Hi,
Following some system upgrade to our Linux 2.6 systems, freesurfer has ceased to work. When running recon-all on bert, we get a message referring to a broken pipe and a crash in mincinfo. We are running Gentoo:
doolin ~ # uname -a Linux doolin 2.6.21-gentoo-r4 #1 SMP Thu Aug 9 09:33:11 PDT 2007 i686 AMD Athlon(tm) 64 X2 Dual Core Processor 4400+ AuthenticAMD GNU/Linux doolin ~ #
Any help and/or advice would be appreciated.
Thank you, Chuck
mri_add_xform_to_header -c /vxfsvol/home/staff/chuck/freesurfer-test/bert/mri/transforms/talairach.xfm /vxfsvol/home/staff/chuck/freesurfer-test/bert/mri/orig.mgz /vxfsvol/home/staff/chuck/freesurfer-test/bert/mri/orig.mgz
INFO: extension is mgz #-------------------------------------------- #@# Nu Intensity Correction Fri Oct 12 10:37:03 PDT 2007
mri_nu_correct.mni --i orig.mgz --o nu.mgz --n 2
/vxfsvol/home/staff/chuck/freesurfer-test/bert/mri /usr/local/freesurfer/bin/mri_nu_correct.mni --i orig.mgz --o nu.mgz --n 2 nIters 2 $Id: mri_nu_correct.mni,v 1.3.2.1 2006/11/24 20:16:38 nicks Exp $ Linux doolin 2.6.21-gentoo-r4 #1 SMP Thu Aug 9 09:33:11 PDT 2007 i686 AMD Athlon(tm) 64 X2 Dual Core Processor 4400+ AuthenticAMD GNU/Linux Fri Oct 12 10:37:03 PDT 2007 Program nu_correct, built from: Package MNI N3, version 1.10, compiled by nicks@kani (i686-pc-linux-gnu) on 2005-11-15 at 21:02:27 tmpdir is ./tmp.mri_nu_correct.mni.28553 /vxfsvol/home/staff/chuck/freesurfer-test/bert/mri mri_convert orig.mgz ./tmp.mri_nu_correct.mni.28553/nu0.mnc mri_convert orig.mgz ./tmp.mri_nu_correct.mni.28553/nu0.mnc reading from orig.mgz... TR=0.00, TE=0.00, TI=0.00, flip angle=0.00 i_ras = (-1, 0, 0) j_ras = (0, 0, -1) k_ras = (0, 1, 0) writing to ./tmp.mri_nu_correct.mni.28553/nu0.mnc...
-------------------------------------------------------- Iteration 1 Fri Oct 12 10:37:11 PDT 2007 nu_correct -clobber ./tmp.mri_nu_correct.mni.28553/nu0.mnc ./tmp.mri_nu_correct.mni.28553/nu1.mnc [chuck@doolin:/vxfsvol/home/staff/chuck/freesurfer-test/bert/mri/] [2007-10-12 10:37:11] running: /usr/local/freesurfer/mni/bin/nu_estimate_np_and_em -parzen -log -sharpen 0.15 0.01 -iterations 50 -stop 0.001 -shrink 4 -auto_mask -nonotify -b_spline 1 -distance 200 -quiet -execute -clobber -nokeeptmp -tmpdir /usr/tmp/nu_correct_28603/ ./tmp.mri_nu_correct.mni.28553/nu0.mnc ./tmp.mri_nu_correct.mni.28553/nu1.imp
nu_estimate_np_and_em: broken pipe nu_estimate_np_and_em: crashed while running mincinfo (termination status=13) nu_correct: crashed while running nu_estimate_np_and_em (termination status=65280) ERROR: nu_correct Linux doolin 2.6.21-gentoo-r4 #1 SMP Thu Aug 9 09:33:11 PDT 2007 i686 AMD Athlon(tm) 64 X2 Dual Core Processor 4400+ AuthenticAMD GNU/Linux
recon-all exited with ERRORS at Fri Oct 12 10:37:11 PDT 2007
chuck@doolin /usr/local/freesurfer $
Hi Chuck,
is that partition full?
Bruce On Fri, 12 Oct 2007, Chuck Theobald wrote:
Hi,
Following some system upgrade to our Linux 2.6 systems, freesurfer has ceased to work. When running recon-all on bert, we get a message referring to a broken pipe and a crash in mincinfo. We are running Gentoo:
doolin ~ # uname -a Linux doolin 2.6.21-gentoo-r4 #1 SMP Thu Aug 9 09:33:11 PDT 2007 i686 AMD Athlon(tm) 64 X2 Dual Core Processor 4400+ AuthenticAMD GNU/Linux doolin ~ #
Any help and/or advice would be appreciated.
