Hi Doug,
After reading your most recent reply, I just found that many of previous freesurfer emails had gone to my junk box automatically, so I missed some of your replies. Sorry!
Looking at the strace output I generated last time, here are the lines where the program tries to stat and mkdir something in /scratch. And, the prev execve() system call.
So, I'm guessing that bin/segmentSubjectT1_autoEstimateAlveusML or something it in turn runs is leading to "/scratch" directory problem.
========================= 13174 execve("/cell_root/software/freesurfer/dev/sys/bin/segmentSubjectT1_autoEstimateAlveusML", ["/cell_root/software/freesurfer/d"..., "HMN042", "/export/extra/pkd-fstest/FG_Temp", "0.333333333333333333333333333333"..., "/cell_root/software/freesurfer/d"..., "/cell_root/software/freesurfer/d"..., "/cell_root/software/freesurfer/d"..., "0.05", "left", "ConjGrad", "v10", "/cell_root/software/freesurfer/d"..., "0"], [/* 86 vars */]) = 0 13182 stat("/scratch/", {st_mode=S_IFDIR|0555, st_size=512, ...}) = 0 13182 stat("/scratch/HMN042_hippoSF_T1_v10_left/", 0x7f4e226015b0) = -1 ENOENT (No such file or directory) 13182 stat("/scratch/HMN042_hippoSF_T1_v10_left/", 0x7f4e226015e0) = -1 ENOENT (No such file or directory) 13182 lstat("/scratch", {st_mode=S_IFDIR|0555, st_size=512, ...}) = 0 13182 lstat("/scratch/HMN042_hippoSF_T1_v10_left", 0x7f4e226004b0) = -1 ENOENT (No such file or directory) 13182 stat("/scratch", {st_mode=S_IFDIR|0555, st_size=512, ...}) = 0 13182 mkdir("/scratch/HMN042_hippoSF_T1_v10_left", 0775) = -1 EROFS (Read-only file system) =========================
I do see, however, MCRv80 references to before the mkdir("/scratch/HMN042_hippoSF_T1_v10_left" call. So, it could be something in the MCRv80 addition that's causing this issue.
And, things like this: ========================= 13182 open("/mathworks/BLR/devel/bat/A3p1/build/3p/install/319992/glnxa64/ITK/share/gdcm/gdcm.dic", O_RDONLY) = -1 ENOENT (No such file or directory) 13182 open("/mathworks/BLR/devel/bat/A3p1/build/3p/install/319992/glnxa64/ITK/share/gdcm/dicomVR.dic", O_RDONLY) = -1 ENOENT (No such file or directory) 13182 open("/mathworks/BLR/devel/bat/A3p1/build/3p/install/319992/glnxa64/ITK/share/gdcm/dicomTS.dic", O_RDONLY) = -1 ENOENT (No such file or directory) 13182 open("/mathworks/BLR/devel/bat/A3p1/build/3p/install/319992/glnxa64/ITK/share/gdcm/DictGroupName.dic", O_RDONLY) = -1 ENOENT (No such file or directory) 13182 open("/mathworks/BLR/devel/bat/A3p1/build/3p/install/319992/glnxa64/ITK/share/gdcm/DicomDir.dic", O_RDONLY) = -1 ENOENT (No such file or directory) =========================
are probably more MCRv80 (matlab runtime) package related and referencing a /mathworks/BLR directory on the machine that it was built, but not on our systems.
Please let me know if you need more information. I will keep my eyes wide open this time.
Thanks!!
--------------------------------------------------------------------------------------------------------------------------------------- Hi Fengji, you have sent several emails. I keep asking you to give us a specific binary that has the problem, but you seem to ignore my request, wait a couple of weeks, then post the question back to the list. I'm happy to take look, but please send me something more specific! thanks doug
----- Fengji
________________________________ From: Fengji Geng Sent: Thursday, January 14, 2016 1:46 PM To: freesurfer@nmr.mgh.harvard.edu Subject: FW: [Freesurfer] FW: problem with hard coded directory (/scratch)
Hi FS's experts, I sent out a couple of emails to seek for help on an error from running recon-all using freesurfer 6.0 beta. Could anyone help with the question? Thank you!
