[c-nsp] Strange thermal issues with 3750E and RPS 2300

[Brian Landers]

We have recently implemented a stack of 3750E switches with RPS 2300
redundant power units at two different sites, and are currently seeing
a flood of "strange" syslog message from both.  I'll be opening a TAC
case, but wanted to reach out to see if anyone else has run into this
in case it's something blindingly stupid.  The errors in question:

Aug 13 17:00:01 xxx.net 52745: Aug 13 21:00:01.675 UTC:
%PLATFORM_ENV-3-RPS_POST_FAILED: RPS POST failed
Aug 13 17:00:02 xxx.net 52746: Aug 13 21:00:01.675 UTC:
%PLATFORM_ENV-1-RPS_PS_THERMAL_CRITICAL: RPS power supply A
temperature has reached critical threshold
Aug 13 17:00:02 xxx.net 52747: Aug 13 21:00:01.675 UTC:
%PLATFORM_ENV-1-RPS_PS_THERMAL_CRITICAL: RPS power supply B
temperature has reached critical threshold
[repeat over and over]

What's odd about this:

- we're seeing the same behavior from two different sites
- both are in temperature-controlled data centers

sw1.xxx#show env temp
TEMPERATURE is OK

RPS Name:
  State: Standby
  PID:
  Serial#:
  Fan: Good
  Temperature: Green

  RPS Power Supply A: Failure-Thermal
    PID             : C3K-PWR-750WAC
    Serial#         : DTN1252E0Q8
    System Power    : Good
    PoE Power       : Good
    Watts           : 300/420 (System/PoE)

  RPS Power Supply B: Failure-Thermal
    PID             : C3K-PWR-750WAC
    Serial#         : DTN1252E0U3
    System Power    : Good
    PoE Power       : Good
    Watts           : 300/420 (System/PoE)


TIA,
B*

--
Brian C Landers
http://www.packetslave.com/
CCIE #23115, RHCE