SyncProject
##Dependencies:## The 3 servers (master, slave and a git server) all run in virtual machines (Debian on VirtualBox), with identical hardware. The scripts to setup/run the experiments and applications, depend on this setup. However, in general, the approaches here are generic to run on all types of OS and hardware.
- For the restlet to work: Java webserver which alles encoded slashes in parameters. (for tomcat, add this to catalina.sh: JAVA_OPTS="$JAVA_OPTS -Dorg.apache.tomcat.util.buf.UDecoder.ALLOW_ENCODED_SLASH=true"
- Operating system with rsync
- A mysql db on the host, used to store experiment results in. DB name: 'Experiments'. Set username to 'syncProject' and password blank. Allow connections from localhost, slave and master
##Getting Started##
- Ideally, add a network interface on the host computer, and use that to create an internal network with the 3 nodes. (in virtualbox, a host-only adapter)
- For internet access, you can use NAT as another network adapter of the 3 nodes
- Make sure all hosts files (on the nodes, as well as the host) point to the right ip
Directories:
- bin: collection of scripts to set up the servers, compile the project, sync the jars/wars to their proper location, and run experiments (see readme in there for more info)
- config: Config directory. Used in all java code, as well as in bin directory scripts. Other folder (e.g. restlet folder) use symlink to connect include a config file. Other applications (e.g. daemon) gets their config file by getting it from the github repo (https://raw.github.com/....)
- daemon: Code for the java daemon, which is runnable on the slave node. Simply build, and then execute using 'java -jar daemon.jar'.
- restlet: The restlet which runs on the master node, and which handles execution and logging of queries