search

networkupstools / nut

The Network UPS Tools repository. UPS management protocol Informational RFC 9271 published by IETF at https://www.rfc-editor.org/info/rfc9271 Please star NUT on GitHub, this helps with sponsorships!
https://networkupstools.org/
Other
2.09k stars 351 forks source link

Add support for Liebert GXT2U w/ SNMP/web Card #416

Open autumnwalker opened 7 years ago

autumnwalker commented 7 years ago

Hi All,

I have acquired a Liebert GXT2U UPS with a "voyager" SNMP/web card installed (note, this is NOT the same as the is-webcard). Trying to configure NUT using snmp-ups driver and receive the following output:

Network UPS Tools - Generic SNMP UPS driver 0.72 (2.7.3)
No matching MIB found for sysOID '.1.3.6.1.4.1.476.1.1.1.14'!
Please report it to NUT developers, with an 'upsc' output for your device.
Going back to the classic MIB detection method.
No supported device detected

Any chance of adding support for that MIB? It's available online to download or I can run something locally (I have access to the card and can walk the device).

jimklimov commented 7 years ago

Hi. Generally, there are several approaches to such issues :)

One is that vendors should implement the standard generic IETF MIB for power devices (the "mibs=ietf" setting in the ups.conf section for the device, or equivalent "-x" flag while testing via command-line). If this card does expose that standard entry point, it may suffice at least for basic monitoring of power status, voltages, etc. - though nothing advanced and specific to the vendor and/or device capabilities, probably not management via upscmd/upsrw either.

The other option with currently standard NUT upstream/master is to copy and adapt one of the drivers/*-mib.c files to define a uniquely-named mapping table for the vendor MIB that you have, mention it in the list of structure references in snmp-ups.c, and recompile the driver. Each *-mib.c is just a C file with a large array of a structure to map SNMP OIDs to NUT variables.

For the latter approach, it would help you to download the vendor MIB to install into your net-snmp configuration directory, and snmpwalk the vendor entry point, so you'd have the commented results (what data is actually exposed and how it's named according to the MIB); also the comments usually present in the MIB text file are helpful to figure out the nuances.

Poking @aquette just in case he has more nuances to share (there was IIRC a helper script to automate some of this new-mapping work).

Unfortunately, the current NUT codebase requires to recompile the snmp-ups driver to incorporate any changes to such SNMP-NUT mappings. There is a DMF branch to support same mappings dynamically via loadable XML files, but the branch is large and review to integrate it into trunk has not yet been completed.

Also note that by your screenshots, your system uses NUT 2.7.3 that may miss quite a bit of improvements made in the current Git HEAD - so consider updating by compiling your own NUT (at least, if IETF does not suffice for you and you have to compile new driver anyway) :)

aquette commented 7 years ago

@autumnwalker the script mentioned by @jimklimov creates the base implementation files (xxx-mib.c/h) and makes it easier to implement and complete such files, as long as you have the MIB available to resolve numeric OIDs into textual (human friendly ones):

autumnwalker commented 7 years ago

Thanks guys! Looks like I am behind a version. I am trying to get this running on my unRAID box using a plugin that was developed there (https://github.com/Ambrotos/plugins/tree/master/nut). Perhaps I should take a step back and stand up vanilla NUT on a VM and get that working first.

I looked at the snmp-subdrivers documentation, but didn't dig into the actual usage where the plugin I am using doesn't have gen-snmp-subdriver.sh with it and I wasn't sure how to go about updating it specifically.

