GATK4 BwaMemIndexImageCreator not producing errors or ".img" output

rswensonrswenson Research Triangle Park, North CarolinaMember
Hi I am running the following inside my environment:

$ gatk/gatk-4.0.12.0/gatk BwaMemIndexImageCreator -I refGenom
Homo_sapiens_assembly38.fasta -O Homo_sapiens_assembly38.fa
sta.fasta.img

This runs without seemingly generating any errors, but does not produce the index image. I also ran with --verbosity INFO and DEBUG, in which neither case, did the level of output change, in which each time, the processing ended without generating the ".img" file. Is there a way to log out or trace, besides redirecting or piping stderr,stdout?

The environment has GATK 4.0.12.0 and Java 8, and is ran inside Cloudera Data Science Workbench , in Kubernetes/Docker atop a Cloudera Cluster with Spark 2.2.0 also running Hail v0.2.

Best Answer

  • rswensonrswenson Research Triangle Park, North Carolina
    Accepted Answer
    This process is now running in the other envs, and still running, which means I expect that they will run fine since in the other env, the process exits almost immediately as it goes to construct BWT. I have this running in a node with 16GB RAM, where in the lab environment running CDSW, it seems my instance has very low memory, under 2GB RAM.

Answers

  • bhanuGandhambhanuGandham Cambridge MAMember, Administrator, Broadie, Moderator admin

    @rswenson

    You could try running this with --stacktrace.

  • rswensonrswenson Research Triangle Park, North CarolinaMember
    There does not seem to be a --stacktrace argument for gatk, do you mean the JRE stacktrace?

    Set the system property GATK_STACKTRACE_ON_USER_EXCEPTION (--java-options '-DGATK_STACKTRACE_ON_USER_EXCEPTION=true') to print the stack trace.


    A USER ERROR has occurred: stacktrace is not a recognized option is what I got running with trying to pass --stacktrace.

    Baffled at why this wouldn't run, in the first place. If I tar bzip2 the FASTA input file, it throws an error stating it cannot detect a FASTA file, ... when I run certain other pipelines, they generated the expected output in folders in the env... permissions + space are not an issue. I will now compare/try this on my local MacOS and in a Linux HPC env...

    Thanks @bhanuGandham
  • rswensonrswenson Research Triangle Park, North CarolinaMember
    Accepted Answer
    This process is now running in the other envs, and still running, which means I expect that they will run fine since in the other env, the process exits almost immediately as it goes to construct BWT. I have this running in a node with 16GB RAM, where in the lab environment running CDSW, it seems my instance has very low memory, under 2GB RAM.
Sign In or Register to comment.