Creating test VMs using GitLab CI

Before you start
1. Create a VM for the test application
2. Prepare a VM with the test application
3. Check that the test application is working
4. Take a snapshot of the VM disk
5. Create a virtual machine with GitLab
6. Configure GitLab
7. Configure Runner
8. Configure CI
9. Check how the application works on the VM created using CI

You can use Yandex.Cloud to automate routine actions, such as running a specific script after each commit to the master branch in a git repository. In the example below, a VM is created and tested after each commit.

To configure CI for VM disk snapshots:

Create a VM for the test application: create a new virtual machine that will be used to make a disk snapshot for creating new virtual machines via CI.
Prepare the VM with the test application: install the web server and the set of components for the test application on the VM. Write a test application that will reverse the words in a text sent to the server.
Check how the application works: send a test request to check the server settings and the results.
Create a snapshot of the VM disk: create a snapshot of the VM disk that CI will use to create new virtual machines.
Create a VM with GitLab: create a virtual machine with GitLab, where the repository will store the CI settings and the functional testing script.
Configure GitLab: create a repository for files and get the necessary configuration parameters.
Configure Runner: set up the Runner utility for performing tasks.
Configure CI: set the CI configuration by specifying the necessary parameters for commands and testing.
Check how the application works on the VM created using CI: make sure that VMs created with CI and the snapshot are created and the test application is running.

Before you start

Before creating a VM:

Go to the Yandex.Cloud management console and select the folder where you want to perform the operations.
Make sure the selected folder has a network with a subnet that the VM can be connected to. To do this, select Virtual Private Cloud on the folder page. If the list contains a network, click on its name to see the list of subnets. If there aren't any networks or subnets, create them.

1. Create a VM for the test application

Create a VM where the test application will be installed, the set of components required for it to work, and a web server:

On the folder page of the management console, click Create resource and select Virtual machine.
In the Name field, enter the VM name: ci-tutorial-test-app.
Select the availability zone to locate the VM in.
Select a public image of Ubuntu 18.04.
In the Computing resources section, select the following configuration:
- Platform: Intel Cascade Lake.
- Guaranteed vCPU share: 5%.
- vCPU: 2.
- RAM: 1 GB.
In the Network settings section, select the subnet to connect the VM to when creating it.
Specify data required for accessing the VM.
- Enter the username in the Login field.
- In the SSH key field, copy the contents of the public key file.
  You need to create a key pair for SSH connection yourself. To generate keys, use third-party tools, such as ssh-keygen utilities on Linux and macOS or PuTTYgen on Windows.
Click Create VM.

Creating the VM may take several minutes. When the VM's status changes to RUNNING, you can begin configuring it.

When a VM is created, it is assigned an IP address and hostname (FQDN). This data can be used for SSH access.

2. Prepare a VM with the test application

On the created VM, install the set of components required for the test application and a web server for processing requests. The application will be written in Python 2.

In the Network section on the VM page of the management console, find the VM's public IP address.
Connect to the VM over SSH. You can use the ssh tool on Linux and macOS and PuTTY for Windows.
```
$ ssh <Login>@<Public_IP_of_the_VM>
```
Run the apt update command to update lists of available packages.
Install the necessary packages (jq JSON processor, git client, PIP package manager, virtualenv virtual environment management system, the set of header files for Python C API, and the nginx web server):
```
$ sudo apt-get --yes install jq git python-pip virtualenv python-dev nginx-full
```
Create the folder where the application will be located and make the user you used for connecting to the VM the owner of the folder:
```
$ sudo mkdir /srv/test-app
$ sudo chown -R $USER /srv/test-app
```
Go to the folder and create a virtualenv virtual environment in it:
```
$ cd /srv/test-app
$ virtualenv test-venv
```
Activate the virtual environment:
```
$ . test-venv/bin/activate
```
Install the Flask framework and the uWSGI web server in the virtual environment:
```
$ pip install flask uwsgi
```
Deactivate the virtual environment:
```
$ deactivate
```
Create the api.py file in the /srv/test-app folder:
```
$ touch api.py
```

Open the api.py file using any text editor, and insert the Python code that:

Accepts a text string as input in the text parameter.
Writes each word from the passed string in the opposite order.
Returns a response:
- In JSON format, if the client application can accept JSON.
- In plain text if the client application does not accept JSON.

# api.py
import json
from flask import Flask, request, make_response as response

app = Flask(__name__)


@app.route("/")
def index():
    text = request.args.get('text')
    json_type = 'application/json'
    json_accepted = json_type in request.headers.get('Accept', '')
    if text:
        words = text.split()
        reversed_words = [word[::-1] for word in words]
        if json_accepted:
            res = response(json.dumps({'text': reversed_words}), 200)
        else:
            res = response(' '.join(reversed_words), 200)
    else:
        res = response('text not found', 501)
    res.headers['Content-Type'] = json_type if json_accepted else 'text/plain'
    return res

Create the wsgi.py file in the /srv/test-app folder:
```
$ touch wsgi.py
```
Open the wsgi.py file using any text editor and insert the code that runs the test application:
```
# wsgi.py
from api import app

if __name__ == "__main__":
    app.run()
```
Create the test-app.ini file in the /srv/test-app folder:
```
$ touch test-app.ini
```

