The deamon dont want to start.

Bug #184678 reported by lipere123
44
This bug affects 1 person
Affects Status Importance Assigned to Milestone
gDesklets
Invalid
Medium
MarioGonzalez

Bug Description

Log messages of /home/ben/.gdesklets/logs/gdesklets%3A0.0.log
/usr/lib/gdesklets/main/__init__.py:118: GtkDeprecationWarning: gtk.threads_init is
deprecated, use gtk.gdk.threads_init instead
  gtk.threads_init()
==========================================================[01/20/08-23:15:20]===
=== Unhandled error! Something bad and unexpected happened. ===
db type could not be determined
in /usr/lib/gdesklets/gdesklets-daemon: line 127 <module>
in /usr/lib/gdesklets/gdesklets-daemon: line 114 _gdesklets_main
in /usr/lib/gdesklets/utils/ErrorFormatter.py: line 118 _new_imp
in /usr/lib/gdesklets/main/Starter.py: line 2 <module>
in /usr/lib/gdesklets/utils/ErrorFormatter.py: line 118 _new_imp
in /usr/lib/gdesklets/config/DaemonConfigger.py: line 1 <module>
in /usr/lib/gdesklets/utils/ErrorFormatter.py: line 118 _new_imp
in /usr/lib/gdesklets/config/StateSaver.py: line 128 <module>
in /usr/lib/gdesklets/config/StateSaver.py: line 30 __init__
in /usr/lib/gdesklets/config/Backend.py: line 68 Backend
in shelve.py: line 225 open
in shelve.py: line 209 __init__
in anydbm.py: line 80 open
/usr/lib/python2.5/anydbm.py
   75 mod = _defaultmod
   76 else:
   77 raise error, "need 'c' or 'n' flag to open new db"
   78 elif result == "":
   79 # db type cannot be determined
> 80 raise error, "db type could not be determined"
   81 else:
   82 mod = __import__(result)
   83 return mod.open(file, flag, mode)

Revision history for this message
lipere123 (benjamin-lipere) wrote :

It start just after have install languages packages (french). But it before this update, it work well. So, i do not understand.

Revision history for this message
lipere123 (benjamin-lipere) wrote :

I have just tested to start the last beta after compilation. Nothing, this bug is not fixed in last beta. Help please!!!!!!!!.
Configuration : AMD 64 Ubuntu Gusty.
-> Daemon do not start.
-> So connection is impossible.

Revision history for this message
Christian Meyer (chrisime) wrote : Re: [Bug 184678] Re: The deamon dont want to start.

this is a python problem (problem with your installation) not a gdesklets
problem.

see: > 80 raise error, "db type could not be determined"

Revision history for this message
lipere123 (benjamin-lipere) wrote :

Thank you for your help.
I use python 2.5

Is it a speficific version of python that i need to use for run gdesklet ??
May be i have doing an automated update that i must not do.

Can you tell me the packages who destroy dependencies of gdesklets ??
I will search for these too.

May be it is an other of the bugs on AMD 64. I have reading on Ubuntu documentation i need to install a unstable version for AMD 64.
So have you an idea, a way of search ???

Thank. lipere123

Revision history for this message
Christian Meyer (chrisime) wrote :

2008/1/22, lipere123 <email address hidden>:
>
> Thank you for your help.
> I use python 2.5

use python 2.4

Is it a speficific version of python that i need to use for run gdesklet ??
> May be i have doing an automated update that i must not do.
>
> Can you tell me the packages who destroy dependencies of gdesklets ??
> I will search for these too.
>
> May be it is an other of the bugs on AMD 64. I have reading on Ubuntu
> documentation i need to install a unstable version for AMD 64.
> So have you an idea, a way of search ???

my experience is, that it mostly only occurs on amd64 systems when people
use python 2.5. try to use python 2.4, it should work.

Revision history for this message
MarioGonzalez (gonzalemario) wrote :

As Christian said, this is not a gdesklets related problem, it's a python problem instead. However lipere123 can you do some test again with the gdesklets 0.36 version? and then please tell us if the bug is still happening. If this so, this should be a bug with a higher importance, that's why, for now, I set the "importance" to Medium.

Changed in gdesklets:
importance: Undecided → Medium
status: New → Incomplete
Revision history for this message
lipere123 (benjamin-lipere) wrote :
Download full text (3.1 KiB)

MarioGonzalez a écrit :
> As Christian said, this is not a gdesklets related problem, it's a
> python problem instead. However lipere123 can you do some test again
> with the gdesklets 0.36 version? and then please tell us if the bug is
> still happening. If this so, this should be a bug with a higher
> importance, that's why, for now, I set the "importance" to Medium.
>
> ** Changed in: gdesklets
> Importance: Undecided => Medium
> Status: New => Incomplete
>
>
Thank you for help.
I have install gDesklets 0.36 from the source and I have that :

"ben@AIMEE:~$ gdesklets
Démarrage du démon gdesklets...
Connexion au démon [ ### ]
==========================================================[03/15/08-12:16:31]===
Registering new control "/usr/local/lib/gdesklets/Controls/Calendar".

