Add new server to GitLab AutoDeploy

2017-05-01T09:26:53Z

84.47.33.253: Installing autostarting Git auto Deploy on Windows.

If not mentioned otherwise, the following values are variables and will change from project to project (or server to server):

* dev-server - Git branch name, usually following the name of the server which the files will be deployed to (can be uat-server or live-server, etc)
* the /var/www/html path is just an example, this will be the root folder of the website we want to deploy to (from nginx/apache sites-enabled)
* the <nowiki>https://gitlab.apoyar.eu/somepath</nowiki> will be the url to the project in question - to be obtained from GitLab

=== And now the steps: ===
==== Linux servers: ====

* Install Git AutoDeploy
** apt-get install software-properties-common
** add-apt-repository ppa:olipo186/git-auto-deploy
** apt-get update
** apt-get install git-auto-deploy
* copy the settings file from Dev: /etc/git-auto-deploy.conf.json
* amend the file contents: url to access (GitLab project), active branch required
* open the /etc/init.d/git-auto-deploy file and make the daemon running as www-data:
** DAEMON_UID="www-data"
** DAEMON_GID="www-data"
* Run systemctl daemon-reload after you’ve made the changes above
* A reboot may be required here
* Make sure the Deploy Bot account has access to the project and is made a Master
* Backup the current website contents: cp -R /var/www/html/ /home/neal/backup/
* Delete the website contents, including the git part (if present):
** cd /var/www/html
** rm –rf *
** rm –rf .git .gitignore .htaccess
* Before you do the below process, you may need to check the owner of the /var/log/git-auto-deploy.log file. It should be www-data. Also check the /var/run/git-auto-deploy/git-auto-deploy.pid If not, change them:
** chown www-data:www-data /var/log/git-auto-deploy.log
** chown www-data:www-data /var/run/git-auto-deploy/git-auto-deploy.pid
* Start the daemon in foreground, but as the user www-data: sudo -H -u www-data bash -c 'git-auto-deploy --pid-file /var/run/git-auto-deploy/git-auto-deploy.pid --log-file /var/log/git-auto-deploy.log --config /etc/git-auto-deploy.conf.json'
* Once static, Ctrl-C the process
* It should create the tree
* Check the contents of the /var/www/html/.gitignore file
* If not present, create one, own it by www-data and push back to the server. If present, leave out the 3 next steps
** vi .gitignore
** chown www-data:www-data .gitignore
** git push origin dev-server
* Copy the ignored files listed above from backup
* Clear the cache and push changes to GitLab:
** cd /var/www/html
** git rm –rf –cached .
** git add .
** git status
** if the above command says nothing to commit, you can leave out the 2 next steps
** git commit
** git push origin dev-server
* Make sure all objects in the .git/objects folder are owned by www-data: chown –R www-data:www-data *
* Change the Deploy Bot to be a Reporter in GitLab
* Start the service: service git-auto-deploy start
* Observe the log file: tail –f /var/log/git-auto-deploy.log
* If you don’t see anything in the log and the foreground app runs fine, reboot the server
* Add the <nowiki>http://hostname:8001</nowiki> as webhook to GitLab project in question: <nowiki>https://gitlab.apoyar.eu/root/bleckmann/settings/integrations</nowiki>
* Reboot the server and check whether it’s listening on port 8001: netstat -anp

==== Windows servers: ====

* All installation files and prerequisites are in the \\ncc1701d\SourceFiles\Git network folder
* Install the Windows Git client (Git-2.11.1-64-bit)
* Install Python for Windows (pywin32-220.win-amd64-py2.7)
* Reboot
* Make sure you can run Python from command line without specifying the installation path (open CMD, then type python -v, you should get the python command prompt. Close it by typing 'quit()')
* Make sure you can run Git from command line. Just type git in the CMD window
* If the above two are not working, just check the Windows PATH environmental variable in the computer's properties
* Now, installing GitAutoDeploy:
** Open the Command Prompt as admin
** cd to 'C:\inetpub\wwwroot'
** run 'git clone https://github.com/olipo186/Git-Auto-Deploy.git'
** cd to newly created subfolder: 'cd Git-Auto-Deploy'
** run the Python requirements checker 'pip install -r requirements.txt' It will check and install all required Python files, if needed.
* From the \\ncc1701d\SourceFiles\Git network path copy these files to the C:\inetpub\wwwroot\Git-Auto-Deploy folder:
** gitautodeploy.bat
** GitAutoDeploy.py
** invisible.vbs
** config.json
* Please note that if the 'Hide extensions for known file types' option is ticked, the above 4 files will be visible without their extension!
* Create a folder which will contain the IIS files. In this example I'll be using 'C:\inetpub\wwwroot\BleckmannAPItest'
* You will point your new IIS website to the above folder.
* Edit the config.json file according to your needs. Particularly change the:
** GitLab url
** Branch to manage (uat-server)
** Local path for the files ('C:/inetpub/wwwroot/BleckmannAPItest')
*** Please note I'm using forwardslash instead of backslash. This is the way it needs to be entered in the config.json file!
* Should you need to serve more than one project, please refer to the complex-config.json file in the same network folder.
* Now get the initial copy of the website from GitLab:
** start CMD and browse to your new empty folder: 'cd C:\inetpub\wwwroot\BleckmannAPItest'
** run git clone https://root@gitlab.apoyar.eu/root/BleckmannAPI.git .
** don't forget the ending dot after the url as this makes sure the files will be saved to the current folder
* Check the contents of the folder
* Add the http://hostname:8001 as webhook to GitLab project in question: https://gitlab.apoyar.eu/root/BleckmannAPI/settings/integrations
* Again, the above url is '''just an example'''
* As for the hostname, it's always good to check that the computer name (for example bleckmannapi.apoyar-networks.com) is pingable from Linuxweb, which is the machine GitLab runs on
* Time for the console test:
** Open CMD
** copy and paste 'python C:\inetpub\wwwroot\Git-Auto-Deploy\GitAutoDeploy.py --config C:\inetpub\wwwroot\Git-Auto-Deploy\config.json'
* You'll get the initialization messages and the script should stop at 'listening on'
* Go to GitLab, project in question, settings, Integrations
* Find the entry you have just added recently and hit Test
* You should get the 'Hook executed succesfully' blue bar at the top.
* Now the startup:
** Open the registry editor: start, run, 'regedit'
** Browse to the: HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Run
** Add a new 'Expandable String Value'
** Name it 'GitAutoDeploy'
** Copy and paste this value: "%SystemRoot%\System32\WScript.exe" "C:\inetpub\wwwroot\Git-Auto-Deploy\invisible.vbs" "C:\inetpub\wwwroot\Git-Auto-Deploy\gitautodeploy.bat"
** Reboot the server and '''log on as administrator'''
** The above is required as the python script won't otherwise run in the background
* Test

Apoyar Wiki - User contributions [en-gb]

Add new server to GitLab AutoDeploy