Folks:
On an installation that had been working, suddenly we're getting fatal errors on starting either tkmedit or tksurfer. These appear to be errors trying to perform X functions. Command lines and responses shown below.
I am not aware of recent changes to the environment (new software installed etc) that would relate to this... but one never knows.
1. Anyone happen to know exactly what solves this?
2. Any clues as to how to troubleshoot would be helpful (eg: is there a log that these programs write into at this early of a stage in startup?)
Thanks,
Graham
================= Initialization =================
[graham@gwlinux ~]$ tcsh [graham@gwlinux ~]$ source /brain-bin/fs_20060528/centos4/freesurfer/SetUpFreeSurfer.csh -------- freesurfer-Linux-centos4-stable-pub-v3.0.3 -------- Setting up environment for FreeSurfer/FS-FAST (and FSL) FREESURFER_HOME /brain-bin/fs_20060528/centos4/freesurfer FSFAST_HOME /brain-bin/fs_20060528/centos4/freesurfer/fsfast SUBJECTS_DIR /data/Structural/FreeSurfer_2007/segmixtest/bxbaseline FUNCTIONALS_DIR /brain-bin/fs_20060528/centos4/freesurfer/sessions MINC_BIN_DIR /brain-bin/fs_20060528/centos4/freesurfer/mni/bin MINC_LIB_DIR /brain-bin/fs_20060528/centos4/freesurfer/mni/lib PERL5LIB /brain-bin/fs_20060528/centos4/freesurfer/mni/lib/perl5/5.8.0 GSL_DIR /brain-bin/fs_20060528/centos4/freesurfer/lib/gsl QTDIR /usr/lib/qt-3.3 TCLLIBPATH /brain-bin/fs_20060528/centos4/freesurfer/lib/tcltktixblt/lib FSL_DIR /brain-bin/fs_20060528/centos4/freesurfer/fsl ================= tkmedit: ================= [graham@gwlinux ~]$ which tkmedit /brain-bin/fs_20060528/centos4/freesurfer/bin//tkmedit [graham@gwlinux ~]$ tkmedit X Error of failed request: BadValue (integer parameter out of range for operation) Major opcode of failed request: 78 (X_CreateColormap) Value in failed request: 0x0 Serial number of failed request: 15 Current serial number in output stream: 17 [graham@gwlinux ~]$
================= tksurfer ================= [graham@gwlinux bxbaseline]$ tksurfer 85432 rh white surfer: current subjects dir: /data/Structural/FreeSurfer_2007/segmixtest/bxbaseline surfer: not in "scripts" dir ==> using cwd for session root surfer: session root data dir ($session) set to: surfer: /data/Structural/FreeSurfer_2007/segmixtest/bxbaseline surfer: Reading header info from /data/Structural/FreeSurfer_2007/segmixtest/bxbaseline/85432/mri/T1.mgz surfer: vertices=184105, faces=368206 surfer: single buffered window
Received X error! Error code : 2 Request code : 78 Minor code : 0
Error text : 'BadValue (integer parameter out of range for operation)'
Received X error! Error code : 12 Request code : 1 Minor code : 0
Error text : 'BadColor (invalid Colormap parameter)'
Graham,
I haven't encountered this particular error. However, tkmedit and tksurfer produce logs called .xdebug_tkmedit and .xdebug_tksurfer which have some startup and debug info.
Is the environment variable 'DISPLAY' set to anything unusual? It should be something like ':0.0'
Nick
On Fri, 2007-01-26 at 15:48 -0800, Graham Wideman wrote:
Folks:
On an installation that had been working, suddenly we're getting fatal errors on starting either tkmedit or tksurfer. These appear to be errors trying to perform X functions. Command lines and responses shown below.
I am not aware of recent changes to the environment (new software installed etc) that would relate to this... but one never knows.
Anyone happen to know exactly what solves this?
Any clues as to how to troubleshoot would be helpful (eg: is there a
log that these programs write into at this early of a stage in startup?)
