WHAT IS SAFEHAUS
Safehaus is a collaborative platform to support open source projects (including i.e. the KisKis Cluster Management System and its subprojects). Incidentally our motto is KisKis. It stands for "Keep it simple! Keep it safe!". We believe that tight simple software products (that minimize scope) designed to do one thing and one thing very well (in line with the UNIX mindset) are inherently more secure, easier to manage, and maintain.

Several contributors and users from around the world use and/or develop Safehaus Software. The software produced at Safehaus is distributed under the Apache Software License version 2.0 as open source software completely free for private and commercial use. So feel free to fork us to start your own project, or rebrand our products even for sale. We have no expectation, everything is gratis in the original tradition of truly free open source. Better yet, if you appreciate what we have done and gain from it, join us and let's work together. Everyone is welcome.
COMMUNITY - HOW WE OPERATE
At Safehaus, our aim is to build the coolest, easiest to use open source community environment. Part of that means making it easy for our contributors to use Safehaus tools, resources and services to develop everyone’s own community around solid products. This page briefly explains what exists for you to use to build your team and to build awesome OS software.

We offer an open source alternative to the crusty excessively process oriented, rules oriented, work environments of our boring day jobs. Heck if you can convince your employer to start and host projects here then you can be hip even at your day job. Basically we've automated the heck out of everything so computers manage the process auto-magically so you don't have to deal with process and managers who just get in the way. It's a place to breath and collaborate with like minded, free thinking, and cool developers.

You may be thinking, “Man I don't know anything about open source”. Well maybe you are right. But you can change that if you can read. So if you want to have some light reading to understand the OS way please take a look at these totally awesome small essays from Eric Raymond:

POLICIES
Basically have fun! We try not to have nonsense heavy policies here. That's for your crusty day job. It's OK that's not the case here and that's the point to Safehaus. We're here to have fun and chillax with technology. If there are problems we resolve them according to our three simple rules:

  • The uber despot is always right
  • You can question the uber despot, but in the end he's still always right
  • When in doubt about the uber despot being right, refer to rule #1

Seriously, we should operate as a meritocracy here but there will be natural leaders. The doers naturally lead but if there's evil running about we will intervene to keep the peace. Hopefully that never needs to happen.

When in doubt we default back to ASF like policies that are pretty well documented here below. The only difference is we allow for benign dictatorships as long as they do not result in a any kind of abuse or ill conduct.

  • Foundation Marks (more on this below)
  • Release Guidelines and Release Info From the Httpd Project
  • Source Header Guidelines
  • 3rd Party Licensing Policy (follow this to the 'T' young padiwan)
  • Do you really want a Project Management Committee? (It's up to your team!)



CODE OF ETHICS, CONDUCT, VALUES AND ALL
We expect people to be good to one another. We have certain guidelines here:
  • Keep it simple, keep it safe (our motto)
  • Be tolerant and respectful of people's religions, nationalities, cultures, and view points
  • "Open Source is a social revolution, not a technical one so" - Original quote by Alex Karasulu
  • Be good to each other (Don't be evil!)
  • Don't be crusty
  • Don't hide knowledge and information, share it
  • Don't worry about making money, worry about doing something well and money will come
  • Communication is the basis for community
  • Code is static and dies without a community, it's the community that has value
  • Do good deeds and help your fellow members
  • Together we're safe, apart we're at risk

Outside of these guidelines, code of ethics etc. everything else is up to you. Think of this place as a hangout for cool people. If you go somewhere where the people are so fun and cool think about how you would have to be.
SAFEHAUS INFRASTRUCTURE
Safehaus members have a great advantage to use several software development tools, platforms while they're developing Safehaus projects. We have the habit and encourage newcomers to get the habit of utilizing these awesome tools so that our lives are easier. Here are some of the tools that we are using:

ATLASSIAN JETBRAINS
  • Intellij IDEA Ultimate Edition

STRUCTURE101

SONARGRAPH
PROJECTS
  • KISKIS CLUSTER MANAGEMENT SYSTEM
    • KISKIS MANAGEMENT SERVER
    • KISKIS AGENT
    • KISKIS BIG DATA PRODUCT PORTFOLIO

