[Nut-upsuser] (bump) Runaway apcsmart process

[Charles Lepple]

Bill,

Apologies for the delay, but I don't think any of the core NUT
developers have serial-based APC units anymore. Is this a built-in
serial port, or is it a USB-to-Serial adapter? If it's the latter, can
you provide information on the hardware and the driver in use?

I don't think we have a way to adjust the debug level of a driver once
it has started, but you might be able to get a bit more info with
"strace -p <pid-of-driver> -o logfile" (which should attach strace to
the running process until you hit Ctrl-C).

-- 
- Charles Lepple

On Sun, Jul 12, 2009 at 12:15 PM, Bill Arlofski<waa-nut at revpol.com> wrote:
>
> Ping... Any thoughts on this?
>
>
> ----- "Bill Arlofski" <waa-nut at revpol.com> wrote:
>> On several occasions I have seen the apcsmart process runaway using
>> 100% CPU
>> while filling my syslog partition by logging the following syslog
>> entries at
>> an incredible rate. I have seen up to 21,800 log entries PER SECOND of
>> the
>> following nature:
>>
>>
>> --[snip]--
>> Jun 25 10:06:38 servername apcsmart[4012]: Communications with UPS
>> lost:
>> Communications with UPS lost - check cabling
>> Jun 25 10:06:38 servername apcsmart[4012]: smartmode: ser_send_char
>> failed:
>> Input/output error
>> Jun 25 10:06:38 servername apcsmart[4012]: smartmode: ser_send_char
>> failed:
>> Input/output error
...
>> --[snip]--
>>
>>
>> There is nothing wrong with the cabling, and nut runs fine until this
>> begins.
>> If I kill apcsmart and then restart the nut ups driver everything goes
>> back to
>> normal for days or weeks or months at a time.
>>
>> I am using nut v2.4.1-r1 on a Gentoo Linux system with a 2.6.27 kernel
>> on an
>> AMD Athlon X2 64bit dual core system.
>>
>> Any thoughts?  What other information can I provide to help diagnose
>> and
>> hopefully solve this issue?
>>
>> Thanks!
>>
>> --
>> Bill Arlofski
>> Reverse Polarity, LLC
>> http://www.revpol.com/