[Nut-upsuser] NUT with Cyber Power 700 AVR

[Thomas Jensen]

On Wed, 24 Feb 2010 12:37:50 +0100, Arjen de Korte
<nut+users at de-korte.org>
wrote:
> Citeren Thomas Jensen <tom.jensen at digitaltoolbox-inc.com>:
> 
>> server:# /usr/local/ups/bin/powerpanel -DDD -a server
>> Network UPS Tools - CyberPower text/binary protocol UPS driver 0.25
>> (2.4.3)
>> Warning: This is an experimental driver.
>> Some features may not function correctly.
>>
>>    0.000000     debug level is '3'
>>    0.014816     Trying binary protocol...
>>    0.222802     send: (2 bytes) => 46 0d
>>    0.678827     read: timeout
>>    0.782803     send: (2 bytes) => 46 0d
>>    1.238833     read: timeout
>>    1.342822     send: (2 bytes) => 46 0d
>>    1.798838     read: timeout
>>    1.822848     Trying text protocol...
>>    1.926817     send: (2 bytes) => 0d 0d
>>    2.282857     read: timeout
>>    2.438895     send: (3 bytes) => 50 34 0d
>>    2.794898     read: timeout
>>    2.950842     send: (3 bytes) => 50 34 0d
>>    3.306846     read: timeout
>>    3.462844     send: (3 bytes) => 50 34 0d
>>    3.818850     read: timeout
>>    3.830840     CyberPower UPS not found on /dev/ttyS0
>>
>> I don't know if it helps at all or not.
> 
> It does help. At least we know it is not a permissions problem and  
> that the driver is attempting to detect the UPS. Sadly, it failed. So  
> most likely, your UPS is behaving differently from what we expect (any  
> combination of baudrate, serial control lines and/or characters used  
> for detection).
> 
>> I did install the CyberPower
>> software for Linux on the machine last night.  I thought this would
help
>> verify if the UPS was communicating with my machine on the serial port.

>> I
>> was able to issue commands and get status from the UPS.  I therefore
>> assumed that both the port and the UPS were working correctly.  I think
>> the
>> following confirms that.
>>
>> server:# lsof /dev/ttyS0
>> COMMAND    PID USER   FD   TYPE DEVICE SIZE NODE NAME
>> pwrstatd 20687 root    4u   CHR   4,64      1365 /dev/ttyS0
>>
>> Any additional help would be appreciated.
> 
> Well, what you could try to do is to capture the communication between  
> the software provided by CyberPower and the UPS. On Windows, this can  
> be done via PortMon
> 
>      http://technet.microsoft.com/en-us/sysinternals/bb896644.aspx
> 
> This can probably also be done under Linux, but I don't know how.

I did find a program that is supposed to be the Linux equivalent for
PortMon.  It is called jpnevulator and is available from Sourceforge at
http://sourceforge.net/projects/jpnevulator/.  I installed it from the
Debian repository.  Naturally, the version in the repository was not the
most current version.  However, I wasn't able to get the program to detect
any traffic on the serial port(s) when I was running the Cyber Power
Powerpanel status commands.

> Posting a 10 second log after startup is usually enough to get a rough  
> idea if it is possible to support the device.

Are you referring to dmesg output?  If so, does it matter in this regards
whether the machine is a virtualized or not.  The UPS is connected to a Xen
Dom0 machine.
> 
> Note that CyberPower has been less than helpful in the past, so if NUT  
> support is important to you, it is probably better to switch to a  
> vendor with a better track record when it comes to supporting NUT.  
> Without vendor support, I feel little urge to invest a lot of (unpaid)  
> time to see what we should do to support them.

I understand your point about prioritizing your time, especially when that
time is unpaid.  I do appreciate the efforts of all the Open Source
community.  I thought NUT would be a useful tool for me when I decided to
purchase a UPS.  When this particular UPS went on sale, I checked the
compatibility list on the NUT website where I found it was listed using
either the cyberpower or powerpanel drivers.  I wasn't aware the
manufacturer has been difficult in the past until you just informed me.

> 
> Best regards, Arjen