Deploy Hexo to ECS

preface

Start operation

This article is not a tutorial that starts with installation. This tutorial is built after you have passed the http://localhost:4000 / visit your blog based tutorial. Therefore, Node installation and configuration, Hexo download, etc. are not involved. If you haven 't finished it, you can do Baidu by yourself or read my previous hydrology article: Building Hexo blog from scratch

1, SSH key generation via Git

// If you have not set the global information of git, you need to set it first. It can be ignored git config --global user.name "yourname" git config --global user.email [email protected] // Generate SSH key ssh-keygen -t rsa -C "[email protected]" // Disable automatic conversion, which will cause a warning when uploading later git config --global core.autocrlf false

Then the key generated by the students of Windows is under the path of C:\Users\yourname\.ssh, but I don't know the little partners of Linux or MAC. After all, I haven't used these two operating systems.

2, Configure GIT server

To connect to the server, you can use tools such as Xshell or FinalShell. Here I have Xshell as an example, another tool is the same.

1. logon server

   Since it's a newly reinstalled system, it's easy to log in. See the white characters on black as before.

   

2. Check if git exists, if not, install it

   Use the git --version command to check whether there is a git command. I will display not found here. It means there is no GIT and needs to be installed.

   

3. Install Git

   yum install curl-devel expat-devel gettext-devel openssl-devel zlib-devel perl-devel -y yum install -y git

   

   The installation will be finished soon!

   

4. Create users and configure their warehouses

   // Create user useradd git // Set password passwd git // Switch users su git // Enter git user's home directory cd /home/git/ // Real directory where the project exists (you can create it elsewhere, of course) mkdir -p projects/blog // create folder mkdir repos && cd repos // Create a bare warehouse git init --bare blog.git

   

   cd blog.git/hooks // Create the hook function. The input is as follows vi post-receive //-------------Split lines do not need to be copied------------- #!/bin/sh git --work-tree=/home/git/projects/blog --git-dir=/home/git/repos/blog.git checkout -f //-------------Split lines do not need to be copied-------------

   [the external link image transfer failed. The source station may have anti-theft chain mechanism. It is recommended to save the image and upload it directly (img-adabrtin-1592708755684)( https://cdn.jsdelivr.net/gh/blogimg/HexoStaticFile2@latest/2020/05/20/38a4628161b8cc3619c17bc6c66eca8e.png ]

   

   Press and hold ESC, then output: wq can save to exit!

   // Add executable permissions for the file you just created chmod +x post-receive // Log out to root exit // add permission chown -R git:git /home/git/repos/blog.git

   

5. Test whether it can be clone

   Create a new folder anywhere on your computer, and switch bash or other command-line tools here. Start executing the following command (note to replace it with your server ip) to test whether the warehouse can be clone down.

   git clone git@server_ip:/home/git/repos/blog.git

   

6. Establish SSH trust relationship (password free login)

   Enter the following command on your own computer (note replace with your file path)

   ssh-copy-id -i C:/Users/yourname/.ssh/id_rsa.pub git@server_ip // Test whether you can log in ssh git@server_ip

   It's over!?

   Well, I don't have the SSH copy ID command, but it doesn't matter. Since I can't use the command, I will manually. If you have no problem with the above two commands, you can skip the process of manually establishing trust relationship.

   Manual settings:

   1. Execute the following commands in turn

      // Switch to git user su git //Enter the home directory. If there is no. ssh directory, it needs to be created. If there is direct access cd ~ //Create. ssh directory mkdir .ssh // Grant authority chmod 700 .ssh/ // Enter the. ssh directory cd .ssh // Edit an authorized_keys and write the public key(id_rsa.pub)content vi authorized_keys // Give permission after adding chmod 600 authorized_keys

      

      

      If the command in the picture is different from the command given by the code, please refer to the given code. The picture is only for reference.

   2. Test whether the connection is successful on your own computer.

      ssh git@server_ip

      A password is required to log in.

      

   Pass SSH after completing the above steps git@server_ IP connection to the server does not require a password. If you still need a password, please check if you missed any step

7. For security, disable the GIT user's shell login rights. So you can only log in with git clone, git push, etc

   Enter the following command on the server side:

   // If not root, switch to root su root // Check whether git shell is in login mode cat /etc/shells // Check to see if it is installed which git-shell vi /etc/shells // Add the Lu Jin shown in the previous 2 steps, usually in / usr / bin / git shell

   

   

   Modify permissions in / etc/passwd

   // Modify passwd file vi /etc/passwd

   // Change the original git:x:1000:1000::/home/git:/bin/bash // Change to git:x:1000:1000::/home/git:/bin/git-shell

   

3, Configure NGINX server

This step is mainly used to open the 80 port server. You don't think you're going to use hexo s, do you?

1. Install NGINX server

   // Installation dependency sudo yum install -y yum-utils // Install nginx service sudo yum install -y nginx

2. Check for successful installation.

   

   See the output version number and represents success.

3. Configure NGINX server

   //Running nginx directly nginx //Stop nginx first nginx -s stop //Switch to the configuration file directory of nginx cd /etc/nginx //Edit file vi nginx.conf // At the same time, otherwise nginx cannot access

   Change user to root as shown in the figure below

   

   Modify the root resolution path (/ home/git/projects/blog /), as shown below

   

   nginx -s reload

   At this time, you can see a prompt - 404 page by directly visiting your server ip address in the browser

4, Configure Hexo Publishing

1. Configure the_ config.yml file

   deploy: type: git #user name repo: git@server_ip:/home/git/repos/blog.git #Git warehouse branch: master #branch

   

2. At package.json Add npm script to

   "scripts": { "deploy": "hexo clean && hexo g -d", "start": "hexo clean && hexo g && hexo s" },

   

3. In this case, npm start is used for local debugging. After debugging, it will be uploaded to the server, and then it can be accessed through the IP of the server!

   

   

Bind domain name

Modify the file of NGINX directly. Then the domain name is resolved to the server with A record.

Configure HTTPS (SSL)

Please refer to Baidu!