Connexion au démon [ ### ]
==========================================================[03/15/08-12:16:31]===
Registering new control "/usr/local/lib/gdesklets/Controls/EventPipe".

Connexion au démon [ ### ]
==========================================================[03/15/08-12:16:31]===
Registering new control "/usr/local/lib/gdesklets/Controls/System".

Connexion au démon [ ### ]
==========================================================[03/15/08-12:16:31]===
Registering new control "/usr/local/lib/gdesklets/Controls/Sensors".

Connexion au démon [ ### ]
==========================================================[03/15/08-12:16:31]===
Registering new control "/usr/local/lib/gdesklets/Controls/HDDTemp".

Connexion au démon [ ### ]
==========================================================[03/15/08-12:16:31]===
Registering new control "/usr/local/lib/gdesklets/Controls/ArrayBuffer".

Connexion au démon [ ###]
==========================================================[03/15/08-12:16:31]===
Registering new control "/usr/local/lib/gdesklets/Controls/Time".

Connexion au démon [ ### ]
==========================================================[03/15/08-12:16:31]===
Registering new control "/usr/local/lib/gdesklets/Controls/URI".

Connecté au démon en 1201 millisecondes.

==========================================================[03/15/08-12:16:32]===
=== Unhandled error! Something bad and unexpected happened. ===

[EXC]
in /usr/local/bin/gdesklets: line 393 <module>
in /usr/local/bin/gdesklets: line 268 parse_command
in /usr/local/bin/gdesklets: line 177 __open_profile
in /usr/local/bin/gdesklets: line 167 __client_daemon
in /usr/local/lib/gdesklets/main/client.py: line 208 set_remove_command
in /usr/local/lib/gdesklets/main/client.py: line 38 __send
in /usr/local/lib/gdesklets/utils/xdr.py: line 75 recv
[EXC]/usr/local/lib/gdesklets/utils/xdr.py

[---] 70 chunk = ""
[---] 71 while (True):
[---] 72 try:
[---] 73 length = ord(s.recv(1))
[---] 74 except:
[ERR]> 75 raise XDRError
[---] 76
[---] 77 if (length): chunk += s.recv(length)
[---] 78
[---] 79 flag = s.recv(1)
[---] 80 if (flag == _CONT): continue
[---] 81 "

Same problem. The demon don't want to start.
Thanks.

                         lipere12...

Read more...

Revision history for this message
Christian Meyer (chrisime) wrote :

The exception occurrs when we don't know why communication with the daemon
failed. There can be several reasons for this:

* a broken installation
* broken bindings
* C libs don't match
* ...

Just to name a few. Maybe there's a clue in the log file ~/.gdesklets/log.
Otherwise you have to find out for yourself what's wrong. Also it'll be
interesting which distro you're using.

Revision history for this message
evilbut (sirclem-13) wrote :

I've had the exact same problem crop up a day or two ago, under i386 Hardy. Tried changing to python 2.4, but the same error came up.

Log messages of /home/tom/.gdesklets/logs/gdesklets%3A0.0.log
/usr/lib/gdesklets/main/__init__.py:118: GtkDeprecationWarning: gtk.threads_init is deprecated, use gtk.gdk.threads_init instead
  gtk.threads_init()

==========================================================[04/28/08-15:04:24]===
=== Unhandled error! Something bad and unexpected happened. ===

[EXC]db type could not be determined
in /usr/lib/gdesklets/gdesklets-daemon: line 127 <module>
in /usr/lib/gdesklets/gdesklets-daemon: line 114 _gdesklets_main
in /usr/lib/gdesklets/utils/ErrorFormatter.py: line 118 _new_imp
in /usr/lib/gdesklets/main/Starter.py: line 2 <module>
in /usr/lib/gdesklets/utils/ErrorFormatter.py: line 118 _new_imp
in /usr/lib/gdesklets/config/DaemonConfigger.py: line 1 <module>
in /usr/lib/gdesklets/utils/ErrorFormatter.py: line 118 _new_imp
in /usr/lib/gdesklets/config/StateSaver.py: line 128 <module>
in /usr/lib/gdesklets/config/StateSaver.py: line 30 __init__
in /usr/lib/gdesklets/config/Backend.py: line 68 Backend
in /usr/lib/python2.5/shelve.py: line 225 open
in /usr/lib/python2.5/shelve.py: line 209 __init__
in /usr/lib/python2.5/anydbm.py: line 80 open
[EXC]/usr/lib/python2.5/anydbm.py

[---] 75 mod = _defaultmod
[---] 76 else:
[---] 77 raise error, "need 'c' or 'n' flag to open new db"
[---] 78 elif result == "":
[---] 79 # db type cannot be determined
[ERR]> 80 raise error, "db type could not be determined"
[---] 81 else:
[---] 82 mod = __import__(result)
[---] 83 return mod.open(file, flag, mode)

Revision history for this message
Adrian Busolini (aab-cs) wrote :

Exactly the same here, it broke on an upgrade from Gutsy to Hardy (32-bit), gdesklets 0.36-1.

