Git isn't Github


A lot of developpers use Git everyday, a high percentage of them probably use Github.

I use Github every work day. I use it to the extent that I use the words Git and Github interchangeably, but this is technically wrong. Github is a hosting service for your Git projects, like Gitlab. these services offer a bunch of bells and whistles, but you can also self host your Git projects. You probably already do if you have a cloned a Git repo before. But you can also host your repos on a remote server that you or someone else owns.

This might seem obvious, but somehow it never really occured to me that I could pay for a server or data storage somewhere (like AWS, GCP, Azure to name a few) and host my Git repos on there.

I realized this by cloning a repo, hosted on a virtual machine in GCP, at work that hadn't been commited to Github.

ANYWAY, the procedure is similar to the one used for cloning a repo from Github.

We need to have SSH access to the remote server. For that we need to generate a ssh key like so:

$ cd ~/.ssh && ssh-keygen -t rsa -b 4096 -C ""

lets break that down a bit:

  • ssh is a short-hand for Secure SHell protocol, ssh is a network protocol to securely communicate between computer 1,
  • ssh-keygen is a CLI tool used to generate key pairs for ssh, this tool can take several options (or flags?)
  • the -t in our command is to specify the type of algorithm used to encrypt our key, in our command we're using the rsa algorithm,
  • the -b option specifies the number of bits in the key to create, in our command we're using 4096 bits,
  • finally, the -C command is a comment, this will add our to our public key (I believe that this is used on Github for authentication reasons, but i could be wrong.)

Great, we have a ssh key pair, but we need to add the generated public key in the authorized_keys folder on the remote server.

There are a couple of ways to do this. For each method you need to know the ip address of the server, and the user. You can define the user when setting up your server or virtual machine. Figuring out the ip might be a bit trickier if you don't have a static ip (in GCP you can request a static ip, or use their CLI that gives you access to all you servers).

ssh-copy-id -i ~/.ssh/my-new-sshkey user@host

The ssh-copy-id does what it says. It copies your ssh id to your remote host.

If your interacting with servers, you'll be ssh'ing into them regularly like so:

ssh user@host -i ~/.ssh/my-new-sshkey

This quickly gets old, and for several ssh type commandes you'll need to supply this info. We can add this information into our ~/.ssh/config:

ForwardAgent yes

Host my-remote-server
User user
IdentityFile ~/.ssh/my-new-sshkey.key

Now you can ssh into your server using this config

ssh my-remote-server

same for ssh-copy-id

ssh-copy-id my-remote-server

Now after having spent quite some time setting up ssh access to our server, we can finally clone/push the repo!

# normal cloning
git clone user@host:my-repo.git
#cloning, but with ssh config set
git clone my-remote-server:my-repo.git

#pushing your local repo to your remote server
git remote add origin user@host:my-repo.git
#with ssh config
git remote add origin my-remote-server:my-repo.git

checkout more about ssh at