Docker container for Scorex-Lagonaki about scorex

[kushti]

That would be awesome!

[oiwn]

Ok, trying to get in. Faced some issues, since there is compilation step (unlike python where you just push the sources inside the container and define entry point). But despite the issues was able to build image and run it. Unfortunately it's not usable yet in this form, need to integrate it with SBT to:

• automatically pull in dependencies and the correct path/name of target jars
• ensure build dependencies match dev. runtime environment. (i used sbt 0.13.8 in image while project required 0.13.9 so when i run container it started to download sbt again)

Also the image size is a bit too large - 800+mb and it takes time to build.

I'll try to do the following:

• pick SBT plugin: https://github.com/marcuslonnberg/sbt-docker or http://www.scala-sbt.org/sbt-native-packager/formats/docker.html
• build as small docker image as possible with Java environment to run compiled scorex.jar (i guess it's enough?)
• - use SBT to generate Docker configs and images
• - add docker task to build image from sbt console

[oiwn]

Guess i found nice solution for smaller image with java:

https://developer.atlassian.com/blog/2015/08/minimal-java-docker-containers/

guys suffered same problem:

Official Java images have historically been mastodontic - cue picture above - I just tried "docker pull java" and I got an image of 816.4MB.

and they was able to reduce the size of image:

fully functional Oracle Java 8 image weighing only 173Mb.

Few questions:

• can Scorex run under OpenJDK or OracleJDK is required?
• Which Java to pick 7 or 8?

And here is how to do the rest using sbt-docker:

https://velvia.github.io/Docker-Scala-Sbt/

[oiwn]

And as far as i see data is persistent between sessions, correct?

https://github.com/ConsensusResearch/Scorex-Lagonaki/blob/master/settings.json

"datadir" : "/tmp/data",

Is so it's required to expose /tmp folder to the docker container.

[oiwn]

#22 pull request

[kushti]

merged