Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 15 Next »


Instructions for installing Airship+Tungsten Fabric using the Regional Controller and the TF Blueprint


Requirements:

  • Regional Controller:  AWS instance t2.medium or any virtual or baremetal  node with 2CPU and 4GB Memory (OS: Ubuntu Xenial 16.04)
  • Airship+Tungsten Fabric host: AWS instance m5.4xlarge or any virtual or baremetal node with 16CPU and 64GB Memory (OS: Ubuntu Xenial 16.04)

Both nodes must be available by ssh.

Overview

Akraino Regional Controller  is necessary part of release 2 deployment procedure. It's Akraino approved blueprint which is common for all of release 2 blueprints and which is using for Edge Site, Blueprint and POD deployment.

After creating the POD Regional Controller creates WORKFLOW that initiates remote installation Airship with TungstenFabric.

More information about Regional Controller:

If you already have Regional Controller you can use it for deployment and if you don't have it you can install it following the instruction bellow

Installing the Regional Controller

You can use  any machine or VM dedicated for this purpose. See instructions on how to start the regional controller.  

https://wiki.akraino.org/display/AK/Starting+the+Regional+Controller

After you have a working Regional Controller you have to login on it and follow the steps bellow

Steps to manual installation of  TungstenFabric Blueprint

All this steps must be done on Regional Controller


Clone the nc/tf repository using 

git clone https://gerrit.akraino.org/r/nc/tf


Setup ssh keys and put them on web -server 

Regional Controller goes to the remote node by ssh, so it needs ssh private key. It can be provided as http URL.  (It's not secure for production, it's only OK for the demo).

Put ssh private key and script deploy.sh on some web server. Ssh public key must be written to the .ssh/authorized_keys on remote node.

Hint: python provisional web server can be used on the localhost. Use python3 -m http.server


 Edit the file setup-env.sh 

Update all the environment variables define the ip addresses nodes, web server baseurl, etc. Define where the Regional Controller is located, as well as the login/password to use
(the login/password shown here are the built-in values and do not need to be changed, if you have not changed them on the Regional Controller):

export RC_HOST=<IP or FQDN name of Regional Controller>
export USER=admin
export PW=admin123


Generate yaml files from templates

source setup-env.sh
cat objects.yaml.env | envsubst > objects.yaml
cat TF_blueprint.yaml.env | envsubst > TF_blueprint.yaml


Clone the api-server repository.

This provides the CLI tools used to interact with the Regional Controller. Add the scripts from this repository to your PATH:

git clone https://gerrit.akraino.org/r/regional_controller/api-server
export PATH=$PATH:$PWD/api-server/scripts


Load the objects

Load objects defined in objects.yaml into the Regional Controller using:

rc_loaddata -H $RC_HOST -u $RC_USER -p $RC_PW -A objects.yaml


Load the blueprint

Load the blueprint  into the Regional Controller using:

rc_cli -H $RC_HOST -u $RC_USER -p $RC_PW blueprint create TF_blueprint.yaml


Get and export UUIDs

Get the UUIDs of the edgesite and the blueprint from the Regional Controller using:

rc_cli -H $RC_HOST -u $RC_USER -p $RC_PW blueprint list
rc_cli -H $RC_HOST -u $RC_USER -p $RC_PW edgesite list

These are needed to create the POD. You will also see the UUID of the Blueprint displayed when you create the Blueprint in step 8 (it is at the tail end of the URL that is printed).
Set and export them as the environment variables ESID and BPID.

export ESID=<UUID of edgesite in the RC>
export BPID=<UUID of blueprint in the RC>


Generate POD.yaml

cat POD.yaml.env | envsubst > POD.yaml


Create the POD

Create the POD using

rc_cli -H $RC_HOST -u $RC_USER -p $RC_PW pod create POD.yaml

This will cause the POD to be created, and the deploy.sh workflow script to be
run on the Regional Controller's workflow engine. This in turn will login to remote node by ssh
and install airship+ tungstenfabric demo on it


Checking POD status

If you want to monitor ongoing progess of the installation, you can issue periodic calls
to monitor the POD with:

rc_cli -H $RC_HOST -u $RC_USER -p $RC_PW pod show $PODID

where $PODID is the UUID of the POD. This will show all the messages logged by the workflow, as well as the current status of the workflow. The status will be WORKFLOW while the workflow is running, and will change to ACTIVE if the workflow completes successfully, or FAILED, if the workflow fails.

  • No labels