The thesis is located inside the report folder
This repository provides a working Kotlin implementation of EpTO as well as JGroups SEQUENCER application to compare the both of them.
- Docker >= 1.12
- OpenJDK or OracleJDK >= 8
- Python >= 3.5
- pip >= 9.0.1
- GNU bash
- Read the Cluster instructions
If you want to run EpTO/JGroups localy execute: cluster/run_benchmarks.py with the --local option on
If you want to run it on your cluster follow the Cluster instructions
If you only want to obtain and run the Java program. Gradle with shadowJar is used to generate a jar file.
Scripts are provided to verify the ordering of EpTO/JGroups, extract various informations from EpTO/JGroups logs and the PSS health (in case of EpTO). They are located in the folder results.
The Failure Trace Archive regroups various failure traces. These traces can be exploited by the python classes cluster/nodes_trace.py and cluster/churn.py.
If you want to visualize the different traces I recommend using Sébastien Vaucher's script that generates plots for traces.[2]
The above classes can also be used to generate simple synthethic churn. To find out how to use them, please refer to the help of cluster/run_benchmarks.py