Which tools do we expect to find on the cluster by default?
- Java
- gcc
- cmake > 3.1
- python
- ...?
Food for thought: might be nice to have some dummy workflow nodes for these guys so that we can make our tools explicitly depend on them. If we change our minds, this might help refactoring/bookkeeping easier.
Also since I'm testing #266 on gcloud with Ubuntu, here are my notes (I get everything in biokepi except GATK/mutect installed; but didn't run anything end-to-end):
Biokepi ./all_downloads test requires:
sudo apt-get install --yes openjdk-9-jre cmake r-base
And razers3 does not compile on a “small” instance:
sudo fallocate -l 2G /swapfile
sudo chmod 600 /swapfile
sudo mkswap /swapfile
sudo swapon /swapfile
Just check it worked:
free -h
total used free shared buff/cache available Mem: 1.7G 83M 1.5G 20M 112M 1.5G Swap: 2.0G 0B 2.0G
This is relevant also: https://github.com/hammerlab/biokepi/issues/283
Also relevant:
- https://github.com/hammerlab/stratocumulus/blob/master/src/lib/deploy.ml#L223
- https://github.com/hammerlab/stratocumulus/blob/master/src/lib/deploy.ml#L122
I've run a bunch of the tools on those Xenial GCloud boxes.