Nut upsuser - Sep 2019 - Debian 10 nut 2.7.4-8 and APC Back UPS 600i on 940-0020B

> On Sun, 8 Sep 2019, tovis r-visor wrote:
>
>> After long struggling, check every parts of UPS and cable, I have
realized
>> that is after boot I reload the driver the system work well:
>> # systemctl restart nut-driver.service
>> I was checked up to real "battery low" and shutdown, switch
off the UPS
every thing goes well.
>> I was activated /etc/rc.local but it is starting before that Netwrok
UPS Tools starting - nothing to restart.
>> What can I do now?
>> Any suggestion?
>
> Normally there is no need to explicitly start nut-driver.servce.  It is
started
> automatically by nut-server.service. How do you start your NUT
> installation?
> Do you use the systemd script provided by NUT or other systemd service
units?
> Could you show us the script/service units?
>
> Roger
>
>
Hi Roger!
It's all about Debian 10 binary packages. I'm nothing change exclude
configuration files. Actually I found these units:
# systemctl list-unit-files | grep nut
nut-client.service   enabled
nut-driver.service   static
nut-monitor.service  enabled
nut-server.service   enabled

I does not change any of them. Simple, after boot, using
$ watch upsc MYUPS
what is give me "LB OL"
from another terminal I restart:
# systemctl restart nut-driver.service
after some seconds the status is changed to simply "OL".
Also for sure I was unplug 230V from UPS the system was shutdown after
about  15 min and at finish the UPS was switched off, as usually (I was
used a 12V/7Ah accumulator to get result faster).
Unfortunately I do not know ho to automatize this "work around".

Any suggestion?
  tovis


_______________________________________________
> Nut-upsuser mailing list
> Nut-upsuser at alioth-lists.debian.net
> https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/nut-upsuser

On Sep 11, 2019, at 5:25 AM, tovis r-visor <tovis at r-visor.hu>
wrote:
> 
> I does not change any of them. Simple, after boot, using
> $ watch upsc MYUPS
> what is give me "LB OL"
> from another terminal I restart:
> # systemctl restart nut-driver.service
> after some seconds the status is changed to simply "OL".
Are there any systemd units that change the serial port settings, e.g.
setserial? They might be changing the cable power settings (which are re-applied
when the driver is restarted). Check the journal to see the order of the units
at startup.

Also, does your UPS support smart signaling? The APC smart protocol driver
(apcsmart) is far more resilient than the genericups driver, IMHO. It may be
worth buying/building a new cable, if needed, to use the smart protocol.

> On Sep 11, 2019, at 5:25 AM, tovis r-visor <tovis at r-visor.hu>
wrote:
>>
>> I does not change any of them. Simple, after boot, using
>> $ watch upsc MYUPS
>> what is give me "LB OL"
>> from another terminal I restart:
>> # systemctl restart nut-driver.service
>> after some seconds the status is changed to simply "OL".
>
> Are there any systemd units that change the serial port settings, e.g.
> setserial? They might be changing the cable power settings (which are
> re-applied when the driver is restarted). Check the journal to see the
> order of the units at startup.
>
> Also, does your UPS support smart signaling? The APC smart protocol driver
> (apcsmart) is far more resilient than the genericups driver, IMHO. It may
> be worth buying/building a new cable, if needed, to use the smart
> protocol.
>
>
Thank you! Up to now I was looking for that. There is a suspicious
message, coming from boot screen:
[ OK ] Started controls configuration of serial ports.
Also seem to be that messages about "Network UPS tools" are after, but
in
the same time, I was seen that scripts (such as rc.local) executed before
the message appear. I'm really confused with systemd start up sequences :(
I found a workaround, disabled all nut service units and put calling of
scripts to the rc.local:
systemctl start nut-monitor.service
systemctl start nut-server.service

This way the false state is gone :)
I understand, that this workaround is dangerous, especially when system is
upgraded (even safe-upgrade), failure could be detected only after reboot.
My APC Back UPS 600i is really old, from early 90's - no smart protocol,
ONLY simple signaling using (DIY) cable APC 940-0020B.

OFF: These UPS -s are really simple, they could be simply
serviced/remounted (no firmware, no mcu). Last cycle YUASA accumulator was
working almost 5 years. The trade off is the simplicity no smart
signaling, no information about accumulator healthy, no sine wave.

I still await for better solution or suggestion
  tovis