Contamination on Gold-Fingers of 541-3819 Memory Risers on X4470, SS7420 and T3-2 may cause Riser-Related Memory Faults

Asset ID:	1-72-1451004.1
Update Date:	2012-06-18
Keywords:

Solution Type Problem Resolution Sure

Solution 1451004.1 : Contamination on Gold-Fingers of 541-3819 Memory Risers on X4470, SS7420 and T3-2 may cause Riser-Related Memory Faults

Applies to:

SPARC T3-2 - Version Not Applicable and later
Sun Fire X4470 Server - Version Not Applicable and later
Sun ZFS Storage 7420 - Version Not Applicable and later
Information in this document applies to any platform.

Symptoms

The symptom of the failure is a memory riser fault with accompanying LED light, seen immediately after a host turn-on or during operation.

Alternatively a "show faulty" command at the SP prompt or an "fmadm faulty" command in the SP faultmanagement shell will indicate a faulty memory riser FRU.

The memory riser FRU will be identified as /SYS/MB/PX/MRY for x86 servers and as /SYS/MB/CMPX/MRY.

Where "X" and "Y" are processor numbers and memory riser numbers respectively.

Example fault locations:

For the X4470 and ZFS storage 7420: /SYS/MB/P1/MR0, /SYS/MB/P1/MR1

For the Sparc T3-2: /SYS/MB/CMP1/MR0, /SYS/MB/CMP1/MR1

For the X4470-based systems [Includes ZFS Storage 7420] the most common symptoms attributable to memory riser problems are:

SPX86-8002-JJ: Memory Branch lane failover








SPX86-8001-XH: Memory Not Present On Branch








SPX86-8001-YD: Memory located on a branch was not successfully initialized

For the SPARC T3-2 systems the most common failure modes attributable to the memory riser are:

SUN4V-8002-PX: Memory link lane failover








SUN4V-8002-Q2: Memory link unrecoverable error








SUN4V-8002-NE: Memory link recoverable CRC errors exceeded acceptable levels

Any other fault which is diagnosed by the SP or Solaris FMA to any of the /SYS/MB/PX/MRY frus should be treated similarly.

This document does not apply to faults that are diagnosed to a DIMM or group of DIMMs, only to faults which are diagnosed to the memory riser.

Changes

N/A

Cause

Contamination on the gold fingers of faulty memory riser boards was noted by a customer. This contamination was analyzed by the board supplier and was attributed to a supplier process problem which was corrected WW11 of 2011. Boards manufactured prior to this work week (datecode 1111) may be subject to this contamination.

Solution

When faults which are diagnosed to memory risers occur, the risers should be replaced.

Attachments

This solution has no attachment