NUT package (2.8.0 and below)

dennypage

The ISSUE is my APC is not shutting down after pfsense and synology shuts off. It keeps running till battery becomes 0.

See post #1, version 2.7.4_6 which addresses this issue.

pablo_deskObar

@pablo_deskObar:

The ISSUE is my APC is not shutting down after pfsense and synology shuts off. It keeps running till battery becomes 0.

See post #1, version 2.7.4_6 which addresses this issue.

Didnt see that :) Also, will the ups power back on automatically after mains is restored after a power kill?

dennypage

@pablo_deskObar:

Also, will the ups power back on automatically after mains is restored after a power kill?

It should. Most all UPSs are designed to.

q54e3w

My pfsense server is running NUT package and I also have a separate ESXi host with NUT installed. I want to shut down the VMs and the ESXI host before pfsense shutdowns to ensure network services remain up to prevent problems closing down the VMs cleanly. How can I delay pfsense powering down until esxi has completed shutting down?
I have a APC SMX UPS attached via AP9631 network card.

dennypage

@q54e3w:

How can I delay pfsense powering down until esxi has completed shutting down?

The HOSTSYNC variable controls how long the master will wait for slaves to initiate their shutdown. You can find information on this in the ups.conf man page.

q54e3w

Thank you - I'll give it a try.

I've noticed the following errors appearing in my logs, any idea why this would be?


...
Feb 18 11:01:52	snmp-ups	44002	dstate_setflags: base variable (battery.runtime.low) is immutable
Feb 18 11:01:20	snmp-ups	44002	dstate_setflags: base variable (battery.runtime.low) is immutable
Feb 18 11:00:48	snmp-ups	44002	dstate_setflags: base variable (battery.runtime.low) is immutable
...

I have the following code

extra arguments


ignorelb
override.battery.charge.warning = 50
override.battery.charge.low = 30
override.battery.runtime.low = 1200

upsmon.conf


HOSTSYNC 600

upsd.conf


LISTEN 192.168.10.1

upsd.users


[remoteuser]
password = mypassword
upsmon slave

the settings appear to be being applied correctly….


upsc ups@localhost
ambient.1.humidity.alarm.high: 60.00
ambient.1.humidity.alarm.low: 30.00
ambient.1.temperature.alarm.high: 40.00
ambient.1.temperature.alarm.low: 10.00
ambient.humidity: 39.00
ambient.temperature: 15.0
battery.charge: 100.00
battery.charge.low: 30
battery.charge.warning: 50
battery.date: 11/15/2017
battery.packs: 1.00
battery.runtime: 4428.00
battery.runtime.low: 1200
battery.voltage: 54.20
device.mfr: APC
device.model: Smart-UPS X 1500
device.type: ups
driver.flag.ignorelb: enabled
driver.name: snmp-ups
driver.parameter.pollinterval: 2
driver.parameter.port: 192.168.10.50
driver.parameter.synchronous: no
driver.version: 2.7.4
driver.version.data: apcc MIB 1.2
driver.version.internal: 0.97
input.frequency: 60.00
input.sensitivity: high
input.transfer.high: 127
input.transfer.low: 106
input.transfer.reason: selfTest
input.voltage: 122.20
input.voltage.maximum: 123.20
input.voltage.minimum: 121.30
output.current: 4.80
output.frequency: 60.00
output.voltage: 122.20
output.voltage.nominal: 120
ups.delay.shutdown: 20
ups.delay.start: 10
ups.firmware: UPS 09.1 (ID20)
ups.id: UPS
ups.load: 48.30
ups.mfr: APC
ups.mfr.date: 06/30/2012
ups.model: Smart-UPS X 1500
ups.status: OL
ups.temperature: 18.60
ups.test.date: 02/14/2018
ups.test.result: Ok

dennypage

@q54e3w:

Thank you - I'll give it a try.

I've noticed the following errors appearing in my logs, any idea why this would be?


...
Feb 18 11:01:52	snmp-ups	44002	dstate_setflags: base variable (battery.runtime.low) is immutable
Feb 18 11:01:20	snmp-ups	44002	dstate_setflags: base variable (battery.runtime.low) is immutable
Feb 18 11:00:48	snmp-ups	44002	dstate_setflags: base variable (battery.runtime.low) is immutable
...

I believe it means that the variable is marked read-only in the SNMP MIB. For more information on this, I would suggest checking the nut users list.