I'll fire up a plain Ubuntu VM with NUT tonight and see if I can get that to speak to my UPS. I will report back with progress (and questions).

aquette commented 7 years ago

note @autumnwalker that you can simply get the gen-snmp-subdriver.sh by using wget https://raw.githubusercontent.com/networkupstools/nut/master/scripts/subdriver/gen-snmp-subdriver.sh Keep in mind to put the needed MIB(s) in the same dir to have it working optimally

aquette commented 5 years ago

@autumnwalker any update on this? If possible, the minimum would be to provide a complete SNMP walk of the unit, so that I can process this and bootstrap a subdriver.

autumnwalker commented 5 years ago

Hi @aquette no update. I have to get back into the card via serial and reset things. I'm unable to access it now as I forget which IPs were configured to have access! I'll dig out my serial cable and try and get to this shortly so I can walk it.

autumnwalker commented 5 years ago

Hey @aquette - admittedly I'm not familiar with SNMP. I loaded up the mib provided by Liebert and did a walk. Below are the results; is this what you are looking for?

lcUpsIdentManufacturer.0 Liebert lcUpsIdentModel.0 GXT2-1500RT120 lcUpsIdentSoftwareVersion.0 2.7.3 lcUpsIdentSpecific.0 .iso.org.dod.internet.private.enterprises.emerson.liebert Corp.liebertUps.luExtensions.luUPStationGxt lcUpsIdentFirmwareVersion.0 GXT2MR14 lcUpsIdentSerialNumber.0 0219100012AF051 lcUpsIdentManufactureDate.0 05JUL02 lcUpsBatTimeRemaining.0 43 lcUpsBatVoltage.0 54 lcUpsBatCapacity.0 51 lcUpsInputFrequency.0 60 lcUpsInputNumLines.0 1 lcUpsInputLine.1 1 lcUpsInputVoltage.1 120 lcUpsOutputFrequency.0 60 lcUpsOutputLoad.0 30 lcUpsOutputNumLines.0 1 lcUpsOutputLine.1 1 lcUpsOutputVoltage.1 120 lcUpsOutputCurrent.1 3 lcUpsOutputWatts.0 324 lcUpsInverterStatus.0 on(2) lcUpsAlarms.0 0 lcUpsTestBattery.0 unknown(1) lcUpsTestBatteryStatus.0 passed(2) lcUpsControlOutputOffDelay.0 0 lcUpsControlOutputOnDelay.0 0 lcUpsControlOutputOffTrapDelay.0 0 lcUpsControlOutputOnTrapDelay.0 0 lcUpsControlUnixShutdownDelay.0 0 lcUpsControlUnixShutdownTrapDelay.0 0 lcUpsControlCancelCommands.0 unknown(1) lcUpsControlRebootAgentDelay.0 0 lcUpsNominalOutputVoltage.0 120 lcUpsNominalInputVoltage.0 120 lcUpsNominalOutputFreq.0 60 lcUpsNominalInputFreq.0 60 lcUpsNominalOutputVaRating.0 1500 lcUpsNominalOutputWattsRating.0 1050 lcUpsOnBypass.0 no(3) lcUpsBypassFrequency.0 60 lcUpsBypassNumLines.0 1 lcUpsBypassLine.1 1 lcUpsBypassVoltage.1 120 lcUpsConfigType.0 online(2) lcUpsConfigBypassInstalled.0 yes(2) lcUpsConfigModuleCount.0 0 lcUpsConfigAudibleStatus.0 enabled(2) lcUpsConfigLowBattTime.0 2 lcUpsConfigAutoRestart.0 0 .1.3.6.1.4.1.476.1.1.3.1.1.0 00 00 01 01 01 02 00 00 00 00 .1.3.6.1.4.1.476.1.1.3.1.2.0 fe 00 00 a4 00 00 8c 00 c0 f8 00 e0 80 00 00 c0 00 e0 c0 00 00 ff 00 00 cf 00 00 00 00 00 00 00 00 00 00 00 f0 00 c0 ee 00 00 Request Failed: Get Response PDU received from 192.168.2.21 Error Indication in response: There is no such variable name in this mib. Errindex: 1

autumnwalker commented 5 years ago

Would it be easier to just see the MIB from Liebert? liebert-mib.zip

autumnwalker commented 5 years ago

I spun up a Ubuntu box and installed NUT via APT. I pulled down gen-snmp-subdriver.sh and had it generate a .mib using mode 1. Now I'm lost - I don't have /drivers or snmp-ups.c in order to continue.

