I'm running a Snakemake RNA-seq pipeline (FastQC → fastp → STAR → samtools → Salmon) inside Docker on Ubuntu 22.04.

The workflow runs correctly for many hours, but eventually one or more fastp jobs stop making progress. The sample is different every time.

When this happens:

• fastp remains running indefinitely

• CPU usage drops to ~2–3%

• Process state is SL+

• Snakemake waits forever because the job never exits

However, running the exact same fastp command manually on the same FASTQ files finishes in ~6 minutes.

I'm running:

snakemake --cores 16 (previously --cores 24)

with:

• fastp: 4 threads (previously 8)

• STAR: 8 threads

The data are stored on a local NVMe drive and bind-mounted into the Docker container. Disk space and RAM are not limiting.

Has anyone seen fastp hang only when executed repeatedly/concurrently under Snakemake? Is this more likely to be a Snakemake scheduling issue, a fastp issue, or something else?