Thanks,
Graham
================= Initialization =================
[graham@gwlinux ~]$ tcsh [graham@gwlinux ~]$ source /brain-bin/fs_20060528/centos4/freesurfer/SetUpFreeSurfer.csh -------- freesurfer-Linux-centos4-stable-pub-v3.0.3 -------- Setting up environment for FreeSurfer/FS-FAST (and FSL) FREESURFER_HOME /brain-bin/fs_20060528/centos4/freesurfer FSFAST_HOME /brain-bin/fs_20060528/centos4/freesurfer/fsfast SUBJECTS_DIR /data/Structural/FreeSurfer_2007/segmixtest/bxbaseline FUNCTIONALS_DIR /brain-bin/fs_20060528/centos4/freesurfer/sessions MINC_BIN_DIR /brain-bin/fs_20060528/centos4/freesurfer/mni/bin MINC_LIB_DIR /brain-bin/fs_20060528/centos4/freesurfer/mni/lib PERL5LIB /brain-bin/fs_20060528/centos4/freesurfer/mni/lib/perl5/5.8.0 GSL_DIR /brain-bin/fs_20060528/centos4/freesurfer/lib/gsl QTDIR /usr/lib/qt-3.3 TCLLIBPATH /brain-bin/fs_20060528/centos4/freesurfer/lib/tcltktixblt/lib FSL_DIR /brain-bin/fs_20060528/centos4/freesurfer/fsl ================= tkmedit: ================= [graham@gwlinux ~]$ which tkmedit /brain-bin/fs_20060528/centos4/freesurfer/bin//tkmedit [graham@gwlinux ~]$ tkmedit X Error of failed request: BadValue (integer parameter out of range for operation) Major opcode of failed request: 78 (X_CreateColormap) Value in failed request: 0x0 Serial number of failed request: 15 Current serial number in output stream: 17 [graham@gwlinux ~]$
================= tksurfer ================= [graham@gwlinux bxbaseline]$ tksurfer 85432 rh white surfer: current subjects dir: /data/Structural/FreeSurfer_2007/segmixtest/bxbaseline surfer: not in "scripts" dir ==> using cwd for session root surfer: session root data dir ($session) set to: surfer: /data/Structural/FreeSurfer_2007/segmixtest/bxbaseline surfer: Reading header info from /data/Structural/FreeSurfer_2007/segmixtest/bxbaseline/85432/mri/T1.mgz surfer: vertices=184105, faces=368206 surfer: single buffered window
Received X error! Error code : 2 Request code : 78 Minor code : 0
Error text : 'BadValue (integer parameter out of range for operation)'Received X error! Error code : 12 Request code : 1 Minor code : 0
Error text : 'BadColor (invalid Colormap parameter)'
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Nick:
Thanks for the quick reply. Relating to the info you added, here's what I see:
1. DISPLAY is set to just ":0" (ie: without the quotes, and no ".0"). Changing it to ":0.0" -- the failures occur as before.
2. The logfiles you mention show the following:
======================================== .xdebug_tkmedit: ------------------ tkmedit started: Fri Jan 26 13:40:02 2007
tkmedit
$Id: tkmedit.c,v 1.274.2.3 2006/04/15 00:13:10 nicks Exp $ $Name: stable3 $
------------------ .xdebug_tksurfer: empty
========================================
So... nothing obvious seems to jump out here.
Graham
Graham,
Does an X app like 'xclock' work? If not, does xclock work if the Freesurfer setup script is not run?
Nick
On Fri, 2007-01-26 at 16:23 -0800, Graham Wideman wrote:
Nick:
Thanks for the quick reply. Relating to the info you added, here's what I see:
- DISPLAY is set to just ":0" (ie: without the quotes, and no ".0").
Changing it to ":0.0" -- the failures occur as before.
- The logfiles you mention show the following:
======================================== .xdebug_tkmedit:
tkmedit started: Fri Jan 26 13:40:02 2007
tkmedit
$Id: tkmedit.c,v 1.274.2.3 2006/04/15 00:13:10 nicks Exp $ $Name: stable3 $
.xdebug_tksurfer: empty
========================================
So... nothing obvious seems to jump out here.
Graham
Nick:
At 1/26/2007 11:23 PM, Nick Schmansky wrote:
Graham,
Does an X app like 'xclock' work? If not, does xclock work if the Freesurfer setup script is not run?
I haven't run xclock, and unfortunately (?) am no longer at work.
However,
-- Machine is running Centos 4.4 -- KDE works fine so far as I can tell -- I can run various GUI editors etc. -- I can run AFNI
So, my main conclusion is that this is FS-specific. Later in the weekend, I'll try launching various apps from a terminal window after running FS setup script, and see if that disrupts anything.