For C-style integration, do not forget to:

Ultimately - do I have to pave / nuke and build NUT from source in order to do this?

autumnwalker commented 5 years ago

So, I found this article and followed the steps to get the dev files on my box. I followed the instructions there combined with the "C-style" integration instructions from above and I still get:

Network UPS Tools - UPS driver controller 2.7.4 Network UPS Tools - Generic SNMP UPS driver 1.12 (2.7.4-734-g1b217f46) No matching MIB found for sysOID '.1.3.6.1.4.1.476.1.1.1.14'! Please report it to NUT developers, with an 'upsc' output for your device. Going back to the classic MIB detection method. No supported device detected Driver failed to start (exit status=1)

aquette commented 5 years ago

hi @autumnwalker bad timing, notification received while moving on vacation :D could you please post a zip with the generated files. Using these files, and doing the integration mentioned by the script implies indeed to modify sources and recompile, which may not be trivial.

autumnwalker commented 5 years ago

Hi @aquette, no problem at all! I hope your move went well and you had a relaxing vacation!

Zip file is here with the .c and .h file generated by the subdriver script. FYI - I re-ran it since my last post so the names are different (NUT does not like hyphens in the name!).

liebert_voyager.zip

aquette commented 5 years ago

hi @autumnwalker finally got the few seconds to look at. thanks for the archive! question: how would you be able to test? Git (branch or master), binary, ...? I'd like some checks from you, since I don't have such HW underhand, despite the sample values in your provided dump.

autumnwalker commented 5 years ago

Hey @aquette! I can test however you'd like me to. I can spin up a VM and install however it would work best.

The target install is with the NUT plugin for Unraid https://forums.unraid.net/topic/60217-plugin-nut-v2-network-ups-tools/, but I have been doing my testing in a Ubuntu 18.04 VM with the Ubuntu NUT package + dev files. I can nuke / pave and pull from Git or go from binaries if you'd like.

aquette commented 5 years ago

hi @autumnwalker so, going the git path, faster for me.

I've made a 1rst stab, available on the liebert-voyager branch. To test it:

Few comments:

autumnwalker commented 5 years ago

Thanks @aquette!

I tried to build NUT from scratch using the source on Git; did the following:

Get the following error:

checking for pkg-config... no
./configure: line 5251: syntax error near unexpected token `dummy_PKG_CONFIG,'
./configure: line 5251: `         PKG_CHECK_MODULES(dummy_PKG_CONFIG, pkg-config,'

Doing an SNMP walk of ".1.3.6.1.2.1.33" returns nothing. Issued command:

snmpwalk -c <community> -v1 <IP> .1.3.6.1.2.1.33

My preliminary review of of your changes for drivers.list (and other files in compare) make sense. My only caveat is the "Voyager SNMP/web Card" name is the best that Liebert could determine based on the markings / serials on the card - so nobody is going to tell us otherwise I guess!

I'm going to try and install the Ubuntu package for NUT and install the liebert-voyager branch over it using the instructions on the Wiki - I'll see if that will allow me to get upsc for you.

clepple commented 5 years ago

@autumnwalker You got most of the build steps, but steps 1-3 on the wiki instructions grab all of the dependencies like pkg-config automatically.

aquette commented 4 years ago

@autumnwalker thanks for the preliminary feedback. Please follow @clepple advice on compiling.

Beside from this, I'm still puzzled by your feedback on the missing IETF MIB info (empty results of snmpwalk on .1.3.6.1.2.1.33) and the lack of other OIDs to feed ups.status (at least OL, OB, LB). The only thing I saw is the BYPASS status

Once good, or beside from this, could you make an additional test:

According to these results, we may have to move on step 2 more quickly, and implement alarms handling to get other status bits.

autumnwalker commented 4 years ago

Compiled via wiki and did a basic nut.conf (netserver) and ups.conf. Ran upsdrvctl start and got the following:

root@nut01:/etc/nut# upsdrvctl start Network UPS Tools - UPS driver controller 2.7.4 Network UPS Tools - Generic SNMP UPS driver 1.12 (2.7.4-736-gf9fa0f63) Detected GXT2-1500RT120 on host 192.168.2.21 (mib: liebert_voyager 0.1)

