Sun Storage[TM] 3510 FC Array Experiences Unexpected SES Events

Asset ID:	1-71-1012828.1
Update Date:	2010-12-21
Keywords:

Solution Type Technical Instruction Sure

Solution 1012828.1 : Sun Storage[TM] 3510 FC Array Experiences Unexpected SES Events

Applies to:

Sun Storage 3510 FC Array
All Platforms

Goal

Description

Customers have reported that every 30 seconds as per polling interval, events were getting reported suggesting a bad power supply which can only be cleared by a power cycle of the Sun Storage[TM] 3510 FC Array.
Such events include: false power supply, not available messages, and amber LEDs on drives, controller, front panel and power supply.

Reported cases were with seen with DC power supplies.

Errors from persistent eventlog:

Thu Jun     8 14:33:28 2006
[Primary]    Notification
SES(C2 I12)Power Supply 1:Power Supply Back On-Line

Thu Jun     8 14:34:25 2006
[Primary]    Alert
SES(C2 I12)Power Supply 0:Device Not Availabled !

Thu Jun     8 14:34:25 2006
[Primary]    Alert
SES(C2 I12)Power Supply 1:Device Not Availabled !

Thu Jun     8 14:35:55 2006
[Primary]    Notification
SES(C2 I12)Power Supply 0:Power Supply Back On-Line

Thu Jun     8 14:35:55 2006
[Primary]    Notification
SES(C2 I12)Power Supply 1:Power Supply Back On-Line

Also seen:

Thu Jun  8 18:00:44 2006
[3F3f] #5542:  SES Device on StorEdge Array SN#8050259 CH2 ID12: ALERT: PS 0 power supply state changed to critical

Thu Jun  8 18:00:44 2006
[3F3f] #5543:  SES Device on StorEdge Array SN#8050259 CH2 ID12: ALERT: PS 1 power supply state changed to critical

Thu Jun  8 18:01:42 2006
[3Fdf] #5544:  SES Device on StorEdge Array SN#8050259 CH2 ID12: NOTICE: PS 0 power supply is no longer critical

Thu Jun  8 18:01:42 2006
[3Fdf] #5545:  SES Device on StorEdge Array SN#8050259 CH2 ID12: NOTICE: PS 1 power supply is no longer critical

In addition, the following amber LEDs were reported:

- Disks 0 and 1
- secondary controller (lower)
- all front panel LEDs except for fan
- either left or right power supply (varies across power cycles).

Name: FC_CHASSIS_BKPLN
Description: SE3510 FC Chassis/backplane
Part Number: 370-5535
Serial Number: 09862F
Revision: 02
Manufacturing Date: Wed Sep 14 09:13:59 2005
Manufacturing Location: Suzhou,China
Manufacturer JEDEC ID: 0x0301
FRU Location: FC MIDPLANE SLOT
Chassis Serial Number: 09862F
FRU Status: Fault

Solution

Steps to Follow

Version 1102 and later provided an SES firmware revision that does not monitor the LM75C temperature sensors on the Storage[TM] 3510 midplane.

Please validate that each troubleshooting step is applicable to your
storage array:

1. Verify the version of SES firmware is at 1102 or above via:

sccli> show ses

2. If below revision 1102, upgrade the SES firmware by installing the latest version in: Patch 113723.

3. If SES firmware cannot be installed, alternative resolution is to power cycle the array.

Internal Comments See CR 6442784

The root cause of the problem is the result of a deficiency in the LM75 temperature sensor
located on the midplane of the SE3510 resulting in the I2C state machine lockup
and the LED freeze. Test @results indicate the noise level of the SCL signal on the
SE3510 midplane triggered the LM75C I2C @state machine lockup due to the inadequate noise
immunity of the National Semiconductor LM75C.

Since the I2C state machine lockup can ONLY occur when the device is addressed, removing the @monitoring of the device will eliminate the potential I2C hang condition of the LM75C. The output @values of the temperature sensors will be derived from the IOM temperature sensor thereby ensuring @backward compatibility with any script that may be monitoring using sccli software.

Change History
Date: 2010-12-21
User Name: susan.copeland@oracle.com
Action: remove sunsolve link

Attachments

This solution has no attachment