NUT package (2.8.0 and below)
-
I'm seeing this in my logs
php-fpm 40655 /rc.newwanip: pfSense package system has detected an IP change or dynamic WAN reconnection - 82.127.34.254 -> 82.127.34.254 - Restarting packages.I was asling yself : what changed ? Why was this line triggered ?
For me, "82.127.34.254" is identical to "82.127.34.254".My WAN is using an RFC 1918 IP (DHCP using an upstream router) - the real WAN IP - 82.127.34.254 - doesn't change, but still, packages get restarted.
For longtime I really thought this was completely unnecessary, nut now I knows that that processes like unbound or openvpn don't like the fact that an interface to which they are bound, go up and down, even when the IP on that interface stays the same. -
@dennypage said in NUT package:
@kevindd992002 I don't believe you can prevent this from occurring. Problem is, pfSense doesn't know what each package is doing with the various interfaces or if they support dynamic discovery of interface changes (most don't). The only way for pfSense to ensure everything is functioning correctly is to restart the packages.
I see. Do you happen to know why the interface IP changes for no reason, in the first place?
-
@kevindd992002 In your case it looks like a change of state for OpenVPN interfaces. I'm sorry I don't know much beyond this. If you want to explore more, I'd suggest asking in the General pfSense Questions or OpenVPN groups.
-
Can I use the 2 USB ports on my XG-7100 1U appliance to monitor my APC Smart-UPS 1500 with this package?
-
@bazzacad If the ups has a usb interface, I expect nut with the usbhid driver would work. You can always check the nut hardware compatibility matrix to confirm. Have you tried it and encountered a problem?
-
@dennypage said in NUT package:
@kevindd992002 In your case it looks like a change of state for OpenVPN interfaces. I'm sorry I don't know much beyond this. If you want to explore more, I'd suggest asking in the General pfSense Questions or OpenVPN groups.
Ok, thanks.
I also have an error with my slave client but I'm not sure if it's OT here or not. Basically, my slave monitor client is a Debian box and it works just fine connecting to this pfsense master. But during boot time I get the same exact error described here. So what does the nut-driver.service do? Since the nut package here only runs as monitor (not server), do I need that service active?
-
@kevindd992002 I'm an RC guy. I don't use systemd.
That said, the link you posted is about a problem with usbhid-ups not being able to find the local ups on the serial bus. This is unrelated to use of upsmon to remotely monitor a ups.
If you are seeing anything from upsdrvctl, it means that the system is misconfigured. In a remote monitor case, only upsmon would be started. The local services, upsdrvctl (usbhid-ups) and upsd are not used and should not be started.
Please don't ask me how to configure systemd. I simply don't know.
-
@dennypage said in NUT package:
@kevindd992002 I'm an RC guy. I don't use systemd.
That said, the link you posted is about a problem with usbhid-ups not being able to find the local ups on the serial bus. This is unrelated to use of upsmon to remotely monitor a ups.
If you are seeing anything from upsdrvctl, it means that the system is misconfigured. In a remote monitor case, only upsmon would be started. The local services, upsdrvctl (usbhid-ups) and upsd are not used and should not be started.
Please don't ask me how to configure systemd. I simply don't know.
Lol, ok. Yeah, that's what I figured. I'm only using the nut package their as a monitor so I don't think the other services should be started at all.
-
@dennypage Everything worked perfectly with my XG-7100 1U and APC Smart-UPS 1500 using the usbhid driver. Is the package able to monitor 2 UPSs? Both APC Smart-UPS 1500. I didn't try that yet?
-
@bazzacad You mean a pfSense system with redundant power supplies, each with a separate ups? NUT itself is capable of this, but the NUT package does not provide any configuration support for it.
-
Just wanted to say thanks for this package. I am now using pfSense to monitor my CyberPower 1500 over USB, while my two unraid servers are network slave clients. Worked great after adding the NAT rule in the OP :-)
-
@cybrnook you are most welcome.
It’s really nice when people say thank you.
-
@dennypage said in NUT package:
@cybrnook you are most welcome.
It’s really nice when people say thank you.
Absolutely! It's my pleasure. Now just to figure out how to add multiple source IP's to the NAT rule, since I have 3 x clients I want to be able to phone home. Maybe it's just as simple as creating one NAT rule per IP :-) And see if you baked in "monuser" for Synology by default :-) since that will be my third.
EDIT: Synology was no problem. Just had to add monuser to the upsd.users with secret, and rename the UPS to "ups" in pfsense. Redid my Unraid systems with the same info, and it's all back online
-
Hi @dennypage - a big thank you from me as well! I came across a blog post today that detailed installing the NUT package on pfSense. Looking at the NUT compatibility list I saw that the the CyberPower UPS that I am using is actually supported. So I went ahead and hooked up the USB connection from the UPS to the firewall, installed the NUT package, and got everything setup and working inside pfSense. After that I also configured my Proxmox box and Synology NAS as slaves (thanks to everyone for all the great tips and advice in this thread!). Everything is working great so far. I went over two years with this same hardware not realizing that I could actually setup and utilize NUT -- very happy I found about the project today and the pfSense package.
