

    - If errors go away after the resetsc you are done.

    - If they go away after executing scfailover, fail back to the original Main SC and confirm the errors persist again.

    - Replace the SC if they do.

3.  Confirm that you are able to determine the suspect list for this issue and prioritize which suspect is most likely to be root cause.

  - See

Document 1019649.1

How to determine the suspect list for Sun Fire[TM] Midrange Server interconnect errors.

4.  Verify that the primary FRU is NOT defective (primary FRU determined by the results of Step 3).

  - If a System Board or I/O Board is implicated, it can be verifief as defective two different ways:

    - By replacing the board.

    - By having a Sun engineer move the suspect board into an empty slot or switch it with another board in the domain and observe the behavior.

        - If the board works in the alternate slot, the RP or the board slot (CP) is implicated (proceed to Step 4).

        - If the board fails to work in the alternate slot, the board is defective, so replace it.

        - If a Repeater (RP) is implicated, it can be verified as defective two different ways:

            - By replacing it.

            - By having an engineer switch the suspect RP with an alternate RP in th system and observe the behavior.

       - If the error follows the RP to it's new location, then the RP is defective, so replace it.

       - If the failure remains at the old RP's slot, then the Centerplane is suspect.

  - The Sun engineer performing any replacement or moving any hardware should be extremely careful to inspect the board and CP pins and sockets.

   Reference:

Document 1019218.1

Sun Fire[TM] Midrange Servers:  How to identify pin or socket damage.

 5.  Verify that the secondary FRU is not defective (secondary FRU determined by the results of Step 3).

  - If a System Board or I/O Board is implicated, it can be verified as defective two different ways:

    - By replacing the board.

    - By having a Sun engineer move the suspect board into an empty  slot or switch it with another board in the domain and observe the behavior.

        - If the board works in the alternate slot, the RP or the board slot (CP) is implicated (proceed to Step 4)

        - If the board fails to work in the alternate slot, the board is defective, so replace it.

        - If a Repeater (RP) is implicated, it can be verified as defective two different ways:

            - By replacing it.

            - By having a Sun engineer switch the suspect RP with an alternate RP in the same system and observe the behavior.

  - If the error follows the RP to it's new location, then the RP is defective, so replace it.

  - If the failure remains at the old RP's slot, then the Centerplane is suspect.

  - The Sun engineer performing any replacement or moving any  hardware should be extremely careful to inspect the board and CP pins and sockets. 

  Reference:

Document 1019218.1

Sun Fire[TM] Midrange Servers: How to identify pin or socket damage.

6.  Collaborate with TSC prior to proceeding to a Centerplane replacement.

  - Make sure to have console data, explorer data, and a detailed explanation of what has been replaced, and when available when collaborating with TSC.

  - Most likely the Centerplane will have to be replaced, but TSC will want make absolutely sure that nothing has been overlooked before proceeding to this invasive replacement action.

NOTE: The testinterconnect command can be utilized to test board interconnections if you obtain a service mode password (setkeyswitch on also accomplishes this testing). For details on testinterconnect command usage refer to

Document 1005014.1

.

Keywords:

interconnect, Interconnect, interconnect test, interconnection,

testinterconnect, Service action required, failure, POST, normalized, Mapped, Global_Oring

References

This solution has no attachment