[Nut-upsuser] NUT and Eaton UPS produce a lot of error messages

[Jim Klimov]

With recent NUT releases, driver disconnections should be contained in its
own `nut-driver at eaton` instance (the `@upsname` part is derived from the
`ups.conf` section name). One of the drivers failing and restarting is not
a proper cause for the data server to recycle.

Also, nowadays some (maybe not all) USB-capable drivers should try to
reconnect without restarting. Note however, that in some NUT iterations it
is possible that this feature misfires if the devices get re-enumerated and
the driver tries to reuse results of original libusb discovery (e.g. some
kernels add +1 to "device" number upon each reconnection, such as due to
USB chip/hub reset, EMI causing protocol reset, or a device going to sleep
and yanked back to work).

In fact, recent work on master branch got `nut-scanner` to not suggest the
`device` option by default to avoid such situations.

Jim


On Fri, Jan 19, 2024 at 9:36 AM Matus UHLAR - fantomas <uhlar at fantomas.sk>
wrote:

> On 19.01.24 09:00, Jim Klimov via Nut-upsuser wrote:
> >Well, from your logs it seems that at 05:17:03 your nut-server (upsd)
> >restarted, so an upsmon reconnection attempt at 05:17:09 had an issue with
> >that (config not all applied? strange a bit) but since 05:17:14 it is
> okay.
> >
> >Maybe a few too many banners shown from upsmon, while its same uptime
> seems
> >to continue. Maybe someone should check on that, seems like a good and
> easy
> >newcomer issue to learn the code, so PRs are welcome :D
> >
> >Messages about lack of "*.pid" files are okay, they should look less scary
> >in newer releases of NUT. They confirm that there is no previous daemon
> >with the same basic configuration currently running, so that your new
> >instance might conflict with it or send it some signals intentionally - so
> >it is just a new start in a clean environment with no competitors.
> >
> >It is a bit unfortunate that both syslog (upsd) and stderr (nut-server)
> >messages land into the same systemd journal. It may be possible to add
> >auto-detection of systemd and not-emit one of those streams then, or add
> an
> >option for that to be quieter and better readable (some people can have a
> >separate syslog setup anyway, maybe on a remote "sink" system, so just
> >cutting one off always is not right).
> >
> >So one question that remains is the missing lines from previous lifetime
> in
> >the short excerpt of upsd: why did it restart at that time? :)
>
> I also have Eaton UPS, just different model. It sometimes disconnects and
> nut has to connect again. I guess this is what happened.
>
> I assume the disonnect message is not visible in "Systemctl status"
> message
> because it does not belong to the current process but previous one.
> Looking for syslog (/var/log/messages) or journal messages (journalctl)
> could
> show that. I guess nothing bad happened there.
>
> >On Fri, Jan 19, 2024 at 6:13 AM Stefan Schumacher via Nut-upsuser <
> >nut-upsuser at alioth-lists.debian.net> wrote:
> >> I recently bought a small UPS by Eaton (Ellipse ECO 800) in order to
> >> prevent my
> >> btrfs-fileserver (running Debian 12 Bookworm 12.4 from shutting down
> >> abruptly while writing
> >> something important during a power loss. I am using the version
> >> 2.8.0.7 provided by Debian.
> >> I have found very gooddocumentation on how to set up the UPS and the
> >> services on the server
> >> connected to it. Unfortunately it's in German
> >> (https://techbotch.org/blog/ups-setup/index.html) which is not a
> >> problem for me but possibly for others trying to understand my set-up.
> >>
> >> I will now describe the steps I took and the configuration options I
> >> set and then post the errors of nut-monitor and nut server. I hope
> >> someone can help fix the underlying problem behind these error
> >> messages.
> >>
> >> lsusb shows the UPS:
> >> Bus 001 Device 003: ID 0463:ffff MGE UPS Systems UPS
> >>
> >> So I made this changes to ups.conf
> >> [Eaton]
> >> driver = usbhid-ups
