All I can say is that it was late when I wrote the script. And I was distracted between the feline overlords (one of whom is attempting to climb into my lap) and the babble box. That’s my story and I’m sticking to it. PEBKAC and ID10T errors were not involved.
Creative Commons licensed ( BY ) flickr photo shared by JeepersMedia
I’m processing 50GB of compressed cheminformatics data on the Pi Swarm, extracting certain pieces of data from substance records from NIH. I created a docker image containing Perl and the SDF Toolkit rather than writing my own parser. Tested a trivial case and pushed it to Docker Hub as nimblestratus/rpi-sdf-toolkit. So far, so good. Then in order to get an idea of how long it would take to process the lot, I wrote a quick script to split up a chunk of the data among the Pi’s.
After determining its next chunk to process, the container is started. However, my script was invoking Docker like this:
|
1 2 3 |
docker -v /mnt:/data run nimblestratus/rpi-sdf-toolkit ARG ARG ARRRRRRRG |
And it printed the version and came back immediately. Did not pass Go. Did not collect $200. Certainly didn’t process any of my data.
What I should have typed to mount a volume was:
|
1 2 |
docker run -v /mnt:/data nimblestratus/rpi-sdf-toolkit |
So…. long story made short: order does matter. Now serving number 35.
