Netra(TM) Supply Rail Failure
supply rail 5 FAULT: failed
NETRA(TM) SUPPLY RAIL FAULT FAILURE
====================================
When Fault LED flashing, a fatal condition could possibly
exist in the system. This condition may be caused by the
following:
- Low fan speed
- Ambient case temperature
- Supply rail voltage
- Ambient System Board/CPU temperature
- Power supply
Problem Example
===============
lom>loghistory
Eventlog:
+22d+0h33m28s supply rail 5 recovered
+22d+1h32m15s supply rail 5 FAULT: failed
+22d+3h20m54s supply rail 5 recovered
+22d+4h55m49s supply rail 5 FAULT: failed
+22d+5h53m16s supply rail 5 recovered
+23d+7h25m59s supply rail 5 FAULT: failed
+23d+7h38m10s supply rail 5 recovered
+23d+9h40m26s supply rail 5 FAULT: failed
+23d+9h44m44s supply rail 5 recovered
+24d+5h40m1s Fault LED OFF
Problem Description
===================
The LOM expects to measure 1.89v for the processor core voltage.
VDD_core is calibrated for 1.9V and not 1.75volts. The PSU controls
it at 1.75v. The LOM EEPROM config is incorrect. This is 2.5% above
the 10% VDD core low voltage warning. The PSU feedback loop will
continue to drive the voltage to 1.75v. It should only dip below the
10% level if it is truly broken. However, some systems may start to
report a low voltage warning with age. The fix in this case would be
to supply a patch for Service to update the LOM EEPROM config file
in the field.
Corrective Action
=================
Patch/Bug Fixes, Error reset
1. Patch-ID# 112140-01
or greater
Keywords: LOM netra t1
Synopsis: Lomlite2 lom_update_eeprom patch
2. Patch-ID# 110208-17
or greater
Keywords: netra lom firmware
Synopsis: Netra Lights Out Management 2.0 patch
3. BUG ID's #4417373
& #4417843
The value against which VDD_core voltage is checked
FJ2-V1.1-258-7884-06.es appears to be set incorrectly.
4. lom> env ->This will only report what's in memory
lom> check ->This will reset the list and keep an error
from being reported
Supply Monitoring Code Fixes:
1. Making the tempsens module/supply rail monitoring
code to correct logical to physical re-numbering as
required.
Fix files:
prom/lomlite2/3.x-h8-3437/firmware/lom.h, v1.6
prom/lomlite2/3.x-h8-3437/firmware/supply.c, v1.5
prom/lomlite2/3.x-h8-3437/firmware/tempsens.c, v1.10
prom/lomlite2/3.x-h8-3437/firmware/ebus.c, v1.12
Reference documents
====================
4480653
Mechanism to deliver LOM firmware config
files
4417373
VDD_core check voltage appears to be wrong in eeprom
4417843
Calibration of VDD_core is wrong in eeprom configuration file
Document Creation: Steven Bock, June 11, 2003