NUT package (2.8.0 and below)

flu

I have email alert activated for nut.
Nut is sending messages like "Communications with UPS UPS lost" and "UPS UPS is unavailable" every time something is changed on any interface. Like editing a interface, restarting a OpenVPN server or even when just the PPPoe lost the DSL connection or is reconnecting.
But the UPS is connected directly to pfsense through a serial port so it is not actually not losing connection.

All these "false positive" error messages make it hard to get alarmed should there really be an issue with the UPS sometime.

Is it possible to fix this behaviour?

DurUser

@flu I think I have seen the same thing happen.

dennypage

@DannyBoy2k You are most welcome. Glad it was so easy!

And very impressed that you went through the entire thread.

dennypage

@flu Yes, this happens because when an interface flaps, pfSense restarts all add-on services. The reason for this is that pfSense itself does not have direct knowledge of what dependencies the services have on interfaces (if any).

flu

@dennypage
many thanks for the explanation.
I guess the nut package could be changed so that it does not send messages like "Communications with UPS UPS lost" and "UPS UPS is unavailable" every time it gets restarted by pfsense?

dennypage

@flu The email comes from NUT itself, rather than the pfSense NUT package. And from NUT's point of view, the connection to the UPS is actually being lost. This will be triggered by something that affects the USB or serial connectivity being restarted or reloaded underneath NUT. NUT doesn't know why the connection failed, it just knows that it failed so it sends the email. You should be able to see the order of events in the system log if you want to understand more about what is happening.

kevindd992002

@dennypage So here's what I noticed recently in my system logs after receiving UPS communication lost and established emails:

Feb 19 04:33:53 	php-cgi 		nut_email.php: Message sent to kevindd992002@yahoo.com OK
Feb 19 04:33:49 	php-cgi 		nut_email.php: Message sent to kevindd992002@yahoo.com OK
Feb 19 04:33:46 	upsmon 	43175 	Communications with UPS ups established
Feb 19 04:33:44 	upsd 	86349 	UPS [ups] data is no longer stale
Feb 19 04:33:41 	upsmon 	43175 	Communications with UPS ups lost
Feb 19 04:33:41 	upsmon 	43175 	Poll UPS [ups] failed - Data stale
Feb 19 04:33:41 	kernel 		ugen0.2: <EATON 5E> at usbus0
Feb 19 04:33:38 	kernel 		ugen0.2: <EATON 5E> at usbus0 (disconnected)
Feb 19 04:33:37 	upsd 	86349 	Data for UPS [ups] is stale - check driver

What does "data is stale" even mean? And why does this happen intermittently?

Gertjan

@kevindd992002 said in NUT package:

What does "data is stale" even mean? And why does this happen intermittently?

Probably because :
Feb 19 04:33:38 kernel ugen0.2: <EATON 5E> at usbus0 (disconnected)
Feb 19 04:33:37 upsd 86349 Data for UPS [ups] is stale - check driver

which says to me that the link between the driver (NUT ?) and the UPS device (cable, whatever) isn't "stable".

kevindd992002

@Gertjan said in NUT package:

@kevindd992002 said in NUT package:

What does "data is stale" even mean? And why does this happen intermittently?

Probably because :
Feb 19 04:33:38 kernel ugen0.2: <EATON 5E> at usbus0 (disconnected)
Feb 19 04:33:37 upsd 86349 Data for UPS [ups] is stale - check driver

which says to me that the link between the driver (NUT ?) and the UPS device (cable, whatever) isn't "stable".

Right but it's just weird that it happens randomly. It's not the UPS and cable because those were replaced already. I also tried using another usb port on the apu2c4 to no avail. There are only 2 usb ports available. Maybe the hub itself? How can you tell for sure though?

dennypage

@kevindd992002 It's either the host end or the UPS end. On the host end, it can be hardware (bridge controller) or software (kernel driver). On the UPS end it would just be a flakey USB implementation.

