Closed ehj000 closed 1 year ago
Does the file access attributes look ok?
Are you thinking on building a new versjon manually yourself to "our own " cvmfs , and present it to the tool in its own loc file ?
The dataases On the global cvmfs Were not built as part of any installation process on Our servers . They were preprocessed and provided by others. Maybe a quick kraken2 - build job on your l ocal Computer / Server can give a clue to if your hypothesis is correct by comparring the outputs ?
Might be some issues with the cvmfs. I tested mapping of reads using BWA against the e. coli reference:
[E::bwa_idx_load_from_disk] fail to locate the index files
Location of index: /cvmfs/data.galaxyproject.org/managed/bwa_mem_index/eschColi_K12/eschColi_K12.fa
Yes, it seems the "/cvmfs/data.galaxyproject.org/" directory is not available on the slurm node.
Hmm. Can we make it available or do we need to host the indexes locally? This will only resolve the problem for this particular tool
This is a problem that affects all tools that use data from "/cvmfs/data.galaxyproject.org/", so we have to make it available again. It worked before, so I don't know why it suddenly has gone missing now.
The Kraken2 tools gives the following error: kraken2: database ("/cvmfs/data.galaxyproject.org/managed/kraken2_databases/2020-11-26T021706Z_standard_kmer-len_35_minimizer-len_31_minimizer-spaces_6_load-factor_0.7") does not contain necessary file taxo.k2d
I have checked the server and taxo.k2d is on this location. I suspect that the database was not built properly upon installation of the tool, but not sure how to re-build the database on the galaxy server (kraken2-build --standard on a "normal" installation)