[Rock-dev] Vizkit: Segfault through wrong parameters blocks omniorb through persistent logfile?

[Jakob Schwendner]

On 02/15/2011 02:31 PM, Martin Zenzes wrote:
> Hi,
>
> during testing, I got a seqFault:
>
> in  /home/likewise-open/DFKI/zenzes/rock-all/gui/vizkit/ruby/lib/vizkit/vizkit.rb:41: [BUG] Segmentation fault
>
> this was due to wrong parameters given by me... But right after the
> segfault (and fixing my code), I get:
still, this kind of error message should be avoided. Maybe you could 
tell, what you did to cause the segfault
>> /home/likewise-open/DFKI/zenzes/rock-all/tools/orocos.rb/lib/orocos/process.rb:260:in>  `spawn': there is already a running task called cyber_glove, are you
>> starting the same component twice ? (ArgumentError)
> while trying to start my module. Restarting corba by
> calling /etc/init.d/omniorb4-nameserver hadn't any effect, I had to
> delete the logfile additionally after restarting onmiorb:
The crash just affected the ruby part, and left the orocos module 
running. In this case we just do a
killall <deployment_name>
to stop the process.

>> rm -f /var/lib/omniorb/omninames-mzenzes-u.{log,bak}
> Since deleting the files is not obvious (restarting should remove all
> loaded tasks?) this may help someone.
imo, there is no need to delete the files. Killing the process is 
usually enough.

Jakob