I'm also downloading FS 3.0.4 (was 3.0.3), and I'll see if that does anything for us.
Thanks,
Graham
Nick:
Still having the same trouble as reported last week. Several more tests to report:
1. I installed 3.0.4 -- same result.
2. In addition to several other GUI apps, I ran xclock as you suggested, looks fine.
3. I ldd'ed tkmedit. To get a realistic output I first did:
source tkmedit ... to run the preparatory script that sets up environment variables as would normally be seen at the point that tkmedit.bin runs. Here's the ldd output:
============================================================== [graham@gwlinux ~]$ ldd /brain-bin/fs_20061125/centos4/freesurfer/bin//tkmedit.bin libtix8.1.8.4.so => /brain-bin/fs_20061125/centos4/freesurfer/lib/tcltktixblt/lib/libtix8.1.8.4.so (0x00f42000) libtk8.4.so => /brain-bin/fs_20061125/centos4/freesurfer/lib/tcltktixblt/lib/libtk8.4.so (0x00bf7000) libtcl8.4.so => /brain-bin/fs_20061125/centos4/freesurfer/lib/tcltktixblt/lib/libtcl8.4.so (0x00111000) libglut.so.3 => /usr/lib/libglut.so.3 (0x00bb4000) libSM.so.6 => /usr/X11R6/lib/libSM.so.6 (0x00b5a000) libICE.so.6 => /usr/X11R6/lib/libICE.so.6 (0x00b65000) libXmu.so.6 => /usr/X11R6/lib/libXmu.so.6 (0x0067f000) libXi.so.6 => /usr/X11R6/lib/libXi.so.6 (0x00dee000) libGLU.so.1 => /usr/X11R6/lib/libGLU.so.1 (0x004f5000) libGL.so.1 => /usr/lib/libGL.so.1 (0x0056e000) libXext.so.6 => /usr/X11R6/lib/libXext.so.6 (0x00b24000) libdl.so.2 => /lib/libdl.so.2 (0x00b1e000) libX11.so.6 => /usr/X11R6/lib/libX11.so.6 (0x001aa000) libgsl.so.0 => /brain-bin/fs_20061125/centos4/freesurfer/lib/gsl/lib/libgsl.so.0 (0x00289000) libgslcblas.so.0 => /brain-bin/fs_20061125/centos4/freesurfer/lib/gsl/lib/libgslcblas.so.0 (0x00907000) libjpeg.so.62 => /usr/lib/libjpeg.so.62 (0x003dd000) libtiff.so.3 => /usr/lib/libtiff.so.3 (0x03c6f000) libz.so.1 => /usr/lib/libz.so.1 (0x00b34000) libm.so.6 => /lib/tls/libm.so.6 (0x00af9000) libcrypt.so.1 => /lib/libcrypt.so.1 (0x02f5d000) libc.so.6 => /lib/tls/libc.so.6 (0x009cc000) libXxf86vm.so.1 => /usr/X11R6/lib/libXxf86vm.so.1 (0x00df8000) libXt.so.6 => /usr/X11R6/lib/libXt.so.6 (0x0062c000) libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0x003fb000) libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x00ddf000) libGLcore.so.1 => /usr/lib/libGLcore.so.1 (0x065ea000) libnvidia-tls.so.1 => /usr/lib/tls/libnvidia-tls.so.1 (0x00ba0000) /lib/ld-linux.so.2 (0x009b3000) ==============================================================
Any ideas?
Graham
Graham,
Your 'ldd' output from tkmedit looks identical to what I see locally, so that unfortunately does not give me a hint to the problem.
Does the utility 'glxgears' run ok?
Also, we run v 1.0-8762 of the NVidia driver (and also some v1.0-8178). Version info is found in /proc/driver/nvidia/version.
Another angle: could the colordepth have changed in your X Windows login? Maybe it was changed from 24-bit to 8-bit?
Sorry, but I don't have any other ideas for now.
Nick
On Mon, 2007-01-29 at 15:19 -0800, Graham Wideman wrote:
Nick:
Still having the same trouble as reported last week. Several more tests to report:
I installed 3.0.4 -- same result.
In addition to several other GUI apps, I ran xclock as you suggested,
looks fine.
