Xerrors

[Steve_3]

Problem: HP-UX 11.0 Workstation logout of CDE. The error I get in /var/dt/Xerrors.
I install a couple of patches for the CDE Runtime and Xserver but still have this problem.

error (pid 1518): Server for display batman:0 terminated unexpectedly 0

Any suggestion?? HELP!!

[Rick Garland]

In /usr/contrib/bin/X11 is a file called dr_dt.

Run this file and see what is reported then fix.

[Alex Glennie]

Steve, does the problem happen for all users ?If not sounds as if something specific to your users setup maybe to blame ? I agree, use the CDE diagnostic dr_dt to clean-up anything obvious, if it still occurs : check $HOME/.dt/errorlogs , also worth ago is adding set -x in your dtprofile and then checking $HOME/.dt/startlog though this normally relates just to startup issues but may show something.

Other things : try moviing out the way the $HOME/.dt directory (CDE wil recreate it) in case some bad resource has been saved and corrupting your users home or current session, also in a similar vain if you have a .Xdefaulys or .cshrc or similar file (include customisations to your .profile and .dtprofile these to can cause similar issues .... I'm affraid with just that error it's a case of elimination.

Lastly if all else fails :



How to run CDE in debug mode :

edit /usr/dt/config/dtrc.d/90_dtlogin_st
# VL_ARGS=""
VL_ARGS="-debug 9"
If [ -x "$DTLOGIN" ]
then
# exec $DTLOGIN $VL_ARGS ......dev/null ....
exec $DTLOGIN $VL_ARGS
/tmp/dtlogin.log 2>&1
else

then reboot....... .

cause the problem again and post the info to the forum ..best of luck !

gut feeling is it sounds as if a bad resource maybe to blame

[Rick Beldin]

error (pid 1518): Server for display
batman:0 terminated unexpectedly
0

This error indicates the Xserver, /usr/bin/X11/X has terminated for some unknown reason. Little information is captured by default on the reason for the failure.

You can trap the Xserver's core file by:

touch /core
chmod 666 /core

The Xserver, when run by CDE or VUE, runs as daemon, which shares roots home directory, but has no permission to write to it. Touching a core file and changing the permissions will allow the Xserver to write its core.

Of course, once the core has been generated, you need to analyze it. The best and easiest tool to use is gdb, which is available for a free download from software.hp.com.

To do a quick analysis of the core file:

1. Make sure that the core is from the
Xserver. Sometimes other processes
overwrite it:

file /core

2. Run gdb to get a stack trace:

/opt/langtools/bin/gdb /usr/bin/X11/X /core
(gdb) bt

This should indicate the failing function which in many cases is enough to pinpoint the failure.

Unfortunately, sometimes this just isn't enough and other, more extreme measures are required to figure out what is going on. In that case, it becomes necessary to log a Response Center call to get someone to focus their efforts on resolving the problem.

Other sources of information that are useful to post are:

/var/X11/Xserver/logs/X0.log
This file contains the list of shared
libraries that the Xserver loaded. A
truncated list sometimes points to a
corrupt shared library.

/var/adm/syslog/syslog.log
Sometimes syslog will contain info
that is useful in isolating the problem.

[Steve_3]

This problem happen again. I am not sure if it is a memory leak or not? Rick, where is this gdb software. I couldn't seem to find it at the hp site.

Thanks,
steve