Open the test-app.ini file using any text editor and insert the configuration of the uWSGI server:

#test-app.ini
[uwsgi]
module = wsgi:app

master = true
processes = 1

socket = test-app.sock
chmod-socket = 660
vacuum = true

die-on-term = true

Assign the www-data user as the owner of the /srv/test-app folder and the files it contains:
```
$ sudo chown -R www-data:www-data /srv/test-app
```

Prepare the service to start your uWSGI server. To do this, make the /etc/systemd/system/test-app.service file look like this:

#/etc/systemd/system/test-app.service
[Unit]
Description=uWSGI instance to serve test API
After=network.target

[Service]
User=www-data
Group=www-data
WorkingDirectory=/srv/test-app
Environment="PATH=/srv/test-app/test-venv/bin"
ExecStart=/srv/test-app/test-venv/bin/uwsgi --ini test-app.ini

[Install]
WantedBy=multi-user.target

Specify the settings of the new virtual server in the nginx configuration by making the /etc/nginx/sites-available/test-app.conf file look like this:

#/etc/nginx/sites-available/test-app.conf
server {
    #server_name test-app.yandex  www.test-app.yandex;

    listen 80;

    location /test/ {
        include uwsgi_params;
        uwsgi_param SCRIPT_NAME /test;
        uwsgi_modifier1 30;
        uwsgi_pass unix:/srv/test-app/test-app.sock;
    }
}

Create a symbolic link that points to the nginx configuration file test-app.conf:

$ sudo ln -s /etc/nginx/sites-available/test-app.conf /etc/nginx/sites-enabled/

Delete the symbolic link that points to the default nginx configuration:
```
$ sudo unlink /etc/nginx/sites-enabled/default
```
Add the service to the system autorun list:
```
$ sudo systemctl enable test-app.service
```

3. Check that the test application is working

To make sure the test application is running and the web server is configured correctly, send a simple request.

In the browser bar, enter the URL for testing the web server and application:
```
http://<Public_IP_of_the_VM>/test/?text=hello_world
```
If everything is working, the screen will display text with reversed words from the text parameter.

4. Take a snapshot of the VM disk

To easily transfer the created application and web server configuration to VMs created with CI, you need to take a snapshot of the test VM disk.

In the Yandex.Cloud management console, select the folder where the VM was created.
Choose Compute Cloud.
Find the ci-tutorial-test-app VM and select it.
Click Stop.
In the window that opens, click Stop.
After stopping the VM, select the Disks tab.
In the disk row, click and select Create snapshot.
In the window that opens, enter a name for the snapshot: test-app-snap.
Click Create snapshot.

5. Create a virtual machine with GitLab

One of the ways to set up CI in Yandex.Cloud is to take advantage of a public image with GitLab pre-installed. GitLab includes a set of tools for managing git repositories and configuring CI.

On the folder page of the management console, click Create resource and select Virtual machine.
In the Name field, enter the VM name: ci-tutorial-gitlab.
Select the availability zone to locate the VM in.
Under Images, click Choose.
In the window that opens, open the DevTools tab.
Choose the GitLab image.
In the Computing resources section, select the following configuration:
- Guaranteed vCPU share: 100%.
- vCPU: 2.
- RAM: 2 GB.
In the Network settings section, select the subnet to connect the VM to when creating it.
Specify data required for accessing the VM.
- Enter the username in the Login field.
- In the SSH key field, copy the contents of the public key file.
  You need to create a key pair for SSH connection yourself. To generate keys, use third-party tools, such as ssh-keygen utilities on Linux and macOS or PuTTYgen on Windows.
Click Create VM.

Creating the VM may take several minutes. When the VM's status changes to RUNNING, you can begin configuring it.

When a VM is created, it is assigned an IP address and hostname (FQDN). This data can be used for SSH access.

6. Configure GitLab

To configure GitLab and prepare CI, you must create a new repository and enter the necessary parameters for authorization in CI.

In a browser, open the GitLab admin panel for the created VM. To do this, open a link that looks like http://<VM-public-IP-address>.
Set the administrator password.
Log in as root with the administrator password.
Click Create a project.
Set the project name: gitlab-test.
Click Create project.
Get an OAuth token from Yandex OAuth. To do this, follow the link and click Allow.
In the browser, open a link that looks like http://<VM-public-IP-address>/root .
Select the gitlab-test project.
On the screen that opens, choose the Settings tab on the left. In the menu that opens, select CI/CD.
Under Variables, click Expand.
Create a new variable:
- For the variable name, specify YC_OAUTH.
- Specify the OAuth token you received as the value of the variable.
- Click Save variables.
Under Runners, click Expand.
Under Set up a specific Runner automatically, you will see the server address to connect to and the token for registering the server in the project. Use these values when registering Runner.

7. Configure Runner

Runner is a tool for performing tasks that a user creates. You need to install Runner on the VM and register it in GitLab. In order for Runner to perform tasks, prepare additional components: install the Yandex.Cloud CLI and create a test to check the created VM.