----- Fengji
________________________________ From: Fengji Geng Sent: Thursday, January 07, 2016 8:15 AM To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] FW: problem with hard coded directory (/scratch)
Hi Douglas, You asked us to give you an example of the problem program. I will try to do it here, but if it is not what you want, please let me know.
we ran recon-all -all -s <subject_name> -hippocampal-subfields-T1T2 <file name of additional scan> <analysisID>. We saw the program tried to create /scratch/$subject_hippoSF_T1_v10_right directory and failing with (Read-only file system) because the user does not have permission to write to /scratch.
I can also send you the strace output if needed.
Thank you!
--------------------------------------------------------------------------------- Message: 4 Date: Mon, 4 Jan 2016 13:37:48 -0500 From: Douglas N Greve greve@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] FW: problem with hard coded directory (/scratch) To: freesurfer@nmr.mgh.harvard.edu Message-ID: 568ABBFC.5000702@nmr.mgh.harvard.edu Content-Type: text/plain; charset=windows-1252; format=flowed
Can you give us specific examples of problem programs?
On 01/04/2016 09:54 AM, Fengji Geng wrote:
Hello Bruce,
Happy New Year!
After we reported the error related to '/scratch' several weeks ago (see the email attached below), we tested the newest version of freesurfer 6.0 recently, it looks likes that a lot of "if (-e /scratch)" have been changed to "if (-dw /scratch)". However, there are still a lot of binaries in the bin directory that match the string "scratch", which cause the same error that I encountered before. I am wondering if it is possible for you to fix '/scratch' in the bin directory.
Thank you!
----- Fengji
That all looks like hippocampal subfield stuff, which is Eugenio's realm.
On 01/15/2016 11:41 AM, Fengji Geng wrote:
Hi Doug,
After reading your most recent reply, I just found that many of previous freesurfer emails had gone to my junk box automatically, so I missed some of your replies. Sorry!
Looking at the strace output I generated last time, here are the lines where the program tries to stat and mkdir something in/scratch. And, the prev execve() system call.
So, I'm guessing that bin/segmentSubjectT1_autoEstimateAlveusML or something it in turn runs is leading to "/scratch" directory problem.
========================= 13174 execve("/cell_root/software/freesurfer/dev/sys/bin/segmentSubjectT1_autoEstimateAlveusML", ["/cell_root/software/freesurfer/d"..., "HMN042", "/export/extra/pkd-fstest/FG_Temp", "0.333333333333333333333333333333"..., "/cell_root/software/freesurfer/d"..., "/cell_root/software/freesurfer/d"..., "/cell_root/software/freesurfer/d"..., "0.05", "left", "ConjGrad", "v10", "/cell_root/software/freesurfer/d"..., "0"], [/* 86 vars */]) = 0 13182 stat("/scratch/", {st_mode=S_IFDIR|0555, st_size=512, ...}) = 0 13182 stat("/scratch/HMN042_hippoSF_T1_v10_left/", 0x7f4e226015b0) = -1 ENOENT (No such file or directory) 13182 stat("/scratch/HMN042_hippoSF_T1_v10_left/", 0x7f4e226015e0) = -1 ENOENT (No such file or directory) 13182 lstat("/scratch", {st_mode=S_IFDIR|0555, st_size=512, ...}) = 0 13182 lstat("/scratch/HMN042_hippoSF_T1_v10_left", 0x7f4e226004b0) = -1 ENOENT (No such file or directory) 13182 stat("/scratch", {st_mode=S_IFDIR|0555, st_size=512, ...}) = 0 13182 mkdir("/scratch/HMN042_hippoSF_T1_v10_left", 0775) = -1 EROFS (Read-only file system) =========================
I do see, however, MCRv80 references to before the mkdir("/scratch/HMN042_hippoSF_T1_v10_left" call. So, it could be something in the MCRv80 addition that's causing this issue.
