making resume work...?

[BioinfoTongLI]

Filed here:
https://github.com/embl-cba/ome-zarr-image-analysis-nextflow/issues/8

But let's continue the discussion here.
@krokicki @bugraoezdemir @sebgoti @tischi (somehow I cannot christian) @folterj

[tischi]

Thanks, @tischi is correct.

Due you think that resume would work without saving any file in the work dir?

[BioinfoTongLI]

Not for now. I am curiose what @krokicki's solution here. Do you use groovy scripting for resuming?

[krokicki]

One way to make resume work here is to use process.cache = 'lenient' in the nextflow.config file. This ignores the changed date stamps on files and only checks to make sure the files are there and correct size.

I typically don't use resume because we haven't found a way to make it work with Spark/Dask clusters. Instead, I have a --skip parameter and just add the steps I want to skip. So --skip=blur,segment would skip the first two steps and just run morphometry. Since all the data is outside the work dir, this is effectively resuming the processing from a given point.