Thank you, Chuck
mri_add_xform_to_header -c /vxfsvol/home/staff/chuck/freesurfer-test/bert/mri/transforms/talairach.xfm /vxfsvol/home/staff/chuck/freesurfer-test/bert/mri/orig.mgz /vxfsvol/home/staff/chuck/freesurfer-test/bert/mri/orig.mgz
INFO: extension is mgz #-------------------------------------------- #@# Nu Intensity Correction Fri Oct 12 10:37:03 PDT 2007
mri_nu_correct.mni --i orig.mgz --o nu.mgz --n 2
/vxfsvol/home/staff/chuck/freesurfer-test/bert/mri /usr/local/freesurfer/bin/mri_nu_correct.mni --i orig.mgz --o nu.mgz --n 2 nIters 2 $Id: mri_nu_correct.mni,v 1.3.2.1 2006/11/24 20:16:38 nicks Exp $ Linux doolin 2.6.21-gentoo-r4 #1 SMP Thu Aug 9 09:33:11 PDT 2007 i686 AMD Athlon(tm) 64 X2 Dual Core Processor 4400+ AuthenticAMD GNU/Linux Fri Oct 12 10:37:03 PDT 2007 Program nu_correct, built from: Package MNI N3, version 1.10, compiled by nicks@kani (i686-pc-linux-gnu) on 2005-11-15 at 21:02:27 tmpdir is ./tmp.mri_nu_correct.mni.28553 /vxfsvol/home/staff/chuck/freesurfer-test/bert/mri mri_convert orig.mgz ./tmp.mri_nu_correct.mni.28553/nu0.mnc mri_convert orig.mgz ./tmp.mri_nu_correct.mni.28553/nu0.mnc reading from orig.mgz... TR=0.00, TE=0.00, TI=0.00, flip angle=0.00 i_ras = (-1, 0, 0) j_ras = (0, 0, -1) k_ras = (0, 1, 0) writing to ./tmp.mri_nu_correct.mni.28553/nu0.mnc...
Iteration 1 Fri Oct 12 10:37:11 PDT 2007 nu_correct -clobber ./tmp.mri_nu_correct.mni.28553/nu0.mnc ./tmp.mri_nu_correct.mni.28553/nu1.mnc [chuck@doolin:/vxfsvol/home/staff/chuck/freesurfer-test/bert/mri/] [2007-10-12 10:37:11] running: /usr/local/freesurfer/mni/bin/nu_estimate_np_and_em -parzen -log -sharpen 0.15 0.01 -iterations 50 -stop 0.001 -shrink 4 -auto_mask -nonotify -b_spline 1 -distance 200 -quiet -execute -clobber -nokeeptmp -tmpdir /usr/tmp/nu_correct_28603/ ./tmp.mri_nu_correct.mni.28553/nu0.mnc ./tmp.mri_nu_correct.mni.28553/nu1.imp
nu_estimate_np_and_em: broken pipe nu_estimate_np_and_em: crashed while running mincinfo (termination status=13) nu_correct: crashed while running nu_estimate_np_and_em (termination status=65280) ERROR: nu_correct Linux doolin 2.6.21-gentoo-r4 #1 SMP Thu Aug 9 09:33:11 PDT 2007 i686 AMD Athlon(tm) 64 X2 Dual Core Processor 4400+ AuthenticAMD GNU/Linux
recon-all exited with ERRORS at Fri Oct 12 10:37:11 PDT 2007
chuck@doolin /usr/local/freesurfer $
Hi,
No, we have sufficient disk space on each machine.
This used to work, but started failing following upgrades to system files. I suspect libc.so.6 or libm.so.6, since mincinfo is linked to both, and both were part of the upgrade, but I don't have anything specific pointing to this as the cause.
Chuck
Bruce Fischl wrote:
Hi Chuck,
is that partition full?
Bruce
you might check the MNI list to see if this is a known issue for them.
On Fri, 12 Oct 2007, Chuck Theobald wrote:
Hi,
No, we have sufficient disk space on each machine.
This used to work, but started failing following upgrades to system files. I suspect libc.so.6 or libm.so.6, since mincinfo is linked to both, and both were part of the upgrade, but I don't have anything specific pointing to this as the cause.
Chuck
Bruce Fischl wrote:
Hi Chuck,
is that partition full?
Bruce
Chuck,
Which freesurfer is installed? The rh9 freesurfer dist is built against libstdc++.so.5 (which many systems dont have), whereas the centos freesurfer dist uses .6.
Nick
On Fri, 2007-10-12 at 14:13 -0700, Chuck Theobald wrote:
Hi,
No, we have sufficient disk space on each machine.
This used to work, but started failing following upgrades to system files. I suspect libc.so.6 or libm.so.6, since mincinfo is linked to both, and both were part of the upgrade, but I don't have anything specific pointing to this as the cause.
Chuck
Bruce Fischl wrote:
Hi Chuck,
is that partition full?
Bruce
I believe we are using the CentOS version. I am downloading the RH9 distro for version 3.0.3 to see if that will work for us.
Chuck
Nick Schmansky wrote:
Chuck,
Which freesurfer is installed? The rh9 freesurfer dist is built against libstdc++.so.5 (which many systems dont have), whereas the centos freesurfer dist uses .6.
Nick
freesurfer@nmr.mgh.harvard.edu