1 $base: "http://arvados.org/cwl#"
3 - name: RunInSingleContainer
6 Indicates that a subworkflow should run in a single container
7 and not be scheduled as separate steps.
11 doc: "Always 'arv:RunInSingleContainer'"
23 local_output_dir: Use regular file system local to the compute node.
24 There must be sufficient local scratch space to store entire output;
25 specify this with `outdirMin` of `ResourceRequirement`. Files are
26 batch uploaded to Keep when the process completes. Most compatible, but
27 upload step can be time consuming for very large files.
29 keep_output_dir: Use writable Keep mount. Files are streamed to Keep as
30 they are written. Does not consume local scratch space, but does consume
31 RAM for output buffers (up to 192 MiB per file simultaneously open for
32 writing.) Best suited to processes which produce sequential output of
33 large files (non-sequential writes may produced fragmented file
34 manifests). Supports regular files and directories, does not support
35 special files such as symlinks, hard links, named pipes, named sockets,
39 - name: RuntimeConstraints
42 Set Arvados-specific runtime hints.
46 doc: "Always 'arv:RuntimeConstraints'"
53 Size of file data buffer for Keep mount in MiB. Default is 256
54 MiB. Increase this to reduce cache thrashing in situations such as
55 accessing multiple large (64+ MiB) files at the same time, or
56 performing random access on a large file.
60 Preferred backing store for output staging. If not specified, the
61 system may choose which one to use.
63 - name: PartitionRequirement
66 Select preferred compute partitions on which to run jobs.
73 - name: APIRequirement
76 Indicates that process wants to access to the Arvados API. Will be granted
77 limited network access and have ARVADOS_API_HOST and ARVADOS_API_TOKEN set
82 doc: "Always 'arv:APIRequirement'"