- I ldd'ed tkmedit. To get a realistic output I first did:
source tkmedit ... to run the preparatory script that sets up environment variables as would normally be seen at the point that tkmedit.bin runs. Here's the ldd output:
============================================================== [graham@gwlinux ~]$ ldd /brain-bin/fs_20061125/centos4/freesurfer/bin//tkmedit.bin libtix8.1.8.4.so => /brain-bin/fs_20061125/centos4/freesurfer/lib/tcltktixblt/lib/libtix8.1.8.4.so (0x00f42000) libtk8.4.so => /brain-bin/fs_20061125/centos4/freesurfer/lib/tcltktixblt/lib/libtk8.4.so (0x00bf7000) libtcl8.4.so => /brain-bin/fs_20061125/centos4/freesurfer/lib/tcltktixblt/lib/libtcl8.4.so (0x00111000) libglut.so.3 => /usr/lib/libglut.so.3 (0x00bb4000) libSM.so.6 => /usr/X11R6/lib/libSM.so.6 (0x00b5a000) libICE.so.6 => /usr/X11R6/lib/libICE.so.6 (0x00b65000) libXmu.so.6 => /usr/X11R6/lib/libXmu.so.6 (0x0067f000) libXi.so.6 => /usr/X11R6/lib/libXi.so.6 (0x00dee000) libGLU.so.1 => /usr/X11R6/lib/libGLU.so.1 (0x004f5000) libGL.so.1 => /usr/lib/libGL.so.1 (0x0056e000) libXext.so.6 => /usr/X11R6/lib/libXext.so.6 (0x00b24000) libdl.so.2 => /lib/libdl.so.2 (0x00b1e000) libX11.so.6 => /usr/X11R6/lib/libX11.so.6 (0x001aa000) libgsl.so.0 => /brain-bin/fs_20061125/centos4/freesurfer/lib/gsl/lib/libgsl.so.0 (0x00289000) libgslcblas.so.0 => /brain-bin/fs_20061125/centos4/freesurfer/lib/gsl/lib/libgslcblas.so.0 (0x00907000) libjpeg.so.62 => /usr/lib/libjpeg.so.62 (0x003dd000) libtiff.so.3 => /usr/lib/libtiff.so.3 (0x03c6f000) libz.so.1 => /usr/lib/libz.so.1 (0x00b34000) libm.so.6 => /lib/tls/libm.so.6 (0x00af9000) libcrypt.so.1 => /lib/libcrypt.so.1 (0x02f5d000) libc.so.6 => /lib/tls/libc.so.6 (0x009cc000) libXxf86vm.so.1 => /usr/X11R6/lib/libXxf86vm.so.1 (0x00df8000) libXt.so.6 => /usr/X11R6/lib/libXt.so.6 (0x0062c000) libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0x003fb000) libgcc_s.so.1 => /lib/libgcc_s.so.1 (0x00ddf000) libGLcore.so.1 => /usr/lib/libGLcore.so.1 (0x065ea000) libnvidia-tls.so.1 => /usr/lib/tls/libnvidia-tls.so.1 (0x00ba0000) /lib/ld-linux.so.2 (0x009b3000) ==============================================================
Any ideas?
Graham
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Nick:
Thanks for the suggestions. It's looking like the problem is not specific to FS after all, but I'll recount the latest clues in case you hit this again.
1. NVidia version is 1.0-8765.
2. glxgears actually fails in same way as FS programs. Hmmm.
3. Color depth, as reported by GUI tools is supposedly "millions of colors" (presumably 24 bit).
Anyhow, looks like I need to pursue this as a generic GL problem.
Graham
Nick and all:
Followup on the earlier thread: tkmedit/tksurfer fail at startup with X BadValue
I was not able to get to the root cause of this problem. However, conducting the installation for the latest nVidia display driver resulted in a system that now does run tkmedit (and glxgears, FWIW).
(I was already running a version of the nVidia driver as up-to-date as the ones you recommended.)
So, not sure whether the installation actually installed anything that fixed a previous bug relative to the particular video card, or whether it replaced some software component that was damaged or missing, or whether it replaced some setting in xorg.conf that was troublesome.
Anyhow, progress has now resumed.
Thanks for the suggestion to try glxgears -- I had thought that other software that was clearly running OK was sufficient evidence that video was OK, but apparently not.
Graham
freesurfer@nmr.mgh.harvard.edu
-
Graham Wideman -
Nick Schmansky