[MINC-users] Running MINC tools from the www user

Mon Apr 21 15:08:47 EDT 2008

Although I haven't moved an inch on solving the issue (thanks Alex for the
help), here's more information on my problem:

1) If I type type the following command in a terminal window, everything
works fine:
/usr/local/minc2/bin/nu_correct -clobber -mapping_dir /Library/MEDICS/tmp
-tmpdir /Library/MEDICS/tmp/ -iterations 100 -shrink 3 -fwhm 0.1 -distance
100 -verbose /Library/MEDICS/tmp/abc1.mnc /Library/MEDICS/tmp/abc.mnc

2) If I precede the previous command with sudo -u _www, I get an error:
sudo -u _www /usr/local/minc2/bin/nu_correct -clobber -mapping_dir
/Library/MEDICS/tmp -tmpdir /Library/MEDICS/tmp/ -iterations 100 -shrink 3
-fwhm 0.1 -distance 100 -verbose /Library/MEDICS/tmp/abc1.mnc
/Library/MEDICS/tmp/abc.mnc

[nu_correct] [_www at pita.crulrg.local:/usr/local/minc2/lib/perl5/site_perl]
[2008-04-21 14:44:23] /usr/local/minc2/bin/nu_estimate_np_and_em -parzen
-log -sharpen 0.1 0.01 -iterations 100 -stop 0.001 -shrink 3 -auto_mask
-nonotify -b_spline 1 -distance 100 -verbose -execute -clobber -nokeeptmp
-tmpdir /Library/MEDICS/tmp/ /Library/MEDICS/tmp/abc1.mnc
/Library/MEDICS/tmp/abc.imp

Uncorrected volume (input): /Library/MEDICS/tmp/abc1.mnc
Intensity mapping (output): /Library/MEDICS/tmp/abc.imp

# Start of NU estimation algorithm
[nu_estimate_np_and_em]
[_www at pita.crulrg.local:/usr/local/minc2/lib/perl5/site_perl]
[2008-04-21 14:44:23] /usr/local/minc2/bin/mincresample -clobber -verbose
-nearest_neighbour -nelements 20 30 30 -step 10.8 -9.14058 -9.14058
/Library/MEDICS/tmp/abc1.mnc /Library/MEDICS/tmp/abc1___.mnc
Transforming slices:....................Done
[nu_estimate_np_and_em]
[_www at pita.crulrg.local:/usr/local/minc2/lib/perl5/site_perl]
[2008-04-21 14:44:23] /usr/local/minc2/bin/mincmath -clobber -verbose -clamp
-const2 1 1.7e+308 /Library/MEDICS/tmp/abc1___.mnc
/Library/MEDICS/tmp/abc_log.mnc.temp
Processing:....................Done
[nu_estimate_np_and_em]
[_www at pita.crulrg.local:/usr/local/minc2/lib/perl5/site_perl]
[2008-04-21 14:44:23] /usr/local/minc2/bin/mincmath -clobber -verbose -zero
-log /Library/MEDICS/tmp/abc_log.mnc.temp /Library/MEDICS/tmp/abc_log.mnc
Processing:....................Done
[nu_estimate_np_and_em]
[_www at pita.crulrg.local:/usr/local/minc2/lib/perl5/site_perl]
[2008-04-21 14:44:23] /bin/rm /Library/MEDICS/tmp/abc_log.mnc.temp
[nu_estimate_np_and_em]
[_www at pita.crulrg.local:/usr/local/minc2/lib/perl5/site_perl]
[2008-04-21 14:44:23] /usr/local/minc2/bin/mincinfo -attvalue
xspace:spacetype -attvalue yspace:spacetype -attvalue zspace:spacetype
/Library/MEDICS/tmp/abc1___.mnc
[nu_estimate_np_and_em]
[_www at pita.crulrg.local:/usr/local/minc2/lib/perl5/site_perl]
[2008-04-21 14:44:23] /usr/local/minc2/bin/volume_stats -quiet -biModalT
/Library/MEDICS/tmp/abc1___.mnc
Assertion failed at line 742 in file templates/CachedArray.cc
nu_estimate_np_and_em: crashed while running volume_stats (termination
status=256)
nu_correct: crashed while running nu_estimate_np_and_em (termination
status=256)

3) The directory /Library/MEDICS/tmp is writable by everyone (mode 777). The
owner of the directory is _www.

