← Back to team overview

opensand team mailing list archive

Python issue during Open Sand platform installation

 

Hi Didier,

This is my set of answers. All logs which are avaiable in this thread are
generated on one and the same VM which is installed on my laptop.



>  > 1. Installation of a daemon from packages
> >
> > According to the OpenSAND manual user has to install daemon first. As
> > a result (for previous and present version of manual guide) there is
> > a small issue with network I presume.
> >
> >  * Shutting down the OpenSAND Daemon
> > sand-daemon
> *>  * /var/run/sand-daemon/pid does not exist
> > [fail]*
>
> That error message appears if the daemon service is restarted while the
> daemon is not running. It should be fixed. Could you please open a new
> ticket in the bugtracker? See instructions given in [1]. Give a clear
> description of the problem and the way to reproduce it.
>
> [1] http://opensand.org/developer/bugtracker.php#report
>
>
Ticket and its description for a bug-tracker is in place. This and next
issue in this section are connected to each other. Short description can be
found in my next comment.


> > During first installation of a daemon it partly fails - I haven't seen
> > configuration panel (maybe daemon use default settings which sounds
> > quite reasonable) but as a user I can use command for reconfiguration
> > (full example in attached log file - log.txt)
> >
> > andyg@andyg-laptop:~/opensand/packages$ sudo dpkg-reconfigure
> > opensand-daemon
> > * Shutting down the OpenSAND Daemon
> > sand-daemon
> *> * /var/run/sand-daemon/pid does not exist
> > [fail]*
> >
> > invoke-rc.d: initscript opensand-daemon, action "stop" failed.
> >  * Removing configuration for the OpenSAND
> > network
> > # Configure OpenSAND network for satellite
> > [ OK ]
> >
> > rsyslog start/running, process 3268
> >  * Configuring the OpenSAND
> > network                                    [ OK ]
> >  * Starting the OpenSAND Daemon sand-daemon                    [ OK ]
> >
> > Similar problem appears with new version of manual (section 2 in a
> > log file for a daemon installation from packages) but I think that
> > whenever user reach an access to configuration panel of a daemon and
> > enter proper network settings, then there will be no problem with
> > network configuration in future attempts for demon installation.
>
> You should have been prompted for the configuration. Did you already
> have installed/configured the package before on that system?


I've download, install and reconfigure daemon many times. When I had no
clue what is going on I've download packages once again and start procedure
again. I've never seen on my laptop the configuration window prompted
during the installation process. My laptop has the first VM created for
OpenSAND tests.

Could you please try to reproduce the problem on another system?
>

Results of a test which is listed above was performed on my laptop
installation of Ubuntu 10.04 VM. I did a lot of testing on that system so I
decided to reproduce this issue with other VM available in my small lab.

*Conclusion* for updated "installation manual" (packages section) and fresh
VM:

Additional tests reveal that during first compilation daemon behaves as you
expect - user is prompted to configuration panel.
However that problem which is highlighted above appears every time you turn
on machine and want to compile daemon from packages. After very first
compilation of a daemon, the Configuration Panel can be reached only
through reconfiguration command.

Full log file is attached to the post: OpanSAND test bed - compilation with
packages.

*Bug description:* During first compilation of the opensand daemon (so far
tested only on packages) we can see an error massage in the terminal
window.  Most probably that error message appears when the daemon service
is restarted while the daemon itself is not running (according to Didier's
diagnose). Scenario is easy to reproduce. Requirements - fresh installation
of Ubuntu 10.04 VM ( in my case installed on Oracle Virtual Box). I haven't
update my system after installation. I went strait to Opensand Installation
Manual (latest changes available since 28th of August) and  perform all
necessary steps before packages installation (thanks to it system was
updated with necessary packages which are required for successful
installation/compilation process). You can check results of my actions in
mentioned above log file.


>
> > 2. Installation of a manager from packages>
>

> After successful installation of a daemon it's time for manager.
> > Installation goes smooth till I want to use command $ sand-manager
> >
> > andyg@andyg-laptop:~/opensand-0.9.0-deb$ sand-manager
> > Traceback (most recent call last):
> >   File "/usr/bin/sand-manager", line 39, in <module>
> >     import gtk
> > ImportError: No module named gtk
> >
> > So i tried to import that gtk module but without success which
> > represents start of the manager. I even updated my python to version
> > 2.7 and install some additional gtk modules from packages but without
> > positive result. Results of my update are listed in section 4 of
> > created log file.
> >
> > Here is another example of an issue with a Python which also haven't
> > bring me any clue what I can do more in order to fix it. Sorry for
> > Polish in the log text. I'm working on converting my terminal to
> > English shell version. It looks good till Traceback line shows up.
>
> According to the logs, you somehow fixed the import gtk problem since
> the sand-manager complains about missing encapsulation modules. Did you
> remember what you tried between the 'import gtk' problem and the
> 'missing encapsulation modules' one?
>
>
I did some research in area of gtk and that particular error. I visited
many forums looking for an answer (I was convince that there is some issue
with my VM) and I've install some additional packages which should support
gtk. Because of encounter problems I decide to reproduce this test scenario
but in the end it wasn't possible. I haven't notice similar issues with
other VM's.
At the moment I have 7 new VM's working on 3 different computers (set of 3
VM's - SAT, GW, ST - on two strongest machines for two independent test
beds and 7th for additional network tests). First test bed with 3 VM's is
fully updated. Remaing 4 VM's are instance of fresh installation which are
in use for packages and source compilation.