  • ODROID ATTACK
  • LOAD BALANCER
  • ODROID VINN INTEGRATION


KISKIS CLUSTER MANAGEMENT SYSTEM


The aim of the KisKis Cluster Management System is to produce an extremely simple, easy to use, end to end, secure private cluster management and control system that gets users up and running with their clustered applications in less than 5-10 minutes. We know several products already exist to manage private on premises IaaS cloud computing environments and our users can also use those products to manage their clusters. Platforms like Apache CloudStack or OpenStack are excellent products that will manage clusters just as well as KisKis would, however we feel these products try to cover too many scenarios and aspects being generally focused on IaaS.

They are designed to handle a wide range of situations, operating systems, software and hardware. The KisKis CMS is designed to just do one thing, manage clusters of clusterable software products, prepackaged into the system ready to be deployed to your own private infrastructure at the press of a button.

We package existing big data frameworks and enable you to provision and de-provision clusters of them over a set of hardware. You just boot the hardware up with USB, CDROM, or BOOTP and unattended installation of the hypervisor is automatically handled. On first boot, after installation, the hardware appears on the management console ready to be used to provision your clusters. You can then dynamically manage your cluster. There's no complex primary/secondary storage scenarios, or networking matters with network hardware. Everything is managed for a simple flat configuration using local storage that should fit most developers and IT professional needs in small companies with a few commodity machines they want to throw at it.

KISKIS MANAGEMENT SERVER
KisKis Management Server is a specialized Ubuntu based operating system which takes control of any kind of physical machines, virtual machines and containers on the network designated as nodes for the cluster you want to form.

Management Server has its own pre-prepared DHCP and DNS server packages specialized for management server¿s use, TFTP and NFS packages for automated OS installation on hosts, Apache Cassandra installation for storing cluster events and information, and an apt repository that contains all packages an agent might need including all Big Data products. On top of these, Management Server has a set of OSGi bundles contained in Apache Karaf used for these purposes;

  • Communicating with agents via a secure AMQP bus
  • Sending commands to agents and receiving results
  • Increasing or decreasing the number of virtual nodes
  • Providing REST endpoints for extensive programmatic interaction with the cluster
  • A comprehensive and easy to use web UI to setup, configure, manage, monitor everything on the cluster


KISKIS AGENT
The Agent is a lightweight tool that sits on each physical host, virtual machines, LXC containers and even on Management Server itself. Its job is to take commands from the management server through AMQP, execute them on its platform, and send back the results.

KisKis Agent is implemented in C++, so it has a very small memory footprint and is able to run on almost any host; a full server class rackmount machine, a workstation, LXC containers, virtual machines, and even ODROIDs, routers.

KISKIS BIG DATA PRODUCT PORTFOLIO
We pre-package almost all open source big data related products, tools, databases, programs so that the users are able to install them on any number of nodes they wish by providing a handful of simple configuration parameters and following simple instructions on Management Server's Web UI. Setting up big data clusters have never been this easy and fast!

Here is our big data product portfolio:



ODROID ATTACK

This project is about taking tens of ODROID-U2 machines (http://tinyurl.com/nrlcwr6) and configuring them in different scenarios to attack on working clusters, basically for test and benchmarking purposes.

LOAD BALANCER

This newly starting project is about creating an open source load balancer which can easily be integrated into multiple different infrastructure architectures and in multiple different scenarios.

ODROID VINN INTEGRATION

The aim of this project is to integrate Odroid devices with a number of Vinn 3G dongles (or any USB 3G dongles) in an hot pluggable, easy to configure and flexible manner. By that, these lowpower, low-cost devices can be portably used to provide web services, act as servers and even form clusters.

THANKS TO
We thank Atlassian for providing free licenses of Jira, Confluence, Stash, Crucible/Fisheye, Bamboo, Crowd, Clover, Gliffy and Team Calendars.
We also thank JetBrains for giving Intellij IDEA Ultimate Edition licenses to be used in open source projects in Safehaus. And thanks to Hello2morrow for Sonargraph and thanks to Structure101, Jenkins and everyone contributing to or supporting open source.