| commit | 28e7cf3b8cf8e46101713b6239d6a62b16ce3849 | [log] [tgz] |
|---|---|---|
| author | Your Name <you@example.com> | Thu Dec 12 14:02:40 2019 +0000 |
| committer | Ignacio Pascual <ignaciojavierpascual@gmail.com> | Wed Feb 26 11:15:18 2020 +0100 |
| tree | b707f1d87dcaea80720dd9a11d2f45ca50ee84b2 | |
| parent | 92f15235ffa86ca89a5674378f83c6761b02826a [diff] |
Initial implementation of offline engine deployment A new option has been added "-x" to execute this use case. It will use a predefined tar.gz package that contains all the necessary files to perform the installation of k8s-calico-nofeature scenario and has to be downloaded in advance. It has been documented the structure of this file and how to fetch the dependencies in an online machine. Apt dependecies are handled by apt-cacher-ng that sets up a proxy in localhost. This is used by the newly created nodes. Kubespray cache (binaries and docker images) is copied to the nodes and the playbook is configured to use those files. In post deployment, the remaining of Docker images are pre- loaded in the target machines. A better solution would be to establish a local docker repository, but it is left out of this patchset. Change-Id: I3155e3bca56435a97f165a266935e7fec49f1a4d
The Cloud Infra Engine is created based on OPNFV Cross Community CI (XCI) project in order to automate the deployment and testing of various cloud infra scenarios. [1]
The supported scenarios are based on
technologies and they can be deployed and tested using different types of resources such as
The possibility of using one of the resources listed above highly depends on the scenario the users want to deploy, test, and use. More information is available in corresponding sections.
Cloud Infra Engine lets users to deploy the scenario of their choosing on to their workstations. Minimum requirements for the host where the Cloud Infra Engine is executed are
Recommended requirements are
The engine currently only supports Ubuntu16.04.
Apart from having sufficient performance and capacity on the host, few packages need to be installed on the host before executing the engine
The user that is executing the engine should also have passwordless sudo enabled.
Cloud Infra Engine is version controlled on Nordix Gerrit so its repository needs to be cloned.
git clone https://gerrit.nordix.org/infra/engine.git
Cloud Infra Engine expects the ssh keys to be created in advance. If you don't have ssh keypair already, you can do that by executing below command.
ssh-keygen -t rsa
Once the keypair is generated, the main script deploy.sh can be executed in order to start deployment of the default scenario on virtual machines that are created by the engine.
cd engine/engine ./deploy.sh -h # get help ./deploy.sh -c
Once the script execution starts, it will prepare the environment, create libvirt resources, provision libvirt vms and install the selected scenario on them. The overall process takes about 40 minutes to complete if ramdisk and deployment images are created in advance.
Please note, in the default / embeded pdf / idf files, we have specified two disks for each vms we are going to create. If you are going to provide pdf / idf files yourself with mulitple disks, please note the current deployment script will not work when the disks have the same size as we specify the bootable disk based on size.
TBD
TBD
TBD
TBD