What I generally recommend to people having these types of issues:

Try a different port on the host that uses a different bridge chip. Favor ports that are USB 2 as USB 3 ports usually have more problems.
If you have a USB hub between the host and the UPS, try removing it.
If you don't have a USB hub between the host and the UPS, try adding one.
Try a different USB hub that uses a different bridge chip.
Try a different UPS model.
Try a different host model.

Regarding the apu2c4, you might find this an interesting read.

Bonne chance

q54e3w

@dennypage I'm seeing the following log entries under 2.4.5, anything to be concerned about?

Apr 16 07:09:28	upsmon	92379	Communications with UPS UPS established
Apr 16 07:09:28	upsd	33793	User local-monitor@127.0.0.1 logged into UPS [UPS]
Apr 16 07:09:23	upsmon	92379	Communications with UPS UPS lost
Apr 16 07:09:23	upsmon	92379	Poll UPS [UPS] failed - Write error: Permission denied
Apr 16 07:09:21	nginx		2020/04/16 07:09:21 [crit] 11971#100993: *62736 SSL_write() failed (13: Permission denied) while processing HTTP/2 connection, client: 192.168.20.115, server: 0.0.0.0:4433
Apr 16 07:08:43	upsmon	92379	Communications with UPS UPS established
Apr 16 07:08:43	upsd	33793	User local-monitor@127.0.0.1 logged into UPS [UPS]
Apr 16 07:08:38	upsmon	92379	Communications with UPS UPS lost
Apr 16 07:08:38	upsmon	92379	Poll UPS [UPS] failed - Write error: Permission denied

dennypage

@q54e3w Can you provide information on your NUT configuration please?

q54e3w

UPS is a network connected Eaton 9PX with Network Card 2.

UPS Type: Remote SNMP
UPS Name: UPS
Notifications: Enabled

Driver Settings
Remote IP: 192.168.10.42
Extra args:
mibs = pw
pollfreq = 15
ignorelb

Advanced Settings
FINALDELAY 180
Addition configuration lines for ups.conf = LISTEN 192.168.10.1
Additional ups.users
[remoteuser]
password = mysupersecretpassword
upsmon slave

This provides the following status

UPS Status
Name: UPS@localhost
Summary status: On line
Runtime (H:M:S): 5:17:00
UPS Load:
36.00%
Battery charge:
100.00%
Battery voltage: 52.00
Input frequency: 60.00

dennypage

@q54e3w said in NUT package:

Poll UPS [UPS] failed - Write error: Permission denied

Is there anything that precedes this entry in the log? Perhaps something similar to this?

q54e3w

@dennypage said in NUT package:

Is there anything that precedes this entry in the log? Perhaps something similar to this?

No, no other errors or messages from upsd/upsmon above or below that one. First sign of trouble is the Poll UPS failed.

dennypage

@q54e3w said in NUT package:

@dennypage said in NUT package:

Is there anything that precedes this entry in the log? Perhaps something similar to this?

No, no other errors or messages from upsd/upsmon above or below that one. First sign of trouble is the Poll UPS failed.

I wasn’t asking about other upsd/upsmon messages, rather I was asking about messages associated with links flapping, DHCP activities, etc. Similar to the logs in the link above.

q54e3w

ahhhh, gotcha. There was another issue with UPnP that I resolved and thats when I noticed these entries. I'll keep an eye on it but as I haven't seen a repeat maybe it was related. Thank you.

kevindd992002

@dennypage Why is it that in the NUT GUI I see different set of parameters shown between my two pfsense boxes (same version of pfsense, same hardware, same NUT version, different UPS but both USB):

I'm missing battery voltage and last test result on the second one. Is this normal?

dennypage

@kevindd992002 said in NUT package:

@dennypage Why is it that in the NUT GUI I see different set of parameters shown between my two pfsense boxes (same version of pfsense, same hardware, same NUT version, different UPS but both USB):

I'm missing battery voltage and last test result on the second one. Is this normal?

Not unusual. NUT only reports what what the UPS reports.

Gertjan

True.
This wouldn't be very assuring :

( image is fake of course )