krank (comparative kränker, superlative am kränksten)
- sick; ill
- (slang) excellent
Turn the krank on your data reduction tasks with the magic of containers.
TODO
- Log in to Ocelote or El Gato
module load singularity
to make thesingularity
command available- Request a large temporary workspace with
xdisk -c create -m 1000
and thencd /xdisk/$USER
so that the image builds outside your (somewhat small) home directory quota singularity build krank.img docker://magaox/krank
- Copy the images you want to analyze into
/xdisk/$USER
with rsync (e.g.rsync ./data filexfer.hpc.arizona.edu:/xdisk/yournetid/data
)
- UA HPC has a storage quota and a file count quota. Trying to retain intermediate files will quickly exhaust the latter, so be sure to delete those after combining.
- Singularity only has two (useful) users,
root
at build time (as whom all build commands are run) and$USER
(i.e. you) at run time. - UA-specific HPC network filesystem mount points are present as empty directories in the image to silence Singularity warnings (and let you access HPC shares when run on UA HPC)
- In Singularity:
singularity shell krank.img
- In Docker:
docker run -it magaox/krank bash
(After commenting out the entrypoint and rebuilding)
- Get access to https://hub.docker.com/r/magaox/krank/
- Install Docker Community Edition locally
git clone
this repository./build.sh
./push.sh
git commit
andgit push
when it's all nice and kentucky
- Dockerfile documentation
- UA HPC Singularity tutorials