Please make an MCVE as mentioned in: https://savannah.gnu.org/bugs/?65017#comment1

and

https://www.gnu.org/software/parallel/man.html#reporting-bugs

If $NUMBER_JOBS is unset, it will evaluate to nothing. Then your example will use '-j nice' which "evaluate the string 'nice' as a perl expressing" which gives 0.

The reason you rarely see GNU Parallel use jobslot > 30 is that it will only use jobslot 31 when 1..30 are all in use. And if your jobs run pretty fast, then job 1 may be done before job 30 is even started.

Hi!

I am running the following command:

find $INPUT_DIR -type f -name "*.zip" | parallel -N $FILES_JOB --jobs $NUMBER_JOBS nice -n $NICE \
      'python3 my_script.py -i {} -w $(printf "%04d" "{%}")'


$INPUT_DIR is a local folder
$NUMBER_JOBS is equals to 30
$NICE is equals to 0

I am passing {%} to my_script.py, which performs some work on the files passed to it, and produces the output in a folder name, based in the parameter -w which receives the value {%} padded with 0s.

The issue comes because, suddenly and only for a minimum amount of time, the script starts receiving slot numbers bigger than the number of jobs.

Please, find below the folders created by my script, based on the value of the parameter -w.

In can be observerd that it creates folders bigger than the value 30, which should be the maximum slot number, based on the parameters above.

0001
0002
0003
0004
0005
0006
0007
0008
0009
0010
0011
0012
0013
0014
0015
0016
0017
0018
0019
0020
0021
0022
0023
0024
0025
0026
0027
0028
0029
0030
0040
0042
0043
0044
0045
0048
0049
0055
0058
0059
0061
0062
0064
0065
0066
0068
0069
0071
0073
0074
0075
0081
0083
0086
0088
0089
0090
0092
0094

Hope it helps!