Ran upsd and got the following:

root@nut01:/etc/nut# upsd Network UPS Tools upsd 2.7.4 fopen /var/run/nut/upsd.pid: No such file or directory listening on 127.0.0.1 port 3493 listening on ::1 port 3493 Connected to UPS [GXT2U]: snmp-ups-GXT2U

Ran upsc and got the following:

root@nut01:/etc/nut# upsc GXT2U Init SSL without certificate database battery.charge: 19 battery.voltage: 54 device.firmware: GXT2MR14
device.firmware.aux: 2.7.3 device.mfr: Liebert device.model: GXT2-1500RT120
device.serial: 0219100012AF051
device.type: ups driver.name: snmp-ups driver.parameter.pollinterval: 2 driver.parameter.port: 192.168.2.21 driver.parameter.synchronous: no driver.version: 2.7.4-736-gf9fa0f63 driver.version.data: liebert_voyager MIB 0.1 driver.version.internal: 1.12 input.frequency: 60 input.frequency.nominal: 60 input.phases: 1 input.voltage: 122 input.voltage.nominal: 120 output.current: 3 output.frequency: 60 output.frequency.nominal: 60 output.phases: 1 output.realpower: 363 output.voltage: 120 output.voltage.nominal: 120 ups.load: 34 ups.power.nominal: 1500 ups.realpower.nominal: 1050 ups.status: ups.test.result: 2

I also tried snmpget -c -v1 .1.3.6.1.4.1.476.1.1.1.1.5.1.0 and got:

SNMPv2-SMI::enterprises.476.1.1.1.1.5.1.0 = INTEGER: 2

I will have to wait until I get home to unplug the UPS and run the snmpget command again.

Looking at "battery.charge: 19" - looks incorrect. Battery should be 100 unless I am misinterpreting the "19".

Huge progress. Thank you for all the effort @aquette and @clepple!

aquette commented 4 years ago

thx for the feedback @autumnwalker , indeed progressing ;) that also confirms partially some of the ups.status points. This will btw cause me some implementation troubles for the "OL" status at least. For the others, when doing your on battery test, could you please also snmpwalk on ".1.3.6.1.4.1.476.1.1.1.1.6" as for the battery and 19, you're indeed right, it's not the percentage but the runtime (in minutes). Fixed now (and published as number of seconds in battery.runtime) along with added the battery remaining charge (as battery.charge).

Remember to "git pull" from your NUT git clone, to get the 2 above fix ;)

autumnwalker commented 4 years ago

Thanks again @aquette. Pulled, recompiled. Looks like the data is correct!

root@nut01:~/nut/drivers# upsc GXT2U@192.168.2.12 Init SSL without certificate database battery.charge: 100 battery.runtime: 1140 battery.voltage: 54 device.firmware: GXT2MR14
device.firmware.aux: 2.7.3 device.mfr: Liebert device.model: GXT2-1500RT120
device.serial: 0219100012AF051
device.type: ups driver.name: snmp-ups driver.parameter.pollinterval: 2 driver.parameter.port: 192.168.2.21 driver.parameter.synchronous: no driver.version: 2.7.4-737-gde5c8642 driver.version.data: liebert_voyager MIB 0.1 driver.version.internal: 1.12 input.frequency: 60 input.frequency.nominal: 60 input.phases: 1 input.voltage: 121 input.voltage.nominal: 120 output.current: 3 output.frequency: 60 output.frequency.nominal: 60 output.phases: 1 output.realpower: 373 output.voltage: 120 output.voltage.nominal: 120 ups.load: 35 ups.power.nominal: 1500 ups.realpower.nominal: 1050 ups.status: ups.test.result: 2

I wasn't able to get to the UPS last night, but I'll try to run the onbatt tests / walks tonight.

autumnwalker commented 4 years ago

Ran the on battery tests. I ran them each without specifying the MIB and also with specifying (-m) the Liebert MIB. Interestingly, the first command returned the same result while powered and while on battery.

