Poredb is a way to manage very large Oxford Nanopore datasets on the basic principle:
-- track the files, rather than moving them around
poredb create myexperiment.db
find . -name "*.fast5" > filelist.txt
poredb import myexperiment.db filelist.txt
poredb fastq myexperiment.db > myexperiment.fastq
Andreas Klosterman for dask suggestion.