NUT package (2.8.0 and below)

[stevel]

Hi

I have recently upgraded to 2.3 - specifically 2.3.2-RELEASE-p1 (i386) - on my Alix device. In the process NUT was upgraded to 2.7.4_3. On visiting the NUT service page I was asked to confirm the upgrade settings. The UPS Type (local serial) and Driver (apcsmart) had been preserved so I didn't think further and hit 'save'. NUT didn't work despite restart. I believe I know why as I have successfully worked around…

By hitting 'save' I overwrote the correct serial port setting with one of the ones presented by the drop-down on the serial port field. The issue seems to be this... the serial port only allowed me to choose from the two built in UARTS on the board /dev/cuau[0-1]. I have a USB-to-serial adapter and this assigns a character device called /dev/cuaU0 (note U/C 'U'). Once I had SSH'd into the box and spotted the issue, editing the conf file and restarting NUT cause everything to spring back into life.

I should also say that I am new to this sort of meddling with conf files. I think these are not really the masters as far as configuration is concerned, i.e. they are auto-generated from elsewhere. i can see, for example that my entries in the other ups*.conf files are not reflected in the NUT UI. Should I be doing this a different way?

BTW, thanks for all the efforts porting this service, I hope this feedback proves useful.

Regards

• Steve

[dennypage]

Sorry about the problem Steve. Best that I can tell, in FreeBSD 10.3 and above call-out serial ports are all supposed to match the pattern "/dev/cuau?" which the device you have doesn't. Is the driver for this adapter included in the FreeBSD distribution or is it something that you had to install separately from the vendor? Can you tell me the name of the driver?

I don't have an issue adding support for uppercase 'U', but I would like to make I understand what's going on first. Thanks.

@stevel:

The issue seems to be this… the serial port only allowed me to choose from the two built in UARTS on the board /dev/cuau[0-1]. I have a USB-to-serial adapter and this assigns a character device called /dev/cuaU0 (note U/C 'U').

[stevel]

Hi,

The driver is called "uchcom.ko".

Cheers

• Steve

[bulldog5]

Dennypage,

I've seen an issue going on where the NUT service gets reset if the WAN connection is lost causing UPS communication loss.  My current setup has my ATT modem in a different area not on UPS backup, so when the power flickers the modem resets.  This is not a permanent setup, eventually the modem will be relocated inside the main rack which it will then be powered by the same UPS as the pfsense box.

However, if my WAN link goes down, why should the NUT service be affected?  If an extended power outage was to happen currently keeping my WAN link down, the pfsense box would never shutdown because its now lost communication with the UPS.

[dennypage]

Hey Bulldog. pfSense restarts services when the WAN interface disconnects/reconnects. Yes, this is unnecessary for some services such as NUT, but there is no way for pfSense to know which services actually need to be restarted.

What you would expect to see is NUT restart once when the interface goes down, and once again when the interface comes back up. Is this what you are seeing? Or are you seeing the NUT service stopped the entire time the interface is down?

[bulldog5]

@dennypage:

Hey Bulldog. pfSense restarts services when the WAN interface disconnects/reconnects. Yes, this is unnecessary for some services such as NUT, but there is no way for pfSense to know which services actually need to be restarted.

What you would expect to see is NUT restart once when the interface goes down, and once again when the interface comes back up. Is this what you are seeing? Or are you seeing the NUT service stopped the entire time the interface is down?

I will need to verify if it stays dead the entire time.

[dennypage]

Steve, I've pushed a fix for this. The release number will be 2.7.4_4. It may take a few days to work it's way through the system.

Thank you for your help in sorting this out. Much appreciated.

@stevel:

The issue seems to be this… the serial port only allowed me to choose from the two built in UARTS on the board /dev/cuau[0-1]. I have a USB-to-serial adapter and this assigns a character device called /dev/cuaU0 (note U/C 'U').

[dennypage]

Or, it could come out the same day it's pushed. :)

Steve, when you have a minute, can you confirm that this fixed the issue for you please?

Thanks.

@dennypage:

Steve, I've pushed a fix for this. The release number will be 2.7.4_4. It may take a few days to work it's way through the system.

[n3by]

hello,

I have an "NoName Polish brand" ups connected on Local USB that work with blazer driver in 2.3.2-RELEASE-p1 (amd64).

I get few emails every day with "UPS is unavailable" and "Communication with UPS lost" even I set cron to restart NUT every 9 minute.

Is it possible to edit and add NUT restart command to UPS monitoring module instead of sending this emails ?
Any idea where to look ?

