[qmtest] exceptions.OSError: [Errno 3] No such process

Philippe Makowski makowski at firebird-fr.eu.org
Fri Dec 21 15:29:33 UTC 2007


> Pierre,
> 
> Pierre Chifflier wrote:
>> Hi,
>> 
>> I'm trying to use qmtest on Debian Sid (Unstable), and encounter some
>> problems when using the gui:
>> Every time I try to run a test using the gui, I got an error
>> qmtest.exception
>>   exceptions.OSError: [Errno 3] No such process
>> 
>> The problem seems to be in setpgid (see the end of this mail for a
>> complete traceback), but I can't tell more.
>> There is a zombie process remaining:
>> pollux   31795  0.1  1.0  20544  8276 pts/6    S+   16:10   0:00  |
>> \_ /usr/bin/python /usr/bin/qmtest gui
>> pollux   31883  0.0  0.0      0     0 pts/6    Z    16:10   0:00  |
>> \_ [bash] <defunct>
>> 
>> I tried using command.ExecTest, command.ShellCommandTest and
>> command.ShellScriptTest without success.
>> 
>> 
>> Now, what's weird: all tests succeeds using the CLI:
>> qmtest run
>> --- TEST RESULTS
>> -------------------------------------------------------------
>> 
>>   shell.test_wzd_action                         : PASS
>> 
>> 
>> Is there a way to make the gui work ? Or have I done something wrong ?
> 
> I can confirm the error, as I ran into it, too, on Fedora Core.
> 
> It appears this is a bug in the way some kernel handles setpgid() calls
> in multi-threaded applications. (FWIW, I filed a bug report:
> https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=188578)
> 
> I haven't yet thought of an 'official' work around, as no-one else seemed
> to be concerned by that. May be for the time being you could just catch
> and ignore that particular exception thrown by os.setpgid() by modifying
> your local QMTest installation ?
I get the same problem under Linux with Python 2.5 and Kernel 2.6.22
can you explain me the patch you suggest ?





More information about the qmtest mailing list