Connect to the VM with GitLab over SSH.

$ ssh <Login>@<Public_IP_of_the_GitLab_VM>

Add the new repository to the package manager:

$ curl -L https://packages.gitlab.com/install/repositories/runner/gitlab-runner/script.deb.sh | sudo bash

Install the Runner that will run your CI scripts:
```
$ sudo apt-get -y install gitlab-runner
```

In the step Please enter the gitlab-ci coordinator URL, specify the address of the GitLab server.
In the step Please enter the gitlab-ci token for this runner, specify the Runner token.
In the step Please enter the gitlab-ci description for this runner, enter the description gitlab test runner.
In the step Please enter the gitlab-ci tags for this runner, don't type anything. Press Enter.
In the step Please enter the executor, specify shell.

$ sudo gitlab-runner register
Runtime platform                                    arch=amd64 os=linux pid=8197 revision=3afdaba6 version=11.5.0
Running in system-mode.

Please enter the gitlab-ci coordinator URL (e.g. https://gitlab.com/):
http://<ci-gitlab-IP-address>/
Please enter the gitlab-ci token for this runner:
<Runner-token>
Please enter the gitlab-ci description for this runner:
[ci-tutorial-gitlab]: gitlab test runner
Please enter the gitlab-ci tags for this runner (comma separated):

Registering runner... succeeded                     runner=wZqzyy9s
Please enter the executor: virtualbox, docker+machine, docker-ssh+machine, kubernetes, docker, docker-ssh, shell, ssh, parallels:
shell
Runner registered successfully. Feel free to start it, but if it's running already the config should be automatically reloaded!

Install the Yandex.Cloud CLI so that the CI script can create VMs:

$ curl https://storage.yandexcloud.net/yandexcloud-yc/install.sh --output install.sh
$ sudo bash install.sh -n -i /opt/yandex-cloud

To perform functional testing, install the pytest package:
```
$ sudo apt-get install python-pytest
```

Create the test.py file with the functional testing script:

Open the home page of the gitlab-test repository.
Click + and choose New file.
In the window that opens, give the file the name test.py.

Copy the following code to the file body:

# test.py
import pytest
import json
import socket as s


@pytest.fixture
def hostname(request):
    with open("instance-creation.out", "r") as fd:
        fqdn = json.loads(fd.read()).get("fqdn")

    return fqdn


@pytest.fixture
def socket(request):
    _socket = s.socket(s.AF_INET, s.SOCK_STREAM)
    def socket_teardown():
        _socket.close()
    request.addfinalizer(socket_teardown)
    return _socket


def test_server_connect(socket, hostname):
    socket.connect((hostname, 80))
    assert socket

Write any commit message and click Commit changes.

8. Configure CI

You need to define the configuration for CI.

Open the home page of the gitlab-test repository:

http://<Public_IP_of_the_GitLab_VM>/root/gitlab-test

Click Set up CI/CD. You will see the screen for adding a new file.

GitLab automatically gives the file the name .gitlab-ci.yml — don't change it. Copy the following configuration to the file:

#.gitlab-ci.yml
stages:
  - build
  - test

build:
  stage: build
  variables:
    snapshot_name: test-app-snap
    folder_id: <folder-id>
    subnet_name: <subnet-name>
  script:
    - export instance_name="ci-tutorial-test-app-$(date +%s)"
    - export PATH="/opt/yandex-cloud/bin:${PATH}"
    - yc config set token $YC_OAUTH
    - yc compute instance create
      --format json
      --name $instance_name
      --folder-id $folder_id
      --zone ru-central1-c
      --network-interface subnet-name=$subnet_name,nat-ip-version=ipv4
      --create-boot-disk name=$instance_name-boot,type=network-ssd,size=15,snapshot-name=$snapshot_name,auto-delete=true
      --memory 1
      --cores 1
      --hostname $instance_name > instance-creation.out
    - sleep 30
  artifacts:
    when: on_success
    paths:
    - instance-creation.out
    expire_in: 10 mins

test_external:
  stage: test
  script:
    - py.test test.py > pytest-external.out
  artifacts:
    paths:
    - pytest-external.out
    expire_in: 1 hour

In the snapshot_name field, enter a name for the snapshot of the first VM.
In the folder_id field, specify the ID of the folder where the VMs are created.
In the subnet_name field, specify the name of the subnet that the VMs will connect to. To find the name in the management console, open the appropriate folder and go to the Virtual Private Cloud page.
Click Commit changes.

9. Check how the application works on the VM created using CI

After the commit, you need to make sure that CI worked correctly. A new VM should appear in the folder where the test application and web server are deployed.

To check the created VM:

Open the Yandex.Cloud management console.
In the folder where the VMs were created, open the Compute Cloud service.
If everything was configured correctly, the list of VMs should include a new VM with a name like ci-tutorial-test-app-1543910277.
Select the VM you created and copy the public address of the created VM.

In the browser, open a link in this format:

http://<public_IP_address_of_the_created_VM>/test/?text=hello_world

The application created in the previous steps should also work on the created VM by returning reversed words from the text parameter.