Re: [gdm-list] GDM 'unable to activate session'

[Luciano <luciano joublanc com>]

Hi Ray, thanks for your reply.

I only have one line in .xsession-errors:

/etc/X11/Sessions/Xsession: line 106: exec: xsm: not found

don't think it's relevant as I have another PC with an identical software config, and it works fine even without xsm ...

In my first message I attached the whole error log with debugging turned on in /etc/gdm/custom.conf, but I don't know if it's visible in the mailing list. It's 800+ lines, but I can paste it here if that helps.

Thanks,

Luciano

On Sun, Jun 23, 2013 at 1:16 PM, Ray Strode <halfline gmail com> wrote:

Hi,

I'm unable to log in to Gnome-3.6.3 via GDM. When I enter my password at the user list screen, the screen goes blank and GDM restarts. I'm attaching a log incluI'm having some issues with GDM which I've posted about on my linux distro's forums but got no replies. I'm hoping somebody here can help.

ding debug messages. It appears that one of the slave processes tells the session to exit i.e. it's not crashing. The corresponding lines are:

138 Jun 22 16:35:41 [gdm-simple-slave] DEBUG(+): GdmSlave: ConsoleKit org.freedesktop.DBus.GLib.UnmappedError.CkVtMonitorError.Code0 raised:    _ GDBus.Error:org.freedesktop.DBus.GLib.UnmappedError.CkVtMonitorError.Code0: Session is already active___

So I think what's going on here is:

1) The same login screen is used for user-switching and for starting new sessions.

2) before starting new sessions it tries to user switch first so a user doesn't get logged in twice

3) here it's trying to user switch to the current VT, the one destined for the new session (it probably could avoid doing that,  but the code seems to be set up that way)

4) since the current VT is already active, consolekit refuses to let the user switch change go through

5) it then it's hits the "migrated is false path" here:

      migrated = switch_to_and_unlock_session (slave, fail_if_already_switched);

        g_debug ("GdmSimpleSlave: migrated: %d", migrated);

        if (migrated) {

                /* We don't stop the slave here because

                   when Xorg exits it switches to the VT it was

                   started from.  That interferes with fast

                   user switching. */

                gdm_session_reset (slave->priv->session);

                slave->priv->start_session_id = 0;

                g_free (slave->priv->start_session_service_name);

                slave->priv->start_session_service_name = NULL;

        } else {

                if (slave->priv->greeter_environment == NULL) {

                        /* auto login */

                        start_session (slave);

                } else {

                        /* Session actually gets started from on_greeter_environment_session_stop */

                        stop_greeter (slave);

                }

        }

and proceedes with login.  So I think those messages, though scary looking, are okay.

Anything in your ~/.xsession-errors file?

--Ray