External Email - Use Caution        

Hello John,

You might be able to tell from the log what directory recon-all was cd'd into when the creation of the soft link failed.  You could try to cd there as your userid (which I assume you used to run recon-all), and see if you can manually create any soft link there.  if you cannot do so, there could be a problem with the directory permissions.  The uersid running recon-all should have permission to read, write, and execute on subdirectories which typically means the permission on them is set to be 755.

While do not officially support cuda, I would first check to see you have the cuda libraries installed on you system.  You could do a search under /usr to see if you have it, e.g.
$ sudo find /usr -name "libcudart*"

- or with the package tools, look to see if any nvidia or cuda packages are installed,  e.g.,
$ rpm -qa | grep -i nvid
$ rpm -qa | grep -i cuda

If you don't find a libcudart.so.5 installed, then I think you will need to install rev 7.5 of the cuda tools on your system as that should be the version with libcudart.so.5

- R.

John Absher
June 22, 2019 at 22:28
External Email - Use Caution

Hi, I am testing a new installation of freesurfer. Two errors occurred, and I'm hoping someone can help me troubleshoot:

First, recon-all exits with error indicating that it is unable to create a symbolic link:

[GHS\026278@vmcntdsktp10 fs_sample1]$ tail /neuro_images_isilon/026278/fs_sample1/sample001/scripts/recon-all.log
ICI = 216.7, FI = 1989.1, variation=32373.324
writing Gaussian curvature to ./lh.white.preaparc.K...done.
writing mean curvature to ./lh.white.preaparc.H...done.
rm -f lh.white.H
ln -s lh.white.preaparc.H lh.white.H
Linux vmcntdsktp10 3.10.0-957.21.3.el7.x86_64 #1 SMP Tue Jun 18 16:35:19 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux

recon-all -s sample001 exited with ERRORS at Sat Jun 22 17:10:41 EDT 2019

To report a problem, see http://surfer.nmr.mgh.harvard.edu/fswiki/BugReporting

The error occurs in CentOS 7 running on a virtual machine environment, with the latest stable freesurfer release, 6.0.

The recon-all process was being run on the sample_001.mgz file to test the installation, and ran for about 7 hours.

Second, when I tried to use the -use-gpu flag, the process seems to be missing a library:

[GHS\026278@vmcntdsktp10 fs_sample1]$ recon-all -subject sample001 -i ./sample-001.mgz -use-gpu -all
Testing for CUDA device:
/usr/local/freesurfer/bin/mri_em_register_cuda: error while loading shared libraries: libcudart.so.5.0: cannot open shared object file: No such file or directory
Linux vmcntdsktp10 3.10.0-957.21.3.el7.x86_64 #1 SMP Tue Jun 18 16:35:19 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux

recon-all -s sample001 exited with ERRORS at Sat Jun 22 10:05:12 EDT 2019

For more details, see the log file
To report a problem, see http://surfer.nmr.mgh.harvard.edu/fswiki/BugReporting

[GHS\026278@vmcntdsktp10 fs_sample1]$

Can you suggest workarounds for one or both of these errors? Thanks,

John


Warm regards,


Please, note my new email address, john.absher@prismahealth.org

John R. Absher, MD, FAAN
Division of Neurology, Department of Medicine

Prisma Health-Upstate
200 Patewood Drive, Suite 350B
City, SC 29615
864-454-4500 (office)
864-350-6655 (mobile)
864-454-4505 (fax)



Inspire health. Serve with compassion. Be the difference.




----------------------------------------------------------------------
CONFIDENTIALITY NOTICE
This message is intended only for the use of the individual or entity to which it is addressed and may contain information that is privileged, confidential and exempt from disclosure under applicable law. If the reader of this message is not the intended recipient, you are hereby notified that any dissemination, distribution, or copying of this communication is strictly prohibited by law. If you have received this communication in error, please notify me immediately.


_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard..edu/mailman/listinfo/freesurfer