On Sun Systems, a Small Number of 256MB DIMMs May Experience Premature Failures with Uncorrectable Memory Error (UE) Messages

---

Impact

Sun has identified a set of 256MB DIMMs that may experience premature failures displayed as uncorrectable memory error (UE) messages. A 12 week date code range of B-die 256MB DIMMs in total are affected.

The described issue typically occurs during the first 50 weeks of memory module usage.

Contributing Factors

This issue can occur in the following releases:

SPARC

• 256MB DIMM with Part Number 501-5401-02 or 501-5401-03

And with a module date code between 0115 and 0127 which are B-die modules built between weeks 15 and 27 of 2001. (Note the module date code is given on a white label on the DIMM)

In order to identify affected DIMMs in a system, the approximate overall serial number range is:

	501540178190000 to 501540178310000

A physical check can be performed on the part number (also given on a white label on the DIMM) to verify it is a B-die. The following part number is an example of a B-die DIMM:

	M323S1742BT2-C1LS0
                 ^
                 |  Denotes B-die

C-die DIMMs would have a "C" in place of the "B" as seen above.

Note: The format for serial numbers (for example, 501540178190000) is :

5015401	Sun Part Number
78	Vendor Code
nnnnnn	Unique Identifier (e.g. "190000")

SPARC Platforms that may be affected:

• Sun Fire (280R, V480, V880)
• Sun Blade 1000/2000
• Netra 20
• Sun Fire Servers (3800/4800/4810/6800)
• Sun Fire 12K Servers
• Sun Fire 15K Servers

Symptoms

Sample error messages include:

	Multiple uncorrectable memory failures in a single machine.

and sample specific strings including:

	Multiple Softerrors
	Sticky Softerrors accumulated ... unix:  [ID 340762 kern.notice] from Memory Module ...
	Uncorrectable system bus (UE) Event on CPU0 Privileged Data Access at ...
	WARNING:  [AFT1] EDU Event on CPU0 ...
	ECC errors ..	
	WARNING:  [AFT1]  WDU Event on ...
	UE   EDU   WDU   Error(s)...

Workaround

Please see Resolution below.

Resolution

Upon failure, the affected modules will be replaced. This issue is addressed via a Field Change Order (FCO).

Modification History


Date: 03-NOV-2003

• Updated Contributing Factors

Date: 23-MAR-2004

• State: Resolved
• Updated Resolution section

Attachments

This solution has no attachment