I do have one basic quick question for everyone: When setting up NUT (e.g. slaves and on the master) how is the UPS critical battery threshold or shutdown time to determined (i.e. when connected devices are given the signal to shut down)? If this parameter adjustable in the configuration files (e.g. on the master) or is there a default value that's used? In other words, with the default out of the box NUT setup and configuration, will connected devices be given plenty of time to shut down? Thanks in advance for your help, I really appreciate it.
-
@tman222 By default, the level at which shutdown will occur is under control of the ups. The ups sends a "low battery" signal, and nut initiates a shutdown.
Generally the low battery signal is sufficient for single host systems, however if you find the ups runs out of battery before shutdown completes you can override the behavior. See the nut variables "battery.charge.low"
and "battery.runtime.low".Important note... some slaves, particularly Synology, take a long time to shutdown. After changing the runtime values, you will want to increase the "HOSTSYNC" variable. The default of 15 is far too short to allow a Synology to complete a shutdown.
Here is what I use on my server:
[Extra Arguments to driver]
override.battery.charge.low = 20
override.battery.runtime.low = 300[Additional configuration lines for upsmon.conf]
HOSTSYNC 120 -
@dennypage - thanks a bunch for this additional information. I was able to find a reference to the variables you mentioned and read up a bit:
https://networkupstools.org/docs/user-manual.chunked/apcs01.html
I had a one quick follow up question: Regarding the two variables you overrode (battery.charge.low and battery.runtime.low): Is it whichever of the two occurs first that then triggers the LB state and causes NUT to issues the shutdown signal? That is to say, in your case, when the battery drops to 20% charge OR has 300 seconds of run time remaining, the LB state is triggered and NUT begins shutdown process begins?
Thanks again for the help and clarification, I really appreciate it.
-
@tman222 said in NUT package:
Regarding the two variables you overrode (battery.charge.low and battery.runtime.low): Is it whichever of the two occurs first that then triggers the LB state and causes NUT to issues the shutdown signal? That is to say, in your case, when the battery drops to 20% charge OR has 300 seconds of run time remaining, the LB state is triggered and NUT begins shutdown process begins?
Thanks again for the help and clarification, I really appreciate it.
Yes, you are correct. 20% or 300 seconds, whichever happens first.
You're welcome!
-
@dennypage said in NUT package:
@kevindd992002 said in NUT package:
Doing a quick test yielded a "done and passed" result for the last test result field. However, when I tried doing a deep test, the CLI returned an OK message but the UPS did nothing. Does that mean that the command is not supported by my UPS? I'm still receiving low battery message notifications in my email regarding this UPS.
Have to try it with your other UPS, but my guess would be that the UPS is declining to initiate the deep test (which usually includes a battery calibration) because it has a low battery condition. Usually a deep test will not initiate on anything other than a completely full battery.
Best suggestion that I could make is to completely power off the UPS and restart it from the ground up. If that doesn't fix it, I would say that it's likely that the UPS or battery is defective.
I have yet to test this with my other working UPS but with the problematic UPS I also get this:
So I get input voltage out of range even though the input is around 228V now, both in the UPS display and using a voltmeter. So what gives? It looks like the UPS is transferring to battery even though it shouldn't?
-
@dennypage said in NUT package:
@kevindd992002 In your case it looks like a change of state for OpenVPN interfaces. I'm sorry I don't know much beyond this. If you want to explore more, I'd suggest asking in the General pfSense Questions or OpenVPN groups.
