Mounts
Fuzzball locates or creates the volumes referenced in your workflow specification YAML as one of the first stages in your workflow. It then mounts the requested volumes to your containers as your jobs run. You can mount the same volume in more than one job. Successive jobs therefore read and write the same data, and each step leaves its results in place for the next one.
A job’s mounts block is keyed by the path inside the container, and its value names a volume
declared in the workflow’s volumes block:
jobs:
analyze:
mounts:
/data: my-volume
Mounting is the only thing a mount declares – there is no read-only option. To mount the same
volume into every job, declare it once under defaults.job.mounts instead of repeating it. See
the Fuzzfile syntax guide for the full grammar.
V4 requires this path-keyed form. The V1 form keyed by volume name (my-volume: {location: /data}) is rejected in a workflow markedversion: v4.