This should not happen. Contact <address@hidden>. at parallel line 3384.

[Sebastian Eiser]

Hi all,
I'm new to using "parallel", but so far I made very good experience
with it on our 16 core machine. Thank you very much for this useful
tool! I found an interesting issue this week when using GNU parallel
20110205, which results in the error/warning given in the subject
line.

This is the cmd I'm running:

$ for year in 2005 2006 2007 2008 2009; do echo $year; find $year/
-type f -name *.he5 | parallel --eta -j+0
/home/seiser/src/gtproduct-omi-old netcdf -m \'O3.COLUMN\' -fiv
\'O3.COLUMN\' -f
\'"CLOUD.COVER<0.2;ANGLE.SOLAR_ZENITH<80;FLAG.QUALITY.PROCESSING=0"\'
-fe \'FLAG.QUALITY.PROCESSING\' -o
~/comparisons/nadir/conv_omi/{/.}.nc {}; done

using --dry-run, everything looks good for _all_ inputs.
e.g. one line:
/home/seiser/src/gtproduct-omi-old netcdf -m 'O3.COLUMN' -fiv
'O3.COLUMN' -f 'CLOUD.COVER<0.2;ANGLE.SOLAR_ZENITH<80;FLAG.QUALITY.PROCESSING=0'
-fe 'FLAG.QUALITY.PROCESSING' -o
/home/seiser/comparisons/nadir/conv_omi/OMI-Aura_L2-OMDOAO3_2005m0109t1600-o02598_v003-2007m1204t191129.nc
2005/009/OMI-Aura_L2-OMDOAO3_2005m0109t1600-o02598_v003-2007m1204t191129.he5

Unfortunately I don't know more details about the error, but from the
output I found a few hints
1. some file names from "{}" seem to get f*cked up:
ERROR: could not find
5/114/OMI-Aura_L2-OMDOAO3_2005m0424t1424-o04126_v003-2007m1204t210249.he5
ERROR: could not find 25-o07271_v003-2007m1206t180121.he5
ERROR: could not find 2.he5
(this is output from the gtproduct-omi-old programme if it doesn't
find the input file)
Using --dry-run this does not occur, all {}s are substituted
correctly. I initially thought the file names are too long, but using
shorter file names didn't help.
2. then, I sometimes get this warning:
Use of uninitialized value in hash element at
/home/seiser/alt/usr/bin/parallel line 3284, <STDIN> line 2542
Use of uninitialized value in hash element at
/home/seiser/alt/usr/bin/parallel line 3287, <STDIN> line 2542
3. and finally, it tells me this (without exiting):
ETA: 1670s 3854left 0.43avg  local:16/1131/100%/0.4s This should not
happen. Contact <parallel@gnu.org>. at parallel line 3384.

It seems that the problem doesn't occur if I reduce the number of
inputs, e.g. not using 5000 files, but just 1000 seems to be fine.
Then, just hint 1. and 2. occur.

Please let me know what you think or what I did wrong. Thanks in
advance! Should you need more information from the output or other
additional diagnostics, let me know and I will supply them quickly.
-Sebastian