Skip to content

Pipeline fails because of huge data output #42

Description

@adriandrr

The pipeline seems to create an enormous amount of output. Therefore, with only a few sample inputs, the pipeline fails because of no disk space left (1.5 TB) ~20 samples a ~8Gb per Read (2x)

  1. Idea: Is it possible to remove intermediate results earlier? I, for example, noticed that fastq.gz files get unzipped and remain unzipped throughout a specific part of the pipeline. Also, once the fastq.gz files are trimmed, can't the untrimmed files be removed? I bet there are a lot of similar space optimization options.

  2. Idea: Would it be possible for the pipeline to estimate the largest amount of output at any given moment, with respect to the input datasets? The user could then decide if a reduction of input would be useful

  3. Idea: Could there be a mode implemented (and maybe also defaulted) where disk space is treated dynamically in a way that only so many samples are processed at the same time, so that at every point, not more than 80% diskspace is used?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions