I have a server with a SATA hard disk mounted on it. I created a directory on the SATA hard disk and ran a Nextflow script that is located on the server, but it failed. There was no .command.err file, but I found the following message in the .command.log : ln: failed to create symbolic link 'FAX37527_pass_barcode25_d51b1ed1_89554596_17.fastq.gz': Operation not supported . The workflow runs successfully when the project directory is on the server’s local storage. What should I do to run the workflow on a mounted hard disk?
Is your attached disk formatted with FAT32? Does it support symbolic links? This is a requirement for Nextflow and is supported by most POSIX filesystems.
Generally, I wouldn’t run a Nextflow pipeline on an attached disk, I would run it on some fast, locally available storage. You can still read the data from the remote disk, but use the faster local storage to analyse the data. Set the working directory to the local disk and/or use process.scratch to work from your internal drives.