I installed python 2.4, and changed the env to python2.4 in the scripts as per Srećko's comment in LP83922, but no dice - I'm still getting the "db type could not be determined" error.

Revision history for this message
anywhere88 (postremus2000) wrote :

Let's try this guys, for me it works.

Install gdesklets 0.36, then edit the following files:

/usr/local/lib/gdesklets/gdesklets
/usr/local/lib/gdesklets/gdesklets-shell
/usr/local/lib/gdesklets/gdesklets-daemon
/usr/local/lib/gdesklets/gdesklets-migration-tool

In each one replace in the first line "python" with "python2.5".

Save and exit, now it works.

(Of course, you must have python2.5 installed).

Revision history for this message
anywhere88 (postremus2000) wrote :

Strange, it seems the way I've posted works only once, and after a reboot or a restartof gdesklets, the errors come back... really strange.

Revision history for this message
evilbut (sirclem-13) wrote :

I've found that if you uninstall gdesklets, delete the /.gdesklets folder, and then reinstall gdesklets, everything works again.

Revision history for this message
Adrian Busolini (aab-cs) wrote :

@anywhere88: I haven't been able to replicate this - I'd already changed these entries to python2.4 (which I had installed), but changing to 2.5 made no difference (and I have py2.5 too).

@evilbut: I just tried this, and yeah, it seems to work perfectly.

  sudo apt-get remove gdesklets
  mv ~/.gdesklets ~/.gdesklets_broken
  sudo apt-get install gdesklets

Attempting to move the old ~/.gdesklets config folder back just produces the same error, but at least it works again now.

Revision history for this message
MarioGonzalez (gonzalemario) wrote :

So, this bug seems to be invalid now. Can someone please tell me if the bug still persists?

Revision history for this message
MarioGonzalez (gonzalemario) wrote :

Adrian, can you attach to this bug in a tarball the broken .gdesklets dir please?

Revision history for this message
Adrian Busolini (aab-cs) wrote :

Here you go; however, I haven't tried gDesklets for a month or so; quite why I'm not sure, but it leads me to a slight uncertainty over whether I've picked out the correct (broken) .gdesklets directory (after the removals & installs). I'm almost certain I have, but unfortunately I can't be entirely sure.

As for the state of the bug, I believe gDesklets was working fine when I last left it. I may well give gDesklets another whirl soon, and I'll be sure to post here if I manage to replicate the bug again, although following the 're-installation' I somewhat doubt I'll experience that issue again.

Thanks

Revision history for this message
MarioGonzalez (gonzalemario) wrote :

In read the log file, it seems you were able to start gdesklets but you were using sidecandy-uptime. That display uses a Sensor which is making problems.

Sensors are deprecated now so, do you still need this desklet? Can you try to make a Control?

Changed in gdesklets:
assignee: nobody → gonzalemario
status: Incomplete → Triaged
Revision history for this message
Bjoern Koch (h.humpel) wrote :

As the author of SideCandy-Uptime I can assure you that it does not use, never has used and never will be using a Sensor ;). It only needs the IExec Control.

Revision history for this message
MarioGonzalez (gonzalemario) wrote :

So this is something to improve, when you read the log, you're not able to decide which message belongs to which display :-\ I re-read the file, and there are more than one display (trying to?) running.

The log is long enough to realize that something is causing troubles but I cannot figure out who or what :-(

Revision history for this message
Bjoern Koch (h.humpel) wrote :

Ufff.... I just looked at the log file again, too. What a mess :/.

Anyway, there are a lot of desklets in it that still use Sensors (all the lt-based desklets like the lm_sensors thingy) as well as the old SideCandy-CPU desklet which puts out a lot of error messages (every time it tries to add a value to the plotter which is like every 500msec or so ;) ).
I can't remember what causes the plotter do write all the error messages, but I know that it started to occur when moving to 0.35.4 (see bug #125262) and I decided to use desklets without plotters only. In 0.36 the error message in the log file has just been altered.

But this all seems to be unrelated to this bug (to me)... :/.

Talking about an improved logging system: a discussion has been started on it a few weeks ago and I just set up a blueprint for it: https://blueprints.launchpad.net/gdesklets/+spec/improved-logging-system

Revision history for this message
Bjoern Koch (h.humpel) wrote :

Any news on this bug ? Can it be closed ?

Revision history for this message
MarioGonzalez (gonzalemario) wrote :

Yes, finally it was a python/ubuntu related bug instead.

Changed in gdesklets:
status: Triaged → Invalid
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.