Hello,
I have been trying to run several versions of the same GLM analysis, trying out different parameters to see if they affect the results. For instance, I am running the GLM with different levels of smoothing, per-session versus per-run motion correction, etc. To speed this up, I am trying to do these GLMs in parallel. However, when I do this, I run into a "stale file handle" problem. As near as I can tell, this occurs because the analyses try to read and write to the same log files at the same time. A typical error message is "/net/rcss2/srv/export/ncf_vcn/share_root/jtaylor/mri-space/studies/fpoinfips01/scripts/glm/log/preproc-160718_xu_loc01_sub02-bold.log: Stale file handle". I have tried putting "-nolog" in the call to selxavg3-sess, but this doesn't prevent the creation of log files when it makes the sub-calls to the various pre-processing functions. Is there any way around this? It would be rather slow to try to run the GLMs sequentially, so I am wondering if anybody has figured out a way around this problem.
Much obliged,
JohnMark