Skip to content
This repository has been archived by the owner on May 27, 2024. It is now read-only.

deNBI/cloud-portal-client

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Cloud-Portal-Client

The Cloud Portal Client is a client written in Python which provides functions to create virtual machines in an OpenStack project.

Features

  • Create and Delete Instances in an OpenStack project
  • Stop, Resume and Reboot Instances in an OpenStack project
  • Get Flavors and Images from an OpenStack project
  • Creating Snapshots
  • Create and attach volumes to an virtual machine
  • Add floating ips to virtual machines

Preparation

Download and source rc file

Download and source OpenStack RC FILE v3 with the following command:

source NameOfRcFile.sh

Configuration

You can view (almost) all existing parameters in the yaml file. For local development: Please copy this file and rename it to config_local.yml and fill in missing parameters. For staging/production setup: Please copy this file and rename it to config_YOUR_LOCATION.yml and fill in missing parameters. Also you need to provide the path to your config file as the first param when starting a server.

Furthermore there are some parameters you must set in the .env file. Copy the .env.in to .env and fill in the missing parameters. When starting with commandline you will need to export some of them manually.

Security Groups

The config file contains two ids for security groups:

gateway_security_group_id: ID of the security group which allows the access to the gateway and thus the forwarding to the instances. Every other security group which gets created will only be accessible from this security group.

forc_remote_id: ID of the security group from the forc gateway, every forc resenv will only be accessible from this security group

Gateway

The client can use a Gateway for starting and stopping machines which allows to use just one floating IP instead of one floating IP per Machine. You can read here how to setup a gateway on an OpenStack instance. You can also find complete scripts in the gateway folder. The client will provide all images with at least one tag, which will be filtered for in the cloud-api. Also the client provides all flavors, which will also be filtered in the cloud-api.

Attention: If you are also using the machine where you run the client as a gateway, it is very important to configure the iptables before installing and using docker, otherwise docker could destroy the rules!

Persistent IP-Tables

If you are using a gateway, it makes sense to make your iptable rules persistent on this gateway. To do this, simply install and use this package:

sudo apt-get install iptables-persistent

Create certificates

To create your own certificates follow the instructions on this Website: thrift certificates

Attention: You need to create your own Server.pem and your client needs the appropriate Client.pem and CA.pem,

Production

Using Docker

Specify in the .env file which release should be used by the client. Then you can start the client with:

$ docker-compose up

or

$ make production

Attention: If you change the port in the yaml file you also need to change the port mapping in the docker-compose.yml!

Development

Linting

$ make lint

will run flake8 on the source code directories.

Documentation

You need thrift to be installed on your machine (see Thrift section).

$ make docs

will create documentaion in the /docs directory.

Docker-Compose

Run the following command in order to start the container setup: Make sure, that your OpenStack RC File is sourced.

docker-compose -f docker-compose.dev.yml up --build

or

make dev

You can additionally start bibigrid:

make dev-bibigrid

and additionally also in detached mode:

make dev-bibigrid-d

to list all make commands use:

make help

Thrift Development

A detailed instruction for installing thrift can be found here. With the portal_client.thrift you can autogenerate your code.

make thrift_py

This command will generate python code from the thrift file.

In order for the cloud-api to use the new/changed methods, VirtualMachineService.py, ttypes.py and constants.py must be copied over.

Because docker can't use relative imports, you also need to change the import of ttypes.py in constants.py and VirtualMachineService.py:

from .ttypes import *
from ttypes import *

Attention: The cloud-api needs the files with the relative imports (from .ttypes)!

A detailed instruction, how to write a thrift file can be found on this link: thrift

To use the methods declared in the thrift file you need to write a handler which implements the Iface from the VirtualMachineService. The handler contains the logic for the methods. Then you can start a server which uses your handler. Example python code for the server:

if __name__ == '__main__':
    with open("s../config.yml", 'r') as ymlfile:
        cfg = yaml.load(ymlfile)
        HOST = cfg['openstack_connection']['host']
        PORT = cfg['openstack_connection']['port']
        CERTFILE = cfg['openstack_connection']['certfile']
    handler=VirtualMachineHandler()
    processor=Processor(handler)
    transport = TSSLSocket.TSSLServerSocket(host=HOST, port=PORT,certfile=CERTFILE)
    tfactory = TTransport.TBufferedTransportFactory()
    pfactory = TBinaryProtocol.TBinaryProtocolFactory()
    server = TServer.TSimpleServer(processor, transport, tfactory, pfactory)
    server.serve()

Deployment via ansible

1.Create your inventory file:

Example:

[test]
REMOTE_IP ansible_user=ubuntu ansible_ssh_private_key_file=PATH_TO_SSH_FILE ansible_python_interpreter=/usr/bin/python3

where

  • REMOTE_IP is the IP of your staging machine

  • PATH_TO_SSH_FILE is the path to the ssh key of the virtual machine

2.Set SSH keyforwarding

In order to checkout the GitHub project you will have to enable SSH Key forwarding in your ~/.ssh/config file.

Host IP
 ForwardAgent yes

where IP is the IP of the machine you want to start the portal.

3.Set SSH Key in Github

The GitHub repository will cloned using an ssh key you set for the GitHub repository. You can read how to set an ssh key for the cloud-portal repository on this website.

4.Install needed libraries

ansible-galaxy install -r ansible_requirements.yml

6.Set all variables

Set all variables that can be found in .env and VirtualMachineService/config/config.yml file. You can have more than one .env file (.env and .env_* are not tracked by git) and specify which you want to copy by using the env_file variable. You can have more than one VirtualMachineService/config/config.yml file (VirtualMachineService/config/config_* are not tracked by git) and specify which you want to copy by using the client_config variable. These options are useful when maintaining multiple client sites.

7.Run the playbook

You can run the playbook using the following command:

ansible-playbook -i inventory_openstack site.yml

where

  • inventory_openstack is your inventory file which you created in the first step.

  • If you also want to start bibigrid use the tag "bibigrid" Choose different files

You can also specify different .env, config.yml and server.pem files.

You can also specify branch, tag, commit that should be checked out with --extra-vars.

For Example:

ansible-playbook -i inventory_openstack --extra-vars "repo_version=master" site.yml

Optional Keys are:

  • repo_version
  • env_file
  • client_server_pem
  • client_config

Note: Default repository is always master. Also by default, the files are taken from the folders as for local start.