I'm not sure if the upsmon parent process died error was the one causing this, but my Synology NUT slave did not shutdown properly when there was a blackout earlier. I was not on-site so I'm just depending on the logs and email notifications I get. First of all, I got pfsense email notifications of the following:
- ups on battery -> Sat, 15 Jun 2019 12:05:24 +0800
- ups on line power -> Sat, 15 Jun 2019 12:17:22 +0800
- Bootup complete -> 13:14:05
And here are the pfsense ups logs:
Jun 14 18:55:20 upsmon 40927 UPS ups on battery Jun 14 19:07:23 upsmon 40927 UPS ups on line power Jun 14 19:07:33 upsmon 40927 UPS ups on battery Jun 14 19:10:15 upsmon 40927 UPS ups battery is low Jun 14 19:10:15 upsd 58260 Client local-monitor@::1 set FSD on UPS [ups] Jun 14 20:10:29 kernel ugen0.2: <American Power Conversion Back-UPS RS 1500G FW878.L4 .I USB FWL4> at usbus0 Jun 14 20:10:53 upsd 78154 Connected to UPS [ups]: usbhid-ups-ups Jun 14 20:10:56 upsd 24337 Connected to UPS [ups]: usbhid-ups-ups Jun 14 20:10:58 upsd 25161 User local-monitor@::1 logged into UPS [ups] Jun 14 20:11:17 upsd 25161 User local-monitor@::1 logged out from UPS [ups] Jun 14 20:11:18 upsd 31254 Connected to UPS [ups]: usbhid-ups-ups Jun 14 20:11:21 upsd 24874 Connected to UPS [ups]: usbhid-ups-ups Jun 14 20:11:23 upsd 25175 User local-monitor@::1 logged into UPS [ups] Jun 14 20:11:24 upsd 25175 User local-monitor@::1 logged out from UPS [ups] Jun 14 20:11:25 upsd 93544 Connected to UPS [ups]: usbhid-ups-ups Jun 14 20:11:28 upsd 84598 Connected to UPS [ups]: usbhid-ups-ups Jun 14 20:11:30 upsd 84882 User local-monitor@127.0.0.1 logged into UPS [ups] Jun 14 20:11:30 upsmon 55413 upsmon parent process died - shutdown impossible Jun 14 20:11:31 upsd 12323 Connected to UPS [ups]: usbhid-ups-ups Jun 14 20:11:33 upsd 13583 User local-monitor@::1 logged into UPS [ups] Jun 14 20:11:34 upsd 50344 Connected to UPS [ups]: usbhid-ups-ups Jun 14 20:11:36 upsd 52255 User local-monitor@::1 logged into UPS [ups] Jun 14 20:11:37 upsd 52255 User local-monitor@::1 logged out from UPS [ups] Jun 14 20:11:39 upsd 65039 Connected to UPS [ups]: usbhid-ups-ups Jun 14 20:11:40 upsd 65095 User local-monitor@::1 logged into UPS [ups] Jun 14 20:11:42 upsd 75747 Connected to UPS [ups]: usbhid-ups-ups Jun 14 20:11:44 upsd 75891 User local-monitor@::1 logged into UPS [ups] Jun 14 20:26:07 upsmon 73025 UPS ups on battery Jun 14 20:26:17 upsmon 73025 UPS ups on line power Jun 15 13:13:40 kernel ugen0.2: <American Power Conversion Back-UPS RS 1500G FW878.L4 .I USB FWL4> at usbus0 Jun 15 13:14:06 upsd 70280 Connected to UPS [ups]: usbhid-ups-ups Jun 15 13:14:10 upsd 99152 Connected to UPS [ups]: usbhid-ups-ups Jun 15 13:14:12 upsd 99790 User local-monitor@::1 logged into UPS [ups] Jun 15 13:14:26 upsd 99790 User local-monitor@::1 logged out from UPS [ups] Jun 15 13:14:28 upsd 12824 Connected to UPS [ups]: usbhid-ups-ups Jun 15 13:14:30 upsd 12827 User local-monitor@::1 logged into UPS [ups] Jun 15 13:14:31 upsd 24418 Connected to UPS [ups]: usbhid-ups-ups Jun 15 13:14:33 upsd 24437 User local-monitor@::1 logged into UPS [ups] Jun 15 13:14:54 upsd 24437 User local-monitor@::1 logged out from UPS [ups] Jun 15 13:14:56 upsd 74194 Connected to UPS [ups]: usbhid-ups-ups Jun 15 13:14:58 upsd 74325 User local-monitor@::1 logged into UPS [ups] Jun 15 13:14:59 upsd 90516 Connected to UPS [ups]: usbhid-ups-ups Jun 15 13:15:01 upsd 90761 User local-monitor@::1 logged into UPS [ups]As you can see, there are no logs showing that the ups changed to battery power and that initiated a shutdown. I'm assuming that the battery just drained out and did not notify any clients to gracefully shutdown, hence the improper shutdown message I got from my Synology UPS.
Any ideas regarding this?
Also, is it normal that I can't find any logs for the monuser client (the Synology NAS) connecting and disconnecting to the pfsense NUT master? I see logs like that when a Windows machine NUT service is connecting to pfsense master but not when a Synology NAS does it. Don't get me wrong though, the Synology NAS connects to it just fine and shuts down properly until today, but I never got the monuser logs.
-
@kevindd992002 said in NUT package:
I have yet to test this with my other working UPS but with the problematic UPS I also get this:
So I get input voltage out of range even though the input is around 228V now, both in the UPS display and using a voltmeter. So what gives? It looks like the UPS is transferring to battery even though it shouldn't?
This in and of itself isn't problematic. Many of the UPSs I've worked with save the last reason state in perpetuity. This just means that it happened sometime in the past.