> >> port = auto
> >> vendorid = 0463
> >> pollfreq = 30
> >>
> >> In nut.conf I set mode=standalone.
> >>
> >> And finally I added these lines to upsd.users:
> >> [upsmon]
> >> password = XXXX
> >> actions = SET
> >> instcmds = ALL
> >>
> >> I did a live test, plugged the cord and waited until the server shut
> >> down at 20%. Worked fine.
> >> Upsc also works - I can query my UPS for specific parameters or show
> them
> >> all.
> >>
> >> The problem is the dozens of errors the systemctl status messages
> >> show. I bought the UPS to increase reliability and now I don't know if
> >> the service is working in case of an emergency. How can I fix this ?
> >>
> >> ● nut-server.service - Network UPS Tools - power devices information
> server
> >> Loaded: loaded (/lib/systemd/system/nut-server.service; enabled;
> >> preset: enabled)
> >> Active: active (running) since Fri 2024-01-19 05:17:03 CET; 5s ago
> >> Main PID: 1303 (upsd)
> >> Tasks: 1 (limit: 38253)
> >> Memory: 640.0K
> >> CPU: 3ms
> >> CGroup: /system.slice/nut-server.service
> >> └─1303 /lib/nut/upsd -F
> >> Jan 19 05:17:03 servername nut-server[1303]: fopen /run/nut/upsd.pid:
> >> No such file or directory
> >> Jan 19 05:17:03 servername nut-server[1303]: Could not find PID file
> >> '/run/nut/upsd.pid' to see if previous upsd instance is already
> >> running!
> >> Jan 19 05:17:03 servername nut-server[1303]: listening on 127.0.0.1 port
> >> 3493
> >> Jan 19 05:17:03 servername nut-server[1303]: listening on ::1 port 3493
> >> Jan 19 05:17:03 servername upsd[1303]: listening on 127.0.0.1 port 3493
> >> Jan 19 05:17:03 servername upsd[1303]: listening on ::1 port 3493
> >> Jan 19 05:17:03 servername nut-server[1303]: Connected to UPS [Eaton]:
> >> usbhid-ups-Eaton
> >> Jan 19 05:17:03 servername upsd[1303]: Connected to UPS [Eaton]:
> >> usbhid-ups-Eaton
> >> Jan 19 05:17:03 servername upsd[1303]: Running as foreground process,
> >> not saving a PID file
> >> Jan 19 05:17:03 servername nut-server[1303]: Running as foreground
> >> process, not saving a PID file
> >>
> >> ● nut-monitor.service - Network UPS Tools - power device monitor and
> >> shutdown controller
> >> Loaded: loaded (/lib/systemd/system/nut-monitor.service; enabled;
> >> preset: enabled)
> >> Active: active (running) since Fri 2024-01-19 03:37:28 CET; 1h 41min ago
> >> Main PID: 847 (upsmon)
> >> Tasks: 2 (limit: 38253)
> >> Memory: 3.4M
> >> CPU: 338ms
> >> CGroup: /system.slice/nut-monitor.service
> >> ├─847 /lib/nut/upsmon -F
> >> └─849 /lib/nut/upsmon -F
> >> Jan 19 03:43:08 servername nut-monitor[849]: UPS Eaton at localhost on
> >> battery
> >> Jan 19 03:43:09 servername nut-monitor[916]: Network UPS Tools upsmon
> 2.8.0
> >> Jan 19 03:43:33 servername nut-monitor[849]: UPS Eaton at localhost on
> line
> >> power
> >> Jan 19 03:43:34 servername nut-monitor[920]: Network UPS Tools upsmon
> 2.8.0
> >> Jan 19 05:17:04 servername nut-monitor[849]: Poll UPS
> >> [Eaton at localhost] failed - Write error: Broken pipe
> >> Jan 19 05:17:04 servername nut-monitor[849]: Communications with UPS
> >> Eaton at localhost lost
> >> Jan 19 05:17:04 servername nut-monitor[1305]: Network UPS Tools upsmon
> >> 2.8.0
> >> Jan 19 05:17:09 servername nut-monitor[849]: Login on UPS
> >> [Eaton at localhost] failed - got [ERR ACCESS-DENIED]
> >> Jan 19 05:17:14 servername nut-monitor[849]: Communications with UPS
> >> Eaton at localhost established
> >> Jan 19 05:17:14 servername nut-monitor[1312]: Network UPS Tools upsmon
> >> 2.8.0
>
>
> --
> Matus UHLAR - fantomas, uhlar at fantomas.sk ; http://www.fantomas.sk/
> Warning: I wish NOT to receive e-mail advertising to this address.
> Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu.
> Spam = (S)tupid (P)eople's (A)dvertising (M)ethod
>
> _______________________________________________
> Nut-upsuser mailing list
> Nut-upsuser at alioth-lists.debian.net
> https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/nut-upsuser
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/nut-upsuser/attachments/20240119/227456c7/attachment.htm>