galaxyproject / pulsar

Distributed job execution application built for Galaxy
https://pulsar.readthedocs.io
Apache License 2.0
37 stars 50 forks source link

Freebayes bam index linking to originating server location not job inputs #169

Closed Slugger70 closed 1 year ago

Slugger70 commented 6 years ago

I'm not sure if this is a Freebayes issue or if it's a deeper problem with Pulsar but I've put it here for comment.

When freebayes makes the links to the bam file and the index, for some reason when running on Pulsar the index gets linked to the originating Galaxy server metadata file location and not the inputs file directory on Pulsar

The two links in the pulsar working directory are:

b_0.bam -> /mnt/pulsar/files/staging/186756/inputs/dataset_290340.dat
b_0.bam.bai -> /mnt/galaxy/files/_metadata_files/021/metadata_21453.dat

where /mnt/galaxy is the location of Galaxy on the originating server.

The offending code in the freebayes.xml file:

#for $bam_count, $input_bam in enumerate( $input_bamfiles ):
        ln -s -f '${input_bam}' 'b_${bam_count}.bam' &&
        ln -s -f '${input_bam.metadata.bam_index}' 'b_${bam_count}.bam.bai' &&
#end for

Both the bam file and the index file get copied by pulsar to job's inputs directory.

jmchilton commented 5 years ago

This is almost certainly not a Freebayes problem.

mvdbeek commented 1 year ago

This is supposed to be fixed by https://github.com/galaxyproject/galaxy/pull/8625. If we setup the galaxy framework tests within the pulsar CI to target a pulsar without shared filesystem we will know for sure.