Score:0

How to interpret weird ipmitool thresholds

us flag

On a CentOS 4 IBM x3650 server, ipmitool -v sensor | grep -A 14 "PECI" gets me this:

Sensor ID              : CPU 1 PECI tics (0x7c)
 Entity ID             : 3.1
 Sensor Type (Analog)  : Temperature
 Sensor Reading        : -55 (+/- 0) unspecified
 Status                : ok
 Lower Non-Recoverable : -64.000
 Lower Critical        : na
 Lower Non-Critical    : 64.000
 Upper Non-Critical    : 0.000
 Upper Critical        : na
 Upper Non-Recoverable : na
 Event Status          : Event Messages Disabled
 Assertion Events      :
 Event Enable          : Event Messages Disabled
 Assertions Enabled    : unc-
--
Sensor ID              : CPU 2 PECI tics (0x7d)
 Entity ID             : 3.2
 Sensor Type (Analog)  : Temperature
 Sensor Reading        : -56 (+/- 0) unspecified
 Status                : ok
 Lower Non-Recoverable : -64.000
 Lower Critical        : na
 Lower Non-Critical    : 64.000
 Upper Non-Critical    : 0.000
 Upper Critical        : na
 Upper Non-Recoverable : na
 Event Status          : Event Messages Disabled
 Assertion Events      :
 Event Enable          : Event Messages Disabled
 Assertions Enabled    : unc-

2 things stick out to me. The lower non-critical threshold is higher than the upper non-critical threshold, so technically the Sensor Reading will always be out of range. But according to ipmitool sdr list | grep PECI, the reading seems to be fine:

CPU 1 PECI tics  | -49 unspecified   | ok
CPU 2 PECI tics  | -52 unspecified   | ok

But I do see that the only Assertion Enabled is unc(Upper Non-Critical). So I guess that means that it's ignoring the Lower Non-Critical threshold? The second thing that sticks out to me is that the Assertions Enabled section says unc- as opposed to unc+. On all other sensors, I only ever see lnr-, lcr-, lnc- and/or, unc+, ucr+, unr+. So although I couldn't find any documentation to support my assumption I assumed they could be interpreted as "assertions are enabled in the event that the reading goes below(-) a lower threshold and/or above(+) an upper threshold," as the case may be. However, wouldn't that mean that unc- would enable assertions when the reading was below(-) the upper threshold? If so, why is the reading of -55 which is below the Upper Non-Critical threshold of 0.000 ok?

I have a similar problem on a CentOS 4 IBM eServer xSeries 346 server. ipmitool -v sensor for the CPU1 VCore shows:

Sensor ID              : CPU 1 VCore (0xb8)
 Entity ID             : 3.1
 Sensor Type (Analog)  : Voltage
 Sensor Reading        : 1.330 (+/- 0) Volts
 Status                : ok
 Lower Non-Recoverable : 1.920
 Lower Critical        : na
 Lower Non-Critical    : 1.920
 Upper Non-Critical    : 0.000
 Upper Critical        : na
 Upper Non-Recoverable : na
 Assertion Events      :
 Assertions Enabled    :

So here the Upper Non-Critical threshold is lower than the Lower Non-Critical threshold, rendering any reading out of range. However, ipmitool sdr list shows:

CPU 1 VCore      | 1.32 Volts        | ok

In this case, however, there are no assertions enabled. Is that why the status is "ok"? What am I to make of these seemingly weird thresholds? Is there any documentation to explain what unc- could mean? Is it a problem specific to CentOS 4 IBMs or maybe these sensors in particular?

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.