Usually a manual restart of UPS monitoring daemon it fix the problem before cron but it is getting annoying to receive so many emails and entries in syslog from NUT… :o

thank you

edit:

I try now to automate restart ups service with:

upssched.conf

CMDSCRIPT /usr/local/bin/custom-upssched-cmd

PIPEFN /var/db/nut/upssched.pipe
LOCKFN /var/db/nut/upssched.lock

AT NOCOMM * START-TIMER ups-no-comm 20
AT COMMBAD * START-TIMER ups-comm-bad 20
AT COMMOK  * CANCEL-TIMER ups-comm-bad COMMOK

custom-upssched-cmd

#!/bin/sh
NAME=`basename $0`
LOGGER="/usr/bin/logger -t $NAME"

case $1 in
    ups-no-comm|ups-comm-bad)
        $LOGGER "Timer event '$1' fired, restarting NUT..."
        /usr/local/etc/rc.d/nut.sh restart
        ;;
    *)
        $LOGGER "Unrecognized command: '$1'"
        ;;
esac

[dennypage]

Continual disconnects generally indicate an issue with the driver, USB hardware, or with the UPS itself. Before trying to implement scripting work-arounds, I would recommend confirming the source of the problem. Lots of suggestions for this…

First thing I would check would be the physicals: good USB cable, direct connection with no hub, etc. Also, try a different USB port and see if this affects the behavior. Run "usbconfig dump_device_desc" during normal operation. What is the vendor id? Does it match what you expect? Run usbconfig again after the disconnect happens. Does the device still appear?

I would also validate the NUT driver. Did the vendor recommend using the blazer driver? Have you tried any of the other usb based drivers? On the Services / UPS /Status page, do all the items in the UPS Detail section make sense? Are they reasonable values? If not, then there is a driver/configuration problem.

[n3by]

After ~12h since my last post when I deleted the cron restart job every 9 minute and I added upssched.conf and custom-upssched-cmd until now I have 0 lost/bad communication event.

here are some data from USB and UPS with my settings

ugen0.2: <usb to="" serial="" inno="" tech=""> at usbus0, cfg=0 md=HOST spd=LOW (1.5Mbps) pwr=ON (100mA)

  bLength = 0x0012
  bDescriptorType = 0x0001
  bcdUSB = 0x0110
  bDeviceClass = 0x0000  <probed by="" interface="" class="">
  bDeviceSubClass = 0x0000
  bDeviceProtocol = 0x0000
  bMaxPacketSize0 = 0x0008
  idVendor = 0x0665
  idProduct = 0x5161
  bcdDevice = 0x0002
  iManufacturer = 0x0001  <inno tech="">
  iProduct = 0x0002  <usb to="" serial="">
  iSerialNumber = 0x0000  <no string="">
  bNumConfigurations = 0x0001</no></usb></inno></probed></usb>

upsmon.conf

DEADTIME 21
POLLFREQ 7
POLLFREQALERT 7
SHUTDOWNCMD "/sbin/halt -p"

ups.conf

pollinterval = 7
default.battery.voltage.high = 13.7
default.battery.voltage.low = 11.5
runtimecal = 180,100,360,50
ignorelb
override.battery.charge.low = 50
override.battery.runtime.low = 600

UPS Detail

Variable 	Value
battery.charge	100
battery.voltage	13.70
battery.voltage.high	13.00
battery.voltage.low	10.40
battery.voltage.nominal	12.0
device.type	ups
driver.name	blazer_usb
driver.parameter.pollinterval	7
driver.parameter.port	auto
driver.parameter.synchronous	no
driver.version	2.7.4
driver.version.internal	0.12
input.current.nominal	2.0
input.frequency	50.0
input.frequency.nominal	50
input.voltage	241.2
input.voltage.fault	241.2
input.voltage.nominal	230
output.voltage	241.2
ups.beeper.status	enabled
ups.delay.shutdown	30
ups.delay.start	180
ups.load	19
ups.productid	5161
ups.status	OL
ups.type	offline / line interactive
ups.vendorid	0665

Blazer is the only driver that work with this UPS, about the vendor of this ActiveJet UPS they have 0 support and documentation and don't replay at emails, I bought this one because it has UTP port IN/OUT isolation which I am using for WAN but I found that is not documented or marked which one is IN and which is OUT…

[dennypage]

USB to serial adapters are often the source of problems.

You might try adding this global directive:

pollinterval=10

to the section entitled "Additional configuration lines for ups.conf" in Services / UPS / Settings / Advanced Settings.

[n3by]

Now after I deleted cron UPS schedule restart I don't have any more problems with lost communication, zero problems in 2 days !

Using upssched.conf & custom-upssched-cmd is the proper way to deal with lost communication if you ask me, no need to schedule a restart for UPS driver if communications is good.

edit:

I was to happy… worked only 2-3 days...

[n3by]

Something is NOT OK !

After 2-3 days without any problem today without any reason UPS communication went offline and no chance to be restored with the changes I made to
custom-upssched-cmd & upssched.conf

I think commands I defined there are not executed, syslog show no trace of NUT service restart.

I restarted manually the service but this problem continued… randomly.
even I changed:

DEADTIME 30
POLLFREQ 10
POLLFREQALERT 10
and
pollinterval = 10

No chance to fix for long term... communication go offline randomly.

Only pfsense reboot solved the UPS communication problem.

Definitely a NUT problem as I did not touched anything on UPS.

Any ideas ?
Do I need something else to set for upssched to work ??

I hope I don't have to return to cron schedule NUT restart… because to schedule a pfsense reboot every night for NUT to work properly it is not a good option.

