[Nut-upsuser] Microdowell (cpsups) driver segfault

[varain at flashmail.com]

Hello Arjen,

Monday, June 25, 2007, 10:17:15 AM, you wrote:

> varain at flashmail.com wrote:
>> Hello Arjen,
>> 
>> Sunday, June 24, 2007, 8:44:04 PM, you wrote:
>> 
>>>>> What I would like to know, is what is read from the UPS. The output of
>>>>> 'upsc' can help here.
>>>> Let me know if you need any other data.  IIRC, it sends that info back
>>>> everytime the ups is polled.
>> 
>>> I meant the readings from the UPS Varain has, to see if the reading of
>>> 'output.voltage' when on battery is approximately the same as
>>> 'input.voltage.nominal'. If they are (almost) the same, this is an
>>> indication that the input voltage to the UPS is low. If it is around 208V,
>>> that probably means that we need to convert the values.
>> 
>>> We know the readings from your UPS are close enough, to assume that it
>>> measures them directly, so we don't need a correction factor and/or
>>> offset.
>> 
>>> Best regards, Arjen
>> 
>> Well, it might not be what you expected, but I have a new problem :
>> 
>> I did something like this :
>> 
>> # upscmd ups1 at localhost test.battery.start
>> 
>> (some time later... )
>> 
>> # upscmd ups1 at localhost test.battery.stop
>> (ups1 is the MicroDowell )
>> 
>> there were no messages , but asking for parameters got me this ...
>> # upsc ups1 at localhost
>> battery.capacity: 7
>> battery.charge: 0.0
>> battery.charge.low: 20
>> battery.packs: 2
>> battery.voltage.nominal: 12
>> driver.name: powerpanel
>> driver.parameter.port: /dev/ttyS0
>> driver.version: 2.0.5
>> driver.version.internal: 0.12
>> input.frequency: 0
>> input.frequency.high: 63
>> input.frequency.low: 47
>> input.transfer.high: 272
>> input.transfer.low: 180
>> input.voltage: 0
>> input.voltage.nominal: 230
>> output.voltage: 0
>> ups.firmware: 2.107
>> ups.load: 0
>> ups.mfr: CYBER POWER
>> ups.model: 1000VA
>> ups.power.nominal: 1000
>> ups.realpower.nominal: 700
>> ups.serial: 000000000000
>> ups.status: OFF
>> ups.temperature: 0

> The fact that many readings are zero, is due to a bug in the nut-2.0.5
> powerpanel driver. It should not parse the data from the UPS if the
> reply it gets is shorter than expected. Instead, You'll probably see in
> the logs 'Status read failed!' while the driver was still running and
> after that should tell that the data is stale.

>> Obviously, the ups is still providing power(it's the only power source for this computer).
>> I've tried to restart the driver and got this :
>> 
>> Network UPS Tools - UPS driver controller 2.0.5
>> Network UPS Tools -  CyberPower text protocol UPS driver 0.12 (2.0.5)
>> Warning: This is an experimental driver.
>> Some features may not function correctly.
>> 
>> warning: sent "\r", expected "#2\r" but got ""
>> warning: sent "P4\r", expected "#<something>" but got ""
>> warning: sent "\r", expected "#2\r" but got ""
>> warning: sent "P4\r", expected "#<something>" but got ""
>> warning: sent "\r", expected "#2\r" but got ""
>> warning: sent "P4\r", expected "#<something>" but got ""
>> warning: sent "\r", expected "#2\r" but got ""
>> warning: sent "P4\r", expected "#<something>" but got ""
>> warning: sent "\r", expected "#2\r" but got ""
>> warning: sent "P4\r", expected "#<something>" but got ""
>> Unable to detect a CyberPower text protocol UPS
>> Driver failed to start (exit status=1)

> Well, apparently the UPS is not talking to us anymore. That's strange.

>> I'm not sure what the problem is ...
>> as for solving it, the "windows solution" of restarting the whole thing (ups+computer) would probably work,
>> but it's not pretty ... (not to mention inconvenient , as the system is in a remote site, with no easy access)

> That's really weird. I don't have a Cyber Power UPS available for
> testing here, bug maybe Doug can try out this command and see what happens.

> As always, testing experimental drivers is risky when the system you're
> working on is remote, but that warning comes too late now.

> Best regards, Arjen

After looking around, it seems the ups is no longer willing to talk to anyone.
I suspect a weakness in the ups firmware.
I'll try to see if the MicroDowell software can help
(they have a linux monitoring daemon... if it can talk to the ups, there's still hope ...)

-- 
Best regards,
 varain                            mailto:varain at flashmail.com