> The problem about 'missing encapsulation modules' happens because at
> least one encapsulation module is required. You need to install one of
> them (ATM/AAL5 for example).
>
> Could you please open another ticket in the bugtracker? See instructions
> given in [1]. Give a clear description of the problem and the way to
> reproduce it.
>

If you still want me to open a ticket for this bug I will do it. I'm not
sure that is necessary to investigate it. I keep this "broken" VM from my
laptop just in case.


>
> To avoid Polish, use LANG=C LC_ALL=C before the commands, eg.:
>   $ LANG=C LC_ALL=C dpkg -i some_package.deb
>

Thanks for a tip :) It's helpful. I have use it and it works perfect.


>  > andyg@andyg-laptop:~/opensand-0.9.0-deb$ sudo dpkg -i
> > libopensand-atm-aal5-plugin_0.9.0_i386.deb
> > Zaznaczenie poprzednio niezaznaczonego pakietu
> > libopensand-atm-aal5-plugin. (Odczytywanie bazy danych ... 230772
> > plików i katalogów obecnie zainstalowanych.)
> > Rozpakowanie libopensand-atm-aal5-plugin (z
> > libopensand-atm-aal5-plugin_0.9.0_i386.deb) ...
> > Konfigurowanie libopensand-atm-aal5-plugin (0.9.0) ...
> > Traceback (most recent call last):
> >   File "/usr/libexec/opensand/atm_support.py", line 39, in <module>
> >     import debconf
> > ImportError: No module named debconf
> >  * Shutting down the OpenSAND Daemon
> > sand-daemon
> >  * /var/run/sand-daemon/pid does not exist
> >             [fail]
> >
> > * Starting the OpenSAND Daemon
> > sand-daemon                              [ OK ]
> >
> > Przetwarzanie wyzwalaczy dla libc-bin...
> > ldconfig deferred processing now taking place
>
> That's strange. You should have the debconf python module. Could you
> please send me the result of the following commands:
>  $ ls /usr/lib/python*/dist-packages/debconf.py
>  $ python --version
>  $ sudo python --version
>  $ env | grep -i python
>  $ sudo env | grep -i python
>


This what I get from my terminal

andyg@andyg-laptop:~$  ls /usr/lib/python*/dist-packages/debconf.py
/usr/lib/python2.6/dist-packages/debconf.py
/usr/lib/python3.1/dist-packages/debconf.py
andyg@andyg-laptop:~$ python --version
Python 2.7
andyg@andyg-laptop:~$ sudo python --version
[sudo] password for andyg:
Python 2.7
andyg@andyg-laptop:~$ env | grep -i python
andyg@andyg-laptop:~$ sudo env | grep -i python
andyg@andyg-laptop:~$

Any idea what might be wrong with this VM ?? As I said before I have a
filing that installation of Python 2.7 wasn't a good idea.


>  > 3. Installation of a daemon from sources
> >
> > When my attempts for SAND installation based on packages failed last
> > week (short report delivered on 15th of August) I decide to give
> > another shot with source packages downloaded through Bazaar. It's
> > definitely longer approach but very educational. Unfortunately I hit
> > a wall when I had to build daemon with following result:
> >
> > andyg@andyg-laptop:~/opensand/opensand-daemon$ python setup.py build
> > Traceback (most recent call last):
> >   File "setup.py", line 2, in <module>
> >     from setuptools import setup, find_packages
> > ImportError: No module named setuptools
>
> I re-wrote that part this morning. Did you installed the dependencies
> listed by [2]?
>
> [2]
> http://opensand.org/support/wiki/doku.php?id=installation_manual#install_required_build_tools


Dependencies are properly installed.


> > It was a first time that I had an impression that there is something
> > wrong with my Python installation. So i decide to update it but as
> > you can see also in section above that move haven't solve this issue.
>
> I think so. The previous commands will maybe help us to understand the
> problem.
>

Regardless to results of those test command above, so far I couldn't
replicate this issue at any other VM in my lab. I'm quite positive that by
accident I corrupt this VM. I perform a lot of testing with it, and
downloaded almost all available updates.

That would be all what I can add to this tread.

With kind regards,

Andy


>
> --
> Mailing list: https://launchpad.net/~opensand
> Post to     : opensand@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~opensand
> More help   : https://help.launchpad.net/ListHelp
>

Follow ups

References