FWIW, if it were me I would just remove the setting of runtime and use charge %. You don’t need both.

spambait

pfsense 2.4.2-RELEASE-p1 (amd64)
nut 2.7.4_5

Polling an APC AP9631 on the LAN via SNMP.

When I reboot the appliance, UPS status shows as down. Log has a bunch of these:

Feb 19 21:40:27 upsmon 66529 Poll UPS [office-UPS] failed - Driver not connected
Feb 19 21:42:22 upsmon 67668 Startup successful
Feb 19 21:42:23 upsd 68233 listening on ::1 port 3493
Feb 19 21:42:23 upsd 68233 listening on 127.0.0.1 port 3493
Feb 19 21:42:23 upsd 68233 Can't connect to UPS [office-UPS] (snmp-ups-office-UPS): No such file or directory
Feb 19 21:42:23 upsd 68498 Startup successful
Feb 19 21:42:24 snmp-ups 68710 Startup successful

Now when I go to restart the service via the GUI In Services -> UPS, it starts up perfectly right away.

apcupsd was installed but I removed it when switching to NUT.

I tried using snmp v1 instead of v3 thinking I screwed up my v3 credentials but that doesn't seem to be the issue. Even with snmpv1 it wont come up after a reboot until I manually intervene.

I tried some of the retry and delay parameters in the additional parameters for ups.conf but they also didn't help with this


maxretry=3
retrydelay=10
pollinterval=5
snmp_retries=10
snmp_timeout=2
pollfreq=10

dennypage

Please give the following a try:

Get rid of pollfreq and pollinterval.

Set these variables in the Extra Arguments to driver section:


snmp_timeout=2
snmp_retries=10

Set these variables in the Advanced section for ups.conf:


retrydelay=30
maxretry=20

Now reboot the box. When the box comes up, go to the UPS status page. [Do not attempt to start the service manually] If the UPS shows as needing attention, wait for the page to refresh automatically. If it doesn't auto refresh, after a couple minutes refresh the page manually. Does the UPS show as running? Or is it still down?

spambait

@dennypage:

Please give the following a try:

Get rid of pollfreq and pollinterval.

Set these variables in the Extra Arguments to driver section:
snmp_timeout=2
snmp_retries=10
Set these variables in the Advanced section for ups.conf:
retrydelay=30
maxretry=20
Now reboot the box. When the box comes up, go to the UPS status page. [Do not attempt to start the service manually] If the UPS shows as needing attention, wait for the page to refresh automatically. If it doesn't auto refresh, after a couple minutes refresh the page manually. Does the UPS show as running? Or is it still down?

These settings work perfectly!

I am now back on SNMPv3 and NUT is performing as expected, across reboots and all. Thank you!

dennypage

That's good news. Now that it's working, you can probably tune the variables down quite a bit. In particular, the settings in Extra Arguments can likely be returned to their default values of 1 & 5. Or simply deleted.

spambait

@dennypage:

That's good news. Now that it's working, you can probably tune the variables down quite a bit. In particular, the settings in Extra Arguments can likely be returned to their default values of 1 & 5. Or simply deleted.

I think actually found a bug (with the AP9631, fw 6.5.0).

I just deployed a second AP9631 out of the box. Factory reset, flashed 6.5.0, factory reset again, set up SNMPv3 and had the same issues I described before where NUT cannot communicate.

I have now noticed that if I have SNMPv1 DISABLED, NUT will fail to connect via SNMPv3 no matter what extra settings I provide. Simply turning on SNMPv1 in the AP9631 (fw 6.5.0) immediately lets them communicate. Very strange. Further testing is needed.

I have replicated this on the previous installation where I can break the SNMPv3 communication between NUT and the AP9631 by disabling SNMPv1, don't even need to reboot.

this_fixes_it.JPG_thumb

this_breaks_it.JPG_thumb

dennypage

@spambait:

I just deployed a second AP9631 out of the box. Factory reset, flashed 6.5.0, factory reset again, set up SNMPv3 and had the same issues I described before where NUT cannot communicate.

I have now noticed that if I have SNMPv1 DISABLED, NUT will fail to connect via SNMPv3 no matter what extra settings I provide. Simply turning on SNMPv1 in the AP9631 (fw 6.5.0) immediately lets them communicate. Very strange. Further testing is needed.