Powered:

root@nut01:~# snmpget -c [community] -v1 [ip] .1.3.6.1.4.1.476.1.1.1.1.5.1.0 SNMPv2-SMI::enterprises.476.1.1.1.1.5.1.0 = INTEGER: 2

root@nut01:~# snmpget -c [community] -v1 [ip] .1.3.6.1.4.1.476.1.1.1.1.5.1.0 -m LIEBERT-UPSTATION-GXT-UPS-MIB LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsInverterStatus.0 = INTEGER: on(2)

root@nut01:~# snmpwalk -c [community] -v1 [ip] .1.3.6.1.4.1.476.1.1.1.1.6 SNMPv2-SMI::enterprises.476.1.1.1.1.6.1.0 = Gauge32: 0

root@nut01:~# snmpwalk -c [community] -v1 [ip] .1.3.6.1.4.1.476.1.1.1.1.6 -m LIEBERT-UPSTATION-GXT-UPS-MIB LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarms.0 = Gauge32: 0

On Battery:

root@nut01:/etc/snmp# snmpget -c [community] -v1 [ip] .1.3.6.1.4.1.476.1.1.1.1.5.1.0 SNMPv2-SMI::enterprises.476.1.1.1.1.5.1.0 = INTEGER: 2

root@nut01:/etc/snmp# snmpget -c [community] -v1 [ip] .1.3.6.1.4.1.476.1.1.1.1.5.1.0 -m LIEBERT-UPSTATION-GXT-UPS-MIB LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsInverterStatus.0 = INTEGER: on(2)

root@nut01:/etc/snmp# snmpwalk -c [community] -v1 [ip] .1.3.6.1.4.1.476.1.1.1.1.6 SNMPv2-SMI::enterprises.476.1.1.1.1.6.1.0 = Gauge32: 4 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.1.3 = INTEGER: 3 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.1.10 = INTEGER: 10 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.1.31 = INTEGER: 31 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.1.35 = INTEGER: 35 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.2.3 = OID: SNMPv2-SMI::enterprises.476.1.1.1.1.6.3.3 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.2.10 = OID: SNMPv2-SMI::enterprises.476.1.1.1.1.6.3.10 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.2.31 = OID: SNMPv2-SMI::enterprises.476.1.1.1.1.6.3.13 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.2.35 = OID: SNMPv2-SMI::enterprises.476.1.1.1.1.6.3.17 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.3.3 = Timeticks: (644584600) 74 days, 14:30:46.00 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.3.10 = Timeticks: (644584615) 74 days, 14:30:46.15 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.3.31 = Timeticks: (644584640) 74 days, 14:30:46.40 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.3.35 = Timeticks: (644584655) 74 days, 14:30:46.55

root@nut01:/etc/snmp# snmpwalk -c [community] -v1 [ip] .1.3.6.1.4.1.476.1.1.1.1.6 -m LIEBERT-UPSTATION-GXT-UPS-MIB LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarms.0 = Gauge32: 4 LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmId.3 = INTEGER: 3 LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmId.10 = INTEGER: 10 LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmId.31 = INTEGER: 31 LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmId.35 = INTEGER: 35 LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmDescr.3 = OID: LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmUtilFailed LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmDescr.10 = OID: LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmOnBattery LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmDescr.31 = OID: LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmInputFreqError LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmDescr.35 = OID: LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsBadBypassPower LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmTime.3 = Timeticks: (644584600) 74 days, 14:30:46.00 LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmTime.10 = Timeticks: (644584615) 74 days, 14:30:46.15 LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmTime.31 = Timeticks: (644584640) 74 days, 14:30:46.40 LIEBERT-UPSTATION-GXT-UPS-MIB::lcUpsAlarmTime.35 = Timeticks: (644584655) 74 days, 14:30:46.55

autumnwalker commented 4 years ago

Hey @aquette, anything else you need from me here? I'm unclear on next steps. Thanks!

niallm90 commented 4 years ago

I also have an older Liebert GTX with the SNMP web card and can confirm this works. Would be keen to see this merged, can I be of any help?

autumnwalker commented 4 years ago

@niallm90 this isn't ready for PROD yet. NUT still does not understand the status of the UPS (i.e. on battery / on mains).

niallm90 commented 4 years ago

Oh, face palm. I'll see if I can get any info form my unit today to help. Unfortunate my batteries are shot so I may not be able to even run the UPS off them with no load.

niallm90 commented 4 years ago

Took me a while to get some new batteries sorted but I've finally been able to and have a bit of a hack around. As far as I can tell their is no nice way to get the status with this UPS. I think I have worked out the logic to get the mode its in, however it doesn’t match the way any of the other SNMP sub-drivers work.

