Skip to content
Switch branches/tags
Go to file

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time


This is the Kadeploy3 readme file. Most of this file's content come from
this publication:

More informations available here:

Kadeploy is a scalable, efficient and reliable deployment system (cluster
provisioning solution) for cluster and grid computing. It provides a set of
tools for cloning, configuring (post installation) and managing cluster
nodes. It can deploy a 300-nodes cluster in a few minutes, without
intervention from the system administrator. It can deploy Linux, *BSD,
Windows, Solaris.

It plays a key role on the Grid'5000 testbed, where it allows users to
reconfigure the software environment on the nodes.

How it works?
This is how Kadeploy works:
(1) Minimal environment setup
  The nodes reboot into a trusted minimal environment that contains all the
  tools required for the deployment (partitioning tools, archive management,...)
  and the required partitioning is performed.
(2) Environment installation
  The environment is broadcast to all the nodes and extracted on the disks.
  Some post-installations operations can also be performed.
(3) Reboot on the deployed environment

Kadeploy3 takes as input an archive containing the operating system to
deploy, called an *environment*, and copies it on the target nodes. As a
consequence, Kadeploy3 does not install an operating system following a
classical installation procedure and the user has to provide an archive of
the operating system's filesystem (as a tarball, for Linux environments).

How does Kadeploy control the boot of the nodes ?
This is how Kadeploy controls the boot process of the nodes in order to be able
to perform it's installation task:
(1) Kadeploy writes PXE profiles on a TFTP or HTTP server
(2) Kadeploy triggers the reboot of compute nodes using SSH, IPMI or a
    manageable PDU
(3) Nodes get their configuration using DHCP
(4) Nodes retrieve their PXE profile using TFTP
(5) Nodes boot on the specified system (which can either be located on the
    node's hard disk or on the network)

The Kadeploy3 software suite
Kadeploy3 is packaged with a set of complementary tools that are briefly
described in this section.

### Management of images
The *Kaenv* tool enables users and administrators to manage a catalog of
deployment images, either shared between users or private to one user.

### Rights management
*Karights* is used to define deployment permissions for users. It also
provides the glue to integrate Kadeploy with a batch scheduler, making it
possible to allow a given user to deploy a set of nodes for the duration of
his job on the cluster.

### Statistics collection
Deployment statistics (durations, success/failures) are collected
continuously, and available through the *Kastat* tool. This data can be
leveraged by system administrators to identify nodes with hardware issues.

### Frontends to low-level tools
Tools such as *Kareboot*, *Kaconsole* and *Kapower* act as frontends to
lower-level tools, such as those based on IPMI, and integrate with the
Kadeploy rights management system, to allow users to reboot, power-off/on,
and access the remote serial console of nodes.

Required services:
- dhcp
- tftp
Recommanded services:
- sql
- dns
- http

License & copyright
Copyright 2008-2013 INRIA, Kadeploy Developers <>

Kadeploy3 is developed at INRIA Nancy - Grand Est. Its development is currently
supported by the ADT Kadeploy project (2011-2013), which is led by the
AlGorille team at LORIA.

Current developers are Luc Sarzyniec (Inria, main developer since 2011),
Emmanuel Jeanvoine (Inria, design & main developer from 2008 to 2011) and
Lucas Nussbaum (Univ. de Lorraine, design since 2011).


Kadeploy is a scalable, efficient and reliable deployment system (cluster provisioning solution) for cluster and grid computing




No packages published