Support APC Smart-UPS Ultra On-Line ambient.temperature via SNMP (PowerNet MIB v4.5.8)

[bbvrpz10k]

For APC Smart-UPS Ultra On-Line units with Network Management Card 3, the current SNMP driver in NUT does not expose ambient.temperature, even though the device provides this metric via SNMP.

The OID used by current APC Network Management Card 3 is available in APC PowerNet MIB v4.5.8 (latest from Schneider Electric ). The existing driver appears to be based on an older version (PowerNet MIB v4.0.4), and does not include the correct OID mapping for ambient temperature.

Expected behavior

When querying a UPS via NUT (e.g., upsc), the output should include:
ambient.temperature: 28
…where the value is sourced from the SNMP OID:
.1.3.6.1.4.1.318.1.1.1.1.1.1.0
This OID is defined in APC PowerNet MIB v4.5.8 and consistently returns the ambient temperature on APC Smart-UPS Ultra On-Line devices with Network Management Card 3.

Actual behavior

Currently ambient.temperature is missing in upsc output for APC Smart-UPS Ultra On-Line devices, even though the SNMP OID exists and returns a valid value.

Environment

• NUT version: latest stable (tested with 2.8.0 SNMP driver)
• UPS model: APC Smart-UPS Ultra On-Line
• Management Card: APC Network Management Card 3
• SNMP version: v3 (and v1/v2c can also return the OID)
• Relevant MIB: PowerNet MIB v4.5.8 (downloadable from Schneider Electric)

PowerNet MIB link:
https://www.se.com/us/en/download/document/APC_POWERNETMIB_EN/

Additional context

If needed, I can provide:

• Tested SNMP walk output for relevant OIDs
• If required, I can additionally validate the behavior using the previous candidate or pre-release version, in order to confirm expected behavior and exclude potential regressions prior to the update.

powernet458.mib.txt

[jimklimov]

I'm inclined to shelf it for a bit (got some dayjob debts) and say "PR welcome", this seems like an easy itch to scratch and we need more community members at least somehow versed in NUT sources and ways to build it :)

https://github.com/networkupstools/nut/wiki/Building-NUT-for-in%E2%80%90place-upgrades-or-non%E2%80%90disruptive-tests should get you started (note you can test the resulting driver from build workspace, without installing the build into your system).

.1.3.6.1.4.1.318.1.1.1.1.1.1.0

This is actually the base ID with the UPS model name. Temperature is elsewhere, re-read the MIB and check with snmpwalk or similar tools.

[bbvrpz10k]

Thanks for the clarification — you’re absolutely right, I mixed up the OID initially.
That was a typo on my side when composing the issue. The correct OID is:
1.3.6.1.4.1.318.1.1.25.1.2.1.6.1.1

I would be happy to work on updating or extending the subdriver, however I should be very clear about my current background: I have no prior experience at all with compiling software, building projects, or working directly with source code on *nix systems.

My professional development experience is almost entirely focused on C# development in Visual Studio on Windows, so the NUT build system, autotools/make-based workflows, and general *nix development practices are completely new to me.

While trying to validate the MIB locally, I ran into issues with snmptranslate. For example:
snmptranslate -m +PowerNet-MIB 1.3.6.1.4.1.318.1.1.25.1.2.1.6.1.1
results in:

Expected "}" (EOF): At line 84580 in /usr/share/snmp/mibs/APC-POWERNET-MIB.txt
: (is a reserved word): At line 84580 in /usr/share/snmp/mibs/APC-POWERNET-MIB.txt
SNMPv2-SMI::enterprises.318.1.1.25.1.2.1.6.1.1

So at the moment I can resolve the numeric OID, but the MIB parsing itself appears to fail with the current APC-POWERNET-MIB.txt as shipped.
For completeness, access to the UPS is done via SNMPv3 with authentication and privacy enabled, for example:

snmpwalk -v3 -l authPriv -u upsmon \
  -a SHA -A "pass222" \
  -x AES -X "pass111" \
  192.168.1.2 \
  1.3.6.1.4.1.318.1.1.25.1.2.1.6.1.1

The OID is definitely present and returns the external/ambient temperature value on APC Smart-UPS Ultra On-Line devices.

I understand the “PR welcome” position and I’m willing to learn and try, especially using the in-place build approach you linked.
If needed, I can also verify the behavior against a previous candidate or pre-release version to help document differences and reduce the risk of regressions.

Any guidance on:

• where this mapping would best live (existing APC subdriver vs a new one),
• or how temperature OIDs are typically normalized to ambient.temperature in NUT,
  would be very helpful as a starting point.

Thanks for your time and for maintaining NUT.

[jimklimov]

Hi, great to hear about enthusiasm, other issues can be overcome :)

Generally speaking, if you follow that Wiki article and the tools and build prerequisites page linked from it, you would get all the tools needed for a bumpless ride, I hope. I copy-paste package installation commands from those docs when preparing new persistent build agents, to make sure the contents remain relevant.

With this quest, source code structure being autotools-driven does not really matter: you just check out a git workspace, run a command or two, and NUT gets built. Then you test and install it.

You might even try your hand at building NUT for Windows semi-natively (with MSYS2 translation layer), although it remains generally experimental.

As for the change, primarily it would be about adding a line to drivers/apc-mib.c - there are several "ambient.temperature" variants near the end of file; whichever one responds first (IIRC, or was it last?) during the initialization walk wins as the data source for this piece of information for the driver run.

So this should be a one-line change for this particular feature. Other rituals include bumping a *_VERSION macro on top of this file (here APCC_MIB_VERSION) and adding a note to NEWS.adoc about the improvement - there are sections about different driver updates in the recent development cycle.

Build, test, and post a PR or repeat, as needed :D

[aarnold-sd68]

This also affects, APC Smart-UPS 1500. I'm assuming because they use the same mibs.