As far as I can tell:

I'm keen to see support for this UPS added and let me know if there is anything I can do to speed this up.

autumnwalker commented 3 years ago

Just re-racked my lab (moved houses) and rigged up my GXT2U again and thought about revisiting this. @niallm90 thank you for putting together that logic. It makes sense to me and I can see those same parameters from mine.

@aquette can you help us build the logic into the driver? Is that even possible?

aquette commented 3 years ago

I'll try to take a look tomorrow, but will need a lot of reactivity from your side to test. Possibly attach snmpwalk for both OB and OL so that I can snmpsimd on these

autumnwalker commented 3 years ago

Thank you @aquette! Let me know what to do and I will run it here.

I can SNMP walk the conditions that @niallm90 indicated for OL and OB. All right now almost all return no response as my UPS is on mains now, not battery. lcUpsInverter shows on(2).

snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.5
SNMPv2-SMI::enterprises.476.1.1.1.1.5.1.0 = INTEGER: 2

Are you trying to build the mib for simulation? The MIB I'm using is here: http://www.oidview.com/mibs/476/LIEBERT-UPS-MIB.html

aquette commented 3 years ago

@autumnwalker would you be able to test from git, using the branch liebert-voyager? please also send in the results of: snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.10 snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.1 snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.13.1 snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.12

autumnwalker commented 3 years ago

Running on latest code from branch liebert-voyager.

Running on mains: snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.10 - no value snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.1 - no value snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.13.1 - SNMPv2-SMI::enterprises.476.1.1.1.1.13.1.0 = INTEGER: 3 snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.12 - no value

I will run on battery shortly and report back.

autumnwalker commented 3 years ago

On battery: snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.10 - no value snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.1 - no value snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.13.1 - SNMPv2-SMI::enterprises.476.1.1.1.1.13.1.0 = INTEGER: 3 snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.12 - no value

Same results.

I also tried: snmpwalk -c [community] -v1 [ip] .1.3.6.1.4.1.476.1.1.1.1.6
SNMPv2-SMI::enterprises.476.1.1.1.1.6.1.0 = Gauge32: 4 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.1.3 = INTEGER: 3 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.1.10 = INTEGER: 10 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.1.31 = INTEGER: 31 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.1.35 = INTEGER: 35 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.2.3 = OID: SNMPv2-SMI::enterprises.476.1.1.1.1.6.3.3 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.2.10 = OID: SNMPv2-SMI::enterprises.476.1.1.1.1.6.3.10 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.2.31 = OID: SNMPv2-SMI::enterprises.476.1.1.1.1.6.3.13 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.2.35 = OID: SNMPv2-SMI::enterprises.476.1.1.1.1.6.3.17 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.3.3 = Timeticks: (39551340) 4 days, 13:51:53.40 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.3.10 = Timeticks: (39551360) 4 days, 13:51:53.60 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.3.31 = Timeticks: (39551385) 4 days, 13:51:53.85 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.3.35 = Timeticks: (39551400) 4 days, 13:51:54.00

aquette commented 3 years ago

I've just pushed some mods, blind mode though (not even had time to try compiling) plz try and report back, thx

autumnwalker commented 3 years ago

Not seeing a change in the branch.

aquette commented 3 years ago

Pushed on my personal repos

aquette commented 3 years ago

now fixed and pushed again on nut repos: https://github.com/networkupstools/nut/tree/liebert-voyager

autumnwalker commented 3 years ago

Thank you @aquette!

