[Nut-upsuser] eaton evolution not shutting down

[Arnaud Quette]

2012/2/22 Laurent Lesage <laurent at lesagepono.be>:
> Bonjour Arnaud,

Bonjour Laurent,

sorry for the lag in answering... long week end.

> Arnaud Quette a écrit :
>
> 2012/2/21 Laurent Lesage <laurent at lesagepono.be>:
>
>
> Hi all,
>
>
> Bonjour Laurent,
>
>
>
> I just changed my old MGE ellipse for a evolution 850 (now eaton). I can
> connect using mge-shut (newmge-shut doesn't work). But I have a few
> issues. I looked around on NUT web-site and googled with no luck.
> Using debian lenny on 2.6.18 kernel (due to some ISDN driver that
> doesn't work with earlier kernels)
> 1) the most annoying is that I cannot shutdown the UPS. here is a trace :
>
>
> Network UPS Tools - MGE UPS SYSTEMS/SHUT driver 0.66 (2.2.2)
> debug level is '2'
> entering upsdrv_initups()
> entering shut_ups_start()
>
> Communication with UPS established
> entering shut_get_descriptor(n 21, 9)
> shut_wait_ack(): ACK received
> entering shut_get_descriptor(n 01, 18)
> shut_wait_ack(): ACK received
> Device Descriptor:
> bLength:         0x12
> bDescriptorType:                    0x01
> bcdUSB:         0x0110
> bDeviceClass:         0x00
> bDeviceSubClass:                    0x00
> bDeviceProtocol:     0x00
> bMaxPacketSize0:                    0x08
> idVendor:         0x0463
> idProduct:         0xffff
> bcdDevice:                        0x0100
> iManufacturer:         0x01
> iProduct:                        0x02
> iSerialNumber:         0x03
> bNumConfigurations:     0x01
>
> entering shut_get_descriptor(n 22, 1811)
> shut_wait_ack(): ACK received
> Unable to get Report Descriptor
>
>
> If I put more "D", I see it stucks on "need more data!" like this :
>
>
> need more data (1691)!
> Receive: (2 bytes) => 04 88
> Receive: (8 bytes) => 10 b1 82 c0 05 84 09 11
> shut_checksum: 7a => OK
> sent: (1 bytes) => 06
> need more data (1683)!
>
>
> 2) I tried to set some variables. The drivers accept to do it (upsrw)
> but the actually do not change. some change but not with the value I set
> in the command and come back to their initial value after a few seconds.

I will need some more details on this point:
- which variables
- a driver debug output when changing values

> Any help appreciated.
>
>
> sure, but I need more info.
> first, issue (1) means that the driver can't start, while (2) implies
> that the driver is started...
> is (1) a repeatable issue?
>
>
> Yes. It happens only when trying to shutdown (upsdrvctl shutdown)

then, are you sure that, when calling "upsdrvctl shutdown", the
previous driver instance has been stopped?
Ie, before calling "upsdrvctl shutdown", you need to call "upsdrvctl stop".


> or is is only that when you're trying to shutdown the UPS (Ie
> 'upsdrvctl shutdown' or 'mge-shut -k ...'), it fails?
>
> can you please send back a full trace, with debug level 5 (Ie -DDDDD),
> in compressed form?
>
>
> here it is (attached). I have to type ''/lib/nut/mge-shut -a evolution -k
> -DDDDD  > shutdown.trace 2>&1'' to get it (upsdrvctl gives no detail).
>
> Finally, note that 2.2.2 is quite oldish now, so you may also consider
> an update.
>
>
>
> Sure.... :-) (It's the "lenny" version for the moment)
>
> I also noticed sthg strange : the variables and values listed with "upsrw"
> seem to be defautl values. I changed some with the windows configurator but
> they stil seem to be the same. If I restart the window MGE software, it gets
> the modified values, not the default ones. So, it seems (it seems only!)
> that the values the "mge-shut" driver gets are wrong. But the variable names
> are meaningful and totally related to the functionnalities of the UPS  - so
> their names must be correct.
>
> for example, upsrw gives (but for example, the "outlet.2.delay.start" should
> be 120, as configured by the windows software - and here it is "-1") :

nothing abnormal.
this is due to the difference between NUT and Windows MGE software in
storing params and settings values:
for example,  "outlet.2.delay.start" set to 120 is stored by the
Windows software, but not applied into the device.
whenever you see this "-1", it is a timer. whenever you set a timer to
a value, countdown start, and in the end, the action (Ie start, stop,
cycle, ...) is executed.

on its side, NUT stores these default value in ups.conf.

a modification has been done on the UPS collection, to allow for
storing the param value (ups.delay.start) and the actual timer value,
when decrementing (ups.timer.start). The same should be done for
outlet too.
I've logged a task for this one:
https://alioth.debian.org/pm/task.php?func=detailtask&project_task_id=492&group_id=30602&group_project_id=318

> ...
> Unable to get Report Descriptor
> Network UPS Tools - MGE UPS SYSTEMS/SHUT driver 0.66 (2.2.2)

well, the situation has improved since version 0.66.
it would be worth trying more recent version. But I will probably lag
in producing a backport for lenny...

cheers,
Arnaud
-- 
Linux / Unix Expert R&D - Eaton - http://powerquality.eaton.com
Network UPS Tools (NUT) Project Leader - http://www.networkupstools.org/
Debian Developer - http://www.debian.org
Free Software Developer - http://arnaud.quette.free.fr/