eScience infrastructure for ecological applications of LiDAR point clouds
Deploying clusters of VMs on bare metal using OpenNebula, provisioning, and running services made easy by fully configurable scalable automation
Provides an easily configurable framework to configure and deploy clusters of VMs on 'bare metal' managed by OpenNebula and deploy services on them
(Re-)deploy defined VM+software configurations with a single command
Scale, instantiate, and tear down clusters of arbitrary size using a simple config file
Emma to deploy fully customisable ansible scripts
supports cluster wide service management
Automated instantiation and deployment of (clusters of) virtual machine(s) on bare metal using the OpenNebula platform, as well as subsequent provisioning and deployment of services incl., e.g. Dask.
marzipan consists of the core marzipan.py python module providing a high level interface to the OpenNebula cloud, as well as an accompanying Docker framework and configurable deployment scripts providing a fully automated instantiation and provisioning environment.
For provisioning marzipan makes use of the emma_marzipan fork ansible playbooks.
marzipan is based off and strongly draws from Lokum, but is updated to make use of current versions of Ansible as well as python 3, and circumvents recurrent synchronicity and timeout issues arsing from the interplay of terraform, the runtastic OpenNebula provider for terraform, and various (legacy) OpenNebula versions.
marzipan has been tested on the SURFsara HPC cloud, but should work for any OpenNebula platform.