Ran git pull Ran ./autogen.sh Ran ./configure --with-user=ups --with-group=ups --with-snmp Ran systemctl restart nut-server

Running on mains: snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.10 - no value snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.1 - no value snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.13.1 - SNMPv2-SMI::enterprises.476.1.1.1.1.13.1.0 = INTEGER: 3 snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.12 - no value

Running on battery: snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.10 - no value snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.1 - no value snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.13.1 - SNMPv2-SMI::enterprises.476.1.1.1.1.13.1.0 = INTEGER: 3 snmpwalk -c [community] -v1 [IP] .1.3.6.1.4.1.476.1.1.1.1.6.3.12 - no value

I also tried: snmpwalk -c [community] -v1 [ip] .1.3.6.1.4.1.476.1.1.1.1.6 SNMPv2-SMI::enterprises.476.1.1.1.1.6.1.0 = Gauge32: 4 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.1.3 = INTEGER: 3 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.1.10 = INTEGER: 10 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.1.31 = INTEGER: 31 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.1.35 = INTEGER: 35 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.2.3 = OID: SNMPv2-SMI::enterprises.476.1.1.1.1.6.3.3 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.2.10 = OID: SNMPv2-SMI::enterprises.476.1.1.1.1.6.3.10 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.2.31 = OID: SNMPv2-SMI::enterprises.476.1.1.1.1.6.3.13 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.2.35 = OID: SNMPv2-SMI::enterprises.476.1.1.1.1.6.3.17 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.3.3 = Timeticks: (48592060) 5 days, 14:58:40.60 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.3.10 = Timeticks: (48592075) 5 days, 14:58:40.75 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.3.31 = Timeticks: (48592100) 5 days, 14:58:41.00 SNMPv2-SMI::enterprises.476.1.1.1.1.6.2.1.3.35 = Timeticks: (48592115) 5 days, 14:58:41.15

I wonder if I have the right code compiled / reboot / etc. properly? Results are all the same.

autumnwalker commented 3 years ago

It just occurred to me that doing the walks means nothing with the updated code.

Ran upsc and ups.status is blank.

upsc GXT2U@192.168.2.12 Init SSL without certificate database battery.charge: 51 battery.runtime: 11280 battery.voltage: 53 device.firmware: GXT2MR14
device.firmware.aux: 2.7.3 device.mfr: Liebert device.model: GXT2-1500RT120
device.serial: 0219100012AF051
device.type: ups driver.name: snmp-ups driver.parameter.pollinterval: 2 driver.parameter.port: 192.168.2.21 driver.parameter.synchronous: no driver.version: 2.7.4-737-gde5c8642 driver.version.data: liebert_voyager MIB 0.1 driver.version.internal: 1.12 input.frequency: 60 input.frequency.nominal: 60 input.phases: 1 input.voltage: 113 input.voltage.nominal: 120 output.current: 4 output.frequency: 60 output.frequency.nominal: 60 output.phases: 1 output.realpower: 490 output.voltage: 120 output.voltage.nominal: 120 ups.load: 46 ups.power.nominal: 1500 ups.realpower.nominal: 1050 ups.status: ups.test.result: 2

clepple commented 3 years ago

@autumnwalker commit de5c8642 (from the end of driver.version, after 2.7.4-##-g) is from 2019 - you likely want a866637578 (from Dec 17, 2020) or better yet, 487c667c43 (latest on liebert-voyager branch, from Dec 18, 2020).

If you have pulled the branch with those commits, you'll need to rebuild and reinstall the snmp-ups driver (make in drivers/ then make install as root), then restart the driver (upsdrvctl restart or equivalent init/systemctl incantation; probably systemctl restart nut-server?).

aquette commented 3 years ago

Note to @clepple the data dump mode of the drivers is very convenient for such case (in tree test): ./drivers/snmp-ups -s test -d2 -x port= -x <other params like community with each -x>

@autumnwalker I'm also interested in debug output level 3, so please add -DDD

autumnwalker commented 3 years ago

@clepple - I ran make and make install as root in drivers/ and restarted - upsc is reporting the same driver.version. I can see that files are updated today though (e.g. snmp_ups-liebert-voyager-mib.o). I can also see that liebert-voyager-mib.c was updated on the 18th (when I did git pull).

