MRCIEU / opengwas-reports

Report module for IEU GWAS pipeline
1 stars 0 forks source link

docker runtime error #20

Closed mcgml closed 4 years ago

mcgml commented 4 years ago

@YiLiu6240 I'm having trouble running render_gwas_report.R using Docker:

command using commit 177aa4b3170d984c704384cc47309162edc4fbe9:

docker run \
--rm \
-v ${MountDir}:${MountDir} \
--cpus="1" \
mrbase-report-module:177aa4b3170d984c704384cc47309162edc4fbe9 \
render_gwas_report.R \
${VcfFileIn} \
--output_dir ${OutputDir} \
--n_cores 1

Output:

docker: Error response from daemon: OCI runtime create failed: container_linux.go:345: starting container process caused "exec: \"render_gwas_report.R\": executable file not found in $PATH": unknown.
YiLiu6240 commented 4 years ago

Could you try using "Rscript render_gwas_report.R" instead of just "render_gwas_report.R"? I removed the working directory from PATH and that could have caused this.

From your example there are other things that could cause problems, e.g. -v ${MountDir}:${MountDir} (see https://github.com/MRCIEU/mrbase-report-module/tree/177aa4b3170d984c704384cc47309162edc4fbe9#docker). At its current functionalities it still requires a reference dataset.

It is not needed to set --n_cores 1 as it is now the default. Although I think it is preferable to give it multiple cores because most of the compute bottleneck is parallelisable.

mcgml commented 4 years ago

Amazing thanks very much, works now. There's not way to tell Cromwell how many cores each task is using (through Docker) so could overload the server if using multiple cores. But the turn-around-time seems fine and is more efficient anyway.