Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 330731 - gconf-sanity-check-2 fails on login
Summary: gconf-sanity-check-2 fails on login
Status: RESOLVED INVALID
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: [OLD] GNOME (show other bugs)
Hardware: x86 Linux
: High major (vote)
Assignee: Gentoo Linux Gnome Desktop Team
URL: N/A
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-08-01 17:41 UTC by Robert Bradbury
Modified: 2010-08-02 16:52 UTC (History)
0 users

See Also:
Package list:
Runtime testing required: ---


Attachments
emerge --info (EmrgInfo.lst,4.26 KB, text/plain)
2010-08-01 17:52 UTC, Robert Bradbury
Details
Gnome .gnomerc-errors file from last attempt to login (gnomerc-errors,41.86 KB, text/plain)
2010-08-01 18:02 UTC, Robert Bradbury
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Robert Bradbury 2010-08-01 17:41:54 UTC
Logging into gnome a pop-up window appears:

There is a problem with the configuration server.
(/usr/libexec/gconf-sanity-check-2 exited with status 256)


Reproducible: Always

Steps to Reproduce:
1. Create a new user (typically via useradd)
2. Copy functional ".profile" from logged (functional) user to new user home directory.
3. Login via X to new user.

Actual Results:  
There is a problem with the configuration server.
(/usr/libexec/gconf-sanity-check-2 exited with status 256)

Also, the new user fails to produce a functional Gnome environment, i.e. no desktop and no panel.  PS reveals:
UID        PID  PPID  C    SZ   RSS PSR STIME TTY          TIME CMD
chrome   25347 25266  0   918  1068   0 11:01 ?        00:00:00 /usr/bin/ck-launch-session /usr/bin/ssh-agent -- /usr/bin/gnome-session
chrome   25396 25347  0 20447  9700   0 11:01 ?        00:00:00  \_ /usr/bin/gnome-session
chrome   25397 25396  0   432   228   0 11:01 ?        00:00:00      \_ /usr/bin/ssh-agent -- /usr/bin/gnome-session
chrome   25607 25396  1 16449 29396   0 11:05 ?        00:00:00      \_ python2 /usr/bin/hp-systray --qt4
chrome   25688 25607  0  4059  9844   0 11:05 ?        00:00:00      |   \_ python2 /usr/bin/hp-systray --qt4
chrome   25689 25688  0  3611  8244   0 11:05 ?        00:00:00      |       \_ python2 /usr/bin/hp-systray --qt4
chrome   25612 25396  0  6991  8044   0 11:05 ?        00:00:00      \_ /usr/libexec/evolution/2.30/evolution-alarm-notify
chrome   25693     1  0  8762 15012   0 11:05 ?        00:00:00 python2 /usr/lib/gdesklets/gdesklets-daemon
chrome   25685     1  0  4221  2724   0 11:05 ?        00:00:00 gnome-screensaver
chrome   25643     1  0  1462  2120   0 11:05 ?        00:00:00 /usr/libexec/gvfsd
chrome   25640     1  0 18848  9708   0 11:05 ?        00:00:00 /usr/libexec/notification-daemon
chrome   25537     1  0  4103  3764   0 11:05 ?        00:00:00 /usr/libexec/gnome-settings-daemon
chrome   25383     1  0   756  1120   0 11:01 ?        00:00:00 /usr/bin/dbus-daemon --fork --print-pid 5 --print-address 7 --session
chrome   25382     1  0   884   836   0 11:01 ?        00:00:00 /usr/bin/dbus-launch --sh-syntax --exit-with-session

The most fundamental components of a gnome session, e.g. gnome-panel, metacity and nautilus are absent!

The "setup" of a gnome session should *never* fail without informing precisely what the problem is!  And the problem should be displayed "on-screen" -- not in a .gnomerc-errors file which the user may not even know exists.  { I realize that this "oversight" should be filed "upstream" with gnome developers. }


Expected Results:  
If I add a "new" user and I setup XSESSION to use "Gnome" logins to the new user ID should not fail to produce a working Gnome session (if Gnome on the system is *known* to work)!  [I am entering this Bug Report using an active Gnome session from a 4+ year old user profile on a system where creating new users fails.]

Creating and logging in under a new Gnome user, on a system which is upgraded from older systems, i.e. standard "emerges" to the bleeding edge, should *always* work, or at least the diagnostic messages as to what is broken should be extremely informative.

All new gnome (e.g. 2.30.X) software subcomponent releases should verify that adding and logging in as a "new" user is feasible *before* they are released to "emerge" status.  This is important given the fragmented release pattern and the propensity for various sub-components to "break" the session capability.  [Do you test various ACPI functions, screensaver function, etc. -- These have been problematic in my experience -- and I'm using Gentoo on a relatively generic HP "desktop" PC.]
Comment 1 Robert Bradbury 2010-08-01 17:52:20 UTC
Created attachment 240957 [details]
emerge --info
Comment 2 Robert Bradbury 2010-08-01 18:02:31 UTC
Created attachment 240959 [details]
Gnome .gnomerc-errors file from last attempt to login

Fatal errors, such as failure to start the metacity, gnome-panel and nautilus should *ALWAYS* be displayed in a login screen/pop-up windows.  Otherwise the novice user has no idea what or how the login failed.

If Linux/Gnome is to succeed as a desktop environment the diagnostics should be extremely clear as to "what went wrong".  Messages of the form:
"Could not send message to GConf daemon: Process /usr/libexec/gconfd-2 received signal 5)
GConf Error: Failed to contact configuration server; some possible causes are that you need to enable TCP/IP networking for ORBit, or you have stale NFS locks due to a system crash. See http://projects.gnome.org/gconf/ for information."

are *NOT* helpful to the novice (or even an experienced user with 15-20 *years* of UNIX/Linux experience).

The errors should be meaningful and precise!
Comment 3 Gilles Dartiguelongue (RETIRED) gentoo-dev 2010-08-02 16:52:25 UTC
> Could not open or create the file "(null)"; this indicates that there may be a > problem with your configuration, as many programs will need to create files in > your home directory. The error was "Failed to create file '/home/chrome/.tmpdir> /gconf-test-locking-file-K8GQGV': No such file or directory" (errno = 2).

This is your problem.