A few months ago I had a user process which had taken control of TCP
port 38900. If that port is not available for gconf to communicate with,
JDS will not start and will exhibit the symptoms you're describing where
existing sessions are okay but new sessions will not start. They just
hang at the little hourglass/watch cursor. I believe port 38900 has
something to do with PAPI and/or APOC and gnome is expecting to either
see nothing on the port or an APOC process. If something else has the
port tied up it just hangs there waiting for a response. My
understanding may be incorrect on those details though.
Try running: lsof -i tcp:38900
Once we killed the offending process, everything began working normally
again.
Hope that helps,
Elliott
Per Lönnborg wrote:
I have tried both. (removing hidden files and so on) Same error.
FYI this sunray-server has an uptime of 477 days. There´s plenty of
users of this system so I rather not like to reboot it. This login
error occured 2 days ago.
I also created a completely new account with no files at all in that
users home-directory. Same error.
It does create the .gnome, .gnome2, .gnome2_private (and the
.Xauthority, .dt, .dtprofile dirs as well) in the users homedirectory.
Can the gconfd be restarted without affect all the logged in users?
I repeat that this error only happens for new logins, all users that
already were logged in before this error started is OK. Including my
own session.
But I am alomost certain that if I log out, i will be in trouble as well!
I put an extra SunRay up for this issue, and that one is also getting
the same error.
This is how it looks after a user has tried to log in with JavaDesktop
but gets the black screen:
bash-3.00# ps -ef |grep popopo
popopo 18881 1 0 11:05:28 ? 0:00 /opt/SUNWut/bin/utaudio
popopo 18930 18776 0 11:05:28 pts/42 0:00
/usr/dt/bin/sdt_shell -c unset DT; DISPLAY=:29;
/usr/dt/bin/d
popopo 18610 25997 0 11:05:14 ? 0:01
/usr/openwin/bin/Xsun :29 -nobanner -auth /var/dt/A:29-WEnOXY
-nobanner -dpms
popopo 18885 18776 0 11:05:28 ? 0:00
/opt/SUNWut/lib/utslaunch
root 5671 115 0 11:25:36 pts/423 0:00 grep popopo
popopo 18928 1 0 11:05:28 ? 0:00 /usr/dt/bin/dsdm
popopo 18932 18930 0 11:05:28 pts/42 0:00 -bash -c unset
DT; DISPLAY=:29; /usr/dt/bin/dtsession_res -me
popopo 18947 18945 0 11:05:28 pts/42 0:00 /usr/bin/gnome-session
popopo 18945 18932 0 11:05:28 pts/42 0:00 /bin/ksh
/usr/dt/config/Xsession2.jds
popopo 18776 18614 0 11:05:18 ? 0:00 /bin/ksh
/usr/dt/bin/Xsession
bash-3.00# pkill -u popopo
bash-3.00#
Oct 15 11:24:58 hostname gconfd (popopo-14331) [ID 702911 user.info
<http://user.info>] starting (version 2.6.1), pid 143331 user ´popopo´
Oct 15 11:24:58 hostname gconfd (popopo-14331) [ID 702911 user.error]
Received signal 11, dumping core. Please report a GConf bug
Can gconfd be restarted without affecting logged-in users?
From 819-0918 Java Desktop System Release 3 Administration Guide:
To terminate the GConf daemon, execute the following command:
# gconftool-2 --shutdown
Nothing about restarting the daemon...
Btw, some version info:
Sun Ray Server software 3.1
Solaris 10 3/05
2008/10/14 Steven M. Dodier <[EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>>
I would suspect something in their environment is causing a core
in java that is OK in CDE.
1) Try cleaning their environment variables to a minimum, or start
with basic .cshrc, .cshrc.local, .login, and .login.local files.
If this works, add sections of the previous files back in to find
the culprit.
OR
2) Gunshot approach. Move all hidden files out of the way. With
user logged out, su - <user>, mkdir old_dotfiles, mv .??*
old_dotfiles/. Setup default . files, and try to have user login
to java again. Of course you will have to move their important
stuff (mail, browser settings and favorites, printer settings,
...) from old_dotfiles back into their home directory.
Per Lönnborg wrote:
I am quite a newbie in SunRay fault diagnosing, but here is the case:
Every new user that tries to login AND uses the Java desktop gets only a
black screen (mouse cursor is the only visible thing).
However, if they change to CDE login it works.
I get this fault messages when the users try to login:
Oct 14 11:24:58 hostname gconfd (plo-14331) [ID 702911 user.info
<http://user.info>] starting
(version 2.6.1), pid 143331 user ´plo´
Oct 14 11:24:58 hostname gconfd (plo-14331) [ID 702911 user.error] Received
signal 11, dumping core. Please report a GConf bug
Thanks,
/Per
------------------------------------------------------------------------
_______________________________________________
SunRay-Users mailing list
[email protected] <mailto:[email protected]>
http://www.filibeto.org/mailman/listinfo/sunray-users
_______________________________________________
SunRay-Users mailing list
[email protected] <mailto:[email protected]>
http://www.filibeto.org/mailman/listinfo/sunray-users
------------------------------------------------------------------------
_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users
_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users