I have replicated this on the previous installation where I can break the SNMPv3 communication between NUT and the AP9631 by disabling SNMPv1, don't even need to reboot.

It sounds like SNMPv3 may not be properly configured/functioning in NUT. Can you post the contents of /usr/local/etc/nut/ups.conf please?

spambait

@dennypage:

It sounds like SNMPv3 may not be properly configured/functioning in NUT. Can you post the contents of /usr/local/etc/nut/ups.conf please?

I am going to find an SNMPv3 client and try replicate this in a moment. If a different v3 client fails in the same way the issue is with AP9631.

If it doesn't then there may indeed be a NUT issue. Maybe it has some routine to check SNMPv1 availability before anything else, even if only SNMPv3 is needed.

I realize the credentials are in there but it's a sandbox setup so this is not really a concern. My full, unedited /usr/local/etc/nut/ups.conf:

retrydelay=30
maxretry=20
snmp_version=v3
privProtocol=AES
authProtocol=SHA
secName=pfSense
authPassword=VtwHzSj6XQlmp9OFFa28NJcUL93qq5zZ4
privPassword=TpWsICXWm4CeKtwUehtvcBpDnBW0KZCC

[S10-AP9631]
driver=snmp-ups
port=172.16.10.5
snmp_timeout=2
snmp_retries=10

dennypage

SNMP parameters are part of the UPS specification. As such, they go in the "Extra Arguments to driver" rather than in the "Additional configuration lines for ups.conf" section. The advanced section is only for Global Directives (see ups.conf), and anything that is not recognized as a global directive is ignored by NUT. I know it's confusing, but it's how NUT's configuration works. Anything that comes from the driver man page (snmp-ups in your case) goes into Extra Arguments to driver.

spambait

@dennypage:

SNMP parameters are part of the UPS specification. As such, they go in the "Extra Arguments to driver" rather than in the "Additional configuration lines for ups.conf" section. The advanced section is only for Global Directives (see ups.conf), and anything that is not recognized as a global directive is ignored by NUT. I know it's confusing, but it's how NUT's configuration works. Anything that comes from the driver man page (snmp-ups in your case) goes into Extra Arguments to driver.

That results in


[S10-AP9631]
driver=snmp-ups
port=172.16.10.5
snmp_timeout=2
snmp_retries=10
snmp_version=v3
privProtocol=AES
authProtocol=SHA
secLevel=authPriv
secName=pfSense
authPassword=VtwHzSj6XQlmp9OFFa28NJcUL93qq5zZ4
privPassword=TpWsICXWm4CeKtwUehtvcBpDnBW0KZCC

Edit: So it seems it was using SNMPv1 all along due to my mis-configuration resulting in the SNMPv3 stuff being ignored. There is no bug. Now that I have the correct structure in ups.conf it is indeed trying to connect SNMPv3 but it is not authenticating. The AP9631's built in baby IDS keeps sending me notifications now that there are failed SNMPv3 authentications from the pfSense IP.

I will switch her back to SNMPv1 for now and troubleshoot my SNMPv3 setup with qtmib

dennypage

That makes sense. For what it's worth, most people use a read-only SNMPv1 group for UPS monitoring. There isn't any sensitive information in the stream.

spambait

@dennypage:

That makes sense. For what it's worth, most people use a read-only SNMPv1 group for UPS monitoring. There isn't any sensitive information in the stream.

The docu said a write community is needed to facilitate shutdowns. I wanted the UPS and NUT to coordinate that NUT will shut down pfSense when there's not much runtime left and then tell the UPS to shut off the outlet in x minutes and come back up when AC power is restored and battery has at least y % charge.

Otherwise pfSense will just shut down and stay down until someone presses the power button or the BIOS power on time is reached.

dennypage

@spambait:

The docu said a write community is needed to facilitate shutdowns. I wanted the UPS and NUT to coordinate that NUT will shut down pfSense when there's not much runtime left and then tell the UPS to shut off the outlet in x minutes and come back up when AC power is restored and battery has at least y % charge.

Ah, yes. Needing write permission to kill power makes sense. Note that the pfSense NUT package version 2.7.4_5 does not support issuing the final power kill order. This will come with NUT package 2.7.4_6 which requires pfSense 2.4.3.

xx_robot_xx

Hello,

How do you configure power-kill after shutdown? I have pfsense 2.4.3 and latest version of nut..