4) Temporary files abc1_.mnc, abc1__.mnc, abc1___.mnc and abc_log.mnc all
get created so I don't suspect a permission problem being the case here.

5) I don't quite understand the MINC source code just yet (hey, that's a lot
of code), where would I find CachedArray.cc?

6) I noticed that in volumeStats.cc, there is the following:

    if (args.cached)
      voxels = new CachedArray<float>(0);
    else
      voxels = new SimpleArray<float>(0);

Is there a way I could just bypass the call to CachedArray and not bother
with trying to save memory (after all, GBytes are cheaper than the time
we're spending to try to make this work... ;)

Again, many thanks,

Burt.

On Wed, Apr 16, 2008 at 5:01 PM, Burt Crépeault <
burt.crepeault at crulrg.ulaval.ca> wrote:

> Hi Alex,
>
> I went ahead and tried your suggestions, unfortunately I'm still stuck. The /tmp folder still has plenty of space and it is writable by anyone. I still get the following error from running the process:
>
>
>    [7]:     [_www at pita.crulrg.local:/Library/WebServer/Documents/] [2008-04-16 12:32:14] running:
>    [8]:       /usr/local/minc2/bin/nu_estimate_np_and_em -parzen -log -sharpen 0.1 0.01 -iterations 100
>                 -stop 0.001 -shrink 3 -auto_mask -nonotify -b_spline 1 -distance 100 -quiet -execute
>
>                 -clobber -nokeeptmp -tmpdir /tmp/nu_correct_1361/
>                 '/Library/MEDICS/data/images/10188/bl/MR.T1.MP-RAGE/1004/Library.MEDICS.data.images.10188.bl.MR.T1.MP-RAGE.1004.1.10004.[mnc.d2m].Br.12_p2.mnc'
>
>                 '/Library/MEDICS/data/images/10188/bl/MR.T1.MP-RAGE/1007/Library.MEDICS.data.images.10188.bl.MR.T1.MP-RAGE.1007.1.10007.[mnc.d2m+mnc.nuc].Br.22_p2.imp'
>    [9]:
>    [10]:     Assertion failed at line 742 in file templates/CachedArray.cc
>
>    [11]:     nu_estimate_np_and_em: crashed while running volume_stats (termination status=256)
>    [12]:     nu_correct: crashed while running nu_estimate_np_and_em (termination status=256)
>
> I've listed my environment variables below:
>
>
>        [TMPDIR] => /tmp
>        [FOO] => /bar
>        [PERL5LIB] => :/usr/local/minc2/lib:/Library/Perl/5.8.8:/usr/local/minc2/lib/perl5/site_perl/:/Users/burt/lib/perl
>        [LD_LIBRARY_PATH] => :/Applications/MATLAB72/dip/Darwin/lib
>
>        [PATH] => /usr/bin:/bin:/usr/sbin:/sbin:/usr/local/minc2/bin:/opt/local/bin:/opt/local/sbin:/bin:/usr/bin:/usr/bin/X11:/usr/X11R6/bin:/usr/local/bin:/sbin:/usr/sbin:/usr/local/mysql/bin
>        [PWD] => /Library/WebServer/Documents
>
>        [SHLVL] => 1
>        [DYLD_LIBRARY_PATH] => :/usr/local/minc2/external-libs
>
> I tried the command at line [8] of my log directly in a terminal and everything runs fine. There is a /tmp/nu_correct_xxxx dir that is being created. However, that directory does not appear when the process is launched from the web app, therefore your assumption that CachedArray.cc is unable to create the temp file (or more likely dir) appears to be correct.
>
>
> This is where I'm stuck at the moment, any other ideas?
>
>
> --
> Burt Crépeault
> Research Assistant
> Centre de Recherche de l'Université Laval - Robert-Giffard
> 418-663-5741, ext 6844

-- 
Burt Crépeault
Research Assistant
Centre de Recherche de l'Université Laval - Robert-Giffard
418-663-5741, ext 6844