And, things like this:
13182 open("/mathworks/BLR/devel/bat/A3p1/build/3p/install/319992/glnxa64/ITK/share/gdcm/gdcm.dic", O_RDONLY) = -1 ENOENT (No such file or directory) 13182 open("/mathworks/BLR/devel/bat/A3p1/build/3p/install/319992/glnxa64/ITK/share/gdcm/dicomVR.dic", O_RDONLY) = -1 ENOENT (No such file or directory) 13182 open("/mathworks/BLR/devel/bat/A3p1/build/3p/install/319992/glnxa64/ITK/share/gdcm/dicomTS.dic", O_RDONLY) = -1 ENOENT (No such file or directory) 13182 open("/mathworks/BLR/devel/bat/A3p1/build/3p/install/319992/glnxa64/ITK/share/gdcm/DictGroupName.dic", O_RDONLY) = -1 ENOENT (No such file or directory) 13182 open("/mathworks/BLR/devel/bat/A3p1/build/3p/install/319992/glnxa64/ITK/share/gdcm/DicomDir.dic", O_RDONLY) = -1 ENOENT (No such file or directory) =========================
are probably more MCRv80 (matlab runtime) package related and referencing a /mathworks/BLR directory on the machine that it was built, but not on our systems.
Please let me know if you need more information. I will keep my eyes wide open this time.
Thanks!!
Hi Fengji, you have sent several emails. I keep asking you to give us a specific binary that has the problem, but you seem to ignore my request, wait a couple of weeks, then post the question back to the list. I'm happy to take look, but please send me something more specific! thanks doug
Fengji
*From:* Fengji Geng *Sent:* Thursday, January 14, 2016 1:46 PM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* FW: [Freesurfer] FW: problem with hard coded directory (/scratch)
Hi FS's experts, I sent out a couple of emails to seek for help on an error from running recon-all using freesurfer 6.0 beta. Could anyone help with the question? Thank you!
Fengji
*From:* Fengji Geng *Sent:* Thursday, January 07, 2016 8:15 AM *To:* freesurfer@nmr.mgh.harvard.edu *Subject:* Re: [Freesurfer] FW: problem with hard coded directory (/scratch)
Hi Douglas, You asked us to give you an example of the problem program. I will try to do it here, but if it is not what you want, please let me know.
we ran recon-all -all -s <subject_name> -hippocampal-subfields-T1T2 <file name of additional scan> <analysisID>. We saw the program tried to create /scratch/$subject_hippoSF_T1_v10_right directory and failing with (Read-only file system) because the user does not have permission to write to /scratch.
I can also send you the strace output if needed.
Thank you!
Message: 4 Date: Mon, 4 Jan 2016 13:37:48 -0500 From: Douglas N Greve greve@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] FW: problem with hard coded directory (/scratch) To: freesurfer@nmr.mgh.harvard.edu Message-ID: 568ABBFC.5000702@nmr.mgh.harvard.edu Content-Type: text/plain; charset=windows-1252; format=flowed
Can you give us specific examples of problem programs?
On 01/04/2016 09:54 AM, Fengji Geng wrote:
Hello Bruce,
Happy New Year!
After we reported the error related to '/scratch' several weeks ago (see the email attached below), we tested the newest version of freesurfer 6.0 recently, it looks likes that a lot of "if (-e /scratch)" have been changed to "if (-dw /scratch)". However, there are still a lot of binaries in the bin directory that match the string "scratch", which cause the same error that I encountered before. I am wondering if it is possible for you to fix '/scratch' in the bin directory.
Thank you!
Fengji
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Douglas N Greve -
Fengji Geng