Sun Blade T6340 large blade configurations (1.2GHz or 1.4GHz x 8 core) show high fallout with power faults from CPU DC/DC converter.

Asset ID:	1-73-1021650.1
Update Date:	2010-11-04
Keywords:

Solution Type FAB (standard) Sure

Solution 1021650.1 : Sun Blade T6340 large blade configurations (1.2GHz or 1.4GHz x 8 core) show high fallout with power faults from CPU DC/DC converter.

Impact

On Sun Fire T6340 Blade motherboards (501-7984-05) with Delta DC-DC Converters D217 (Locations U7401 and U7501), this fault occurs intermittently after running memory tests in POST, or memory tests and processor tests in SunVTS in system exerciser mode. Any D217 DC-DC Converter having the datecode of 0907 (or later) exhibit this issue more frequently.

Contributing Factors

All Sun Fire T6340 Blades with 1.4GHz x 8 Core or 1.2GHz x 8 Core system boards are impacted by this issue. The 1.4GHz x 8 Core systems exhibit this issue more often.

Symptoms

If you type "showlogs" in alom, you may see the MB_DC_POK faults in blades with 8 Core CPUs.

Example:

Fault | critical: "SP detected fault at time Tue Oct 27 18:17:32 2009. Host Power Failure: MB_DC_POK Fault"

Root Cause

D217 DC/DC Converter noise problem tripping single channel over current threshold which subsequently triggers blade power shutdown.

On system boards already built the supplier reprogrammed the DC-DC Converters D217 (U7401 base address 0xca and U7501 base address 0xda) to turn off this bit# 6-11 at offset 0x0300 via deviation WO_42285 as of October 30, 2009.

For all new builds the supplier is reprogramming D217 parts with the bit# 6-11 at 0x0300 turned off.

The motherboards with the new programmed D217 DC-DC Converters will be dash rolled to 501-7984-06 for future reference.

Corrective Action

Workaround:

Workaround may be accomplished by turning this feature off in converter controller program settings. Over-current protection is covered with a summed channel average current sense function.

Resolution:

Update the Customer's sysfw to v7.2.4.f (see CR#6895793 and CR#6879684), which contains the ILOM patch to turn off the individual phase OCP detection. This firmware patch is currently available as <>

Comments

Primarion controller chip's register bits 6 - 11 at 0x300 were set to assert POK faults even with a single phase over current detection. Noise was causing this bit to be set frequently when memory tests were run either in POST or Sunvts.

ILOM needs to modify the D217 settings on SP boot so these faults do not occur.

References:

Resolution Patches: <>

For information about FAB documents, its release processes, implementation strategies and billing information, go to the following URL:

http://tns.central/fab

For Sun Authorized Service Providers go to:

http://tns.central/SPE/fab/

In addition to the above you may email:

[email protected]

Modification History
Changes made since initial publication.

30-Nov-2009

Changed from Preliminary to Resolved due to availability of T6340 firmware.

Internal Contributor/submitter
[email protected]

Internal Eng Responsible Engineer
[email protected] [email protected]

Internal Services Knowledge Engineer
[email protected]

Internal Eng Business Unit Group
Systems Group - SVS (SPARC Volume Systems, Horizontal Systems, includes T2000/Ontario)

Internal Sun Alert & FAB Admin Info
18-Nov-2009: Completed draft - sending to Extended Review.
20-Nov-2009: No feedback from Ext Rvw - sending to Publish.

Attachments

This solution has no attachment