@aquette apologies. Where do you want me to add -DDD?

I might need some ELI5 here.

clepple commented 3 years ago

@autumnwalker my mistake, I guess we are missing a dependency to automatically update the version when you run make from drivers (I am usually building from one directory up).

If I understand Arno's comment, he's saying to stop the rest of the NUT machinery (systemctl stop nut-server and possibly systemctl stop nut-client) and just test the driver without installing, using the standalone command he mentioned. (The systemd service scripts tell upsdrvctl to start the driver in a similar fashion, but for that, the driver needs to be installed to the same locations as the .debs. This new test mechanism sidesteps that, and lets you run the driver straight from where it is built with make, and doesn't require upsd. On the other hand, without upsd, you can't use upsc, but apparently the test mode prints similar output to upsc.) Something like this:

path/to/nut/drivers/snmp-ups -s test -d2 -DDD -x port=<ip> -x community=<community>

clepple commented 3 years ago

@aquette while I was mentally walking through the dependency chain for fixing driver.version in this case (#938), it occurred to me that you could also just bump the version in the sub-driver when you make a change.

autumnwalker commented 3 years ago

Running as root.

root@nut01:~/nut# ./drivers/snmp-ups -s test -d2 -DDD -x port=192.168.2.12 -x community=private Network UPS Tools - Generic SNMP UPS driver 1.12 (2.7.4-737-gde5c8642) 0.000000 [D1] debug level is '3' 0.000982 Can't chdir to /var/run/nut: Permission denied

Even here - it seems to be running the de5c864 driver version. When I ran git pull from the nut directory (liebert-voyager branch) shouldn't that have grabbed the updated driver?

jimklimov commented 3 years ago

Are you sure git also successfully checked out that new branch?

On Mon, Dec 21, 2020, 01:09 autumnwalker notifications@github.com wrote:

Running as root.

root@nut01:~/nut# ./drivers/snmp-ups -s test -d2 -DDD -x port=192.168.2.12 -x community=private Network UPS Tools - Generic SNMP UPS driver 1.12 (2.7.4-737-gde5c8642) 0.000000 [D1] debug level is '3' 0.000982 Can't chdir to /var/run/nut: Permission denied

Even here - it seems to be running the de5c864 https://github.com/networkupstools/nut/commit/de5c8642590d7e991b866eef42bcbd0aeaf0a9dd driver version. When I ran git pull from the nut directory (liebert-voyager branch) shouldn't that have grabbed the updated driver?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/networkupstools/nut/issues/416#issuecomment-748692358, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAMPTFENGWZDC6TTTZMVWB3SV2G45ANCNFSM4DGSVWKQ .

autumnwalker commented 3 years ago

@jimklimov

root@nut01:~/nut# git branch

root@nut01:~/nut# git status On branch liebert-voyager Your branch is up to date with 'origin/liebert-voyager'.

Changes not staged for commit: (use "git add ..." to update what will be committed) (use "git checkout -- ..." to discard changes in working directory)

    modified:   configure.ac

no changes added to commit (use "git add" and/or "git commit -a")

clepple commented 3 years ago

Even here - it seems to be running the de5c864 driver version. When I ran git pull from the nut directory (liebert-voyager branch) shouldn't that have grabbed the updated driver?

@autumnwalker It sounds like you might have the latest (you can confirm with git show, which prints the git version hash for the source), but the build rule to update the driver.version string is incomplete. A make clean from the top nut directory would fix it on the next make, but that is overkill - hence the new issue (#938).

aquette commented 3 years ago

Try ./drivers/snmp-ups -s test -d2 -x root -DDD -x port=192.168.2.12 -x community=private Prefer the zip the result, more readable ;)

clepple commented 3 years ago

Maybe -u root instead of -x root?

I thought the pathnames and usernames specified in the wiki should have taken care of any permissions issues, especially when starting the driver as root.