[doktornotor]

Dude, how about changing your "NoName Polish brand" UPS for something that works, instead of wasting time and blaming NUT?

[n3by]

Nu chance yet "duck", Santa is not authorized to transport this devices yet.

If you use an UPS you will now that this problem is old and common to NUT UPS.

https://forum.pfsense.org/index.php?topic=33860.0
https://forum.pfsense.org/index.php?topic=78977.0
…
just search it if you want more...
https://duckduckgo.com/?q=nut+ups+lost+communication

If a reboot of pfsense fixed this problem after 2-3 days when it worked OK then are low chances to blame only the UPS hardware.

Cron seem a quick&dirty way to fix the problem and everybody recommend this because it is easy to implement... do you want to restart a process that it is working ok every 15 min, just to be sure ?

I think a better way to quick fix is to use monitor that will restart the driver/package only when it is required, ( this is the purpose of upssched.conf ) this is what I try to have it working properly, it already send you notification at defined event why not use that to restart the NUT driver if it is required & only when it is required.

p.s.
having this ( upssched.conf ) working properly will help everybody not only me !

[dennypage]

@n3by:

If a reboot of pfsense fixed this problem after 2-3 days when it worked OK then are low chances to blame only the UPS hardware.

Just the opposite. If a reboot is required to fix the issue, this guarantees that it is not a NUT problem. This leaves you with the USB controller kernel driver, the USB controller itself, the cabling and the UPS. The kernel driver is extremely unlikely, so we'll set that aside. This leaves us with the hardware.

Please do the following:

1. Confirm that you have a good quality USB cable.
2. Confirm that you are not using a hub.
3. Confirm that you have tried a different USB port on the host.
4. Check to see if usbconfig shows the device as still being present after the UPS goes offline.
5. Try unplugging the USB cable for 10 seconds instead of rebooting.

I suspect that what you will find in #4 is that the UPS no longer appears in the USB tree. In #5 I suspect you will find that unplugging the USB connection will cause the UPS to return to the tree.

Btw, the threads you referred to are not really pertinent to your issue. Please do the above steps first before going out to research more complicated situations. After the above steps, if you want to research more information on your issues, I would google for "NUT USB 0665 5161".

[n3by]

this time after ~1:30'after pfsense restart it happened again:

1 - Original USB UPS cable, not sure how good it is, I will try change it but for the moment I don't have another one.
2 - NO HUB - direct connection
3 - YES, the same.
4 - YES it is present, also present after unplugged and replugged cable.

ugen0.2: <usb to="" serial="" inno="" tech=""> at usbus0, cfg=0 md=HOST spd=LOW (1.5Mbps) pwr=ON (100mA)

  bLength = 0x0012
  bDescriptorType = 0x0001
  bcdUSB = 0x0110
  bDeviceClass = 0x0000  <probed by="" interface="" class="">
  bDeviceSubClass = 0x0000
  bDeviceProtocol = 0x0000
  bMaxPacketSize0 = 0x0008
  idVendor = 0x0665
  idProduct = 0x5161
  bcdDevice = 0x0002
  iManufacturer = 0x0001  <inno tech="">
  iProduct = 0x0002  <usb to="" serial="">
  iSerialNumber = 0x0000  <no string="">
  bNumConfigurations = 0x0001</no></usb></inno></probed></usb>

5- Not fixed automatically… fixed only after saved again the UPS Settings.

Any ideas why is not working properly custom-upssched-cmd ?

#!/bin/sh
NAME=`basename $0`
LOGGER="/usr/bin/logger -t $NAME"

case $1 in
	ups-no-comm)
		$LOGGER "Timer event '$1' fired, restart NUT..."
		/usr/local/etc/rc.d/nut.sh restart
        ;;
    *)
        $LOGGER "Unrecognized command: '$1'"
        ;;
esac

upssched.conf

CMDSCRIPT /usr/local/bin/custom-upssched-cmd

PIPEFN /tmp/upssched.pipe
LOCKFN /tmp/upssched.lock

AT NOCOMM * START-TIMER ups-no-comm 30
AT COMMBAD * START-TIMER ups-no-comm 30
AT COMMOK  * CANCEL-TIMER ups-no-comm COMMOK


[n3by]

It is a problem of user rights ?


[dennypage]

@n3by:

this time after ~1:30'after pfsense restart it happened again:

4 - YES it is present, also present after unplugged and replugged cable.
5- Not fixed automatically… fixed only after saved again the UPS Settings.

Re-saving the UPS settings just stops and starts NUT. Simpler to just restart the service.

You may be experiencing multiple communication issues. Previously, you had an event where restarting the service did not work, and you had to reboot the system, yes? It is this type of event that I'm asking you to run usbconfig for. Do you still have pollinterval set to 10 for testing? If not, please put it back as described above. If you are setting anything else, please remove it.

If you want to diagnose the source of the problem, I will do my best to help you. In the end, we will likely find that the UPS is the source of the problem. I know that this is not what you want to hear, but it's almost certainly the case. I'm not able to help you with upssched other than to tell you that working around communications errors is absolutely not what it's intended for.