It's a separate issue from failed download, the file GCA_000001635.9_GRCm39_genomic.fna.gz is downloaded completely, however, when running same pipeline again or different stages (i.e. pre or reads) it will start download again, resulting in multiple identical files and extended running times.