23 Remotes

Remote repositories are versions of your project that are hosted on the Internet or another network. A single project can have 1, 2, or even hundreds of remotes. You pull others’ changes from remotes and push your changes to remotes.

23.1 Listing what remotes exist

git remote lists the names of available remotes, but usually it is more useful to see what URLs each note corresponds to (with -v).

git remote -v
## origin   https://github.com/jennybc/happy-git-with-r (fetch)
## origin   https://github.com/jennybc/happy-git-with-r (push)

23.2 Adding a new remote

git clone automatically adds a new remote, so often you do not need to do this manually initially. However, after the initial clone, it is often useful to add additional remotes.

Use git remote add to add a new remote:

git remote add happygit https://github.com/jennybc/happy-git-with-r.git

Note: when you add a remote you give it a nickname (here happygit), which you can use in git commands in place of the entire URL.

git fetch happygit

Sidebar on nicknames: there is a strong convention to use origin as the nickname of your main remote. At this point, it is common for the main remote of a repo to be hosted on GitHub (or GitLab or Bitbucket). It is tempting to use a more descriptive nickname (such as github), but you might find that following convention is worth it. It makes your setup easier for others to understand and for you to transfer information that you read in documentation, on Stack Overflow, or in blogs.

A common reason to add a second remote is when you have done a “fork and clone” of a repo and your personal copy (your fork) is set up as the origin remote. Eventually you will want to pull changes from the original repository. It is common to use upstream as the nickname for this remote.

git remote add upstream https://github.com/TRUE_OWNER/REPO.git

23.3 Fetching data from remotes

To get new data from a remote use git fetch <remote_name>. This retrieves the data locally, but importantly it does not change the state of your local files in any way. To incorporate the data into your repository, you need to merge or rebase your project with the remote project.

# Fetch the data
git fetch happygit

# Now merge it with our local main
git merge happygit/main main

# git pull is a shortcut which does the above in one command
git pull happygit main

For more detail on git pull workflows, see 29.

23.4 Pushing to remotes

Use git push <remote> <branch> to push your local changes to the <branch> branch on the <remote> remote.

# push my local changes to the origin remote's main branch
git push origin main

# push my local changes to the happygit remote's test branch
git push happygit test

23.5 Renaming and changing remotes

git remote rename can be used to rename a remote:

git remote rename happygit hg

git remote set-url can be used to change the URL for a remote. This is sometimes useful if you initially set up a remote using HTTPS, but now want to use SSH instead (or vice versa).

git remote set-url happygit git@github.com:jennybc/happy-git-with-r.git

One fairly common workflow is you initially cloned a repository on GitHub locally (without forking it), but now want to create your own fork and push changes to it. As described earlier, it is common to call the source repository upstream and to call your fork origin. So, in this case, you need to first rename the existing remote (from origin to upstream). Then add your fork as a new remote, with the name origin.

git remote rename origin upstream
git remote add origin git@github.com:jimhester/happy-git-with-r.git

23.6 Upstream tracking branches

It is possible to set the branch on the remote each of your local remotes corresponds to. git clone sets this up automatically, so for your own main branch this is not something you will run into. However by default if you create a new branch and try to push to it you will see something like this:

git checkout -b mybranch
git push
# fatal: The current branch foo has no upstream branch.
# To push the current branch and set the remote as upstream, use
#     git push --set-upstream origin foo

You can do as the error message says and explicitly set the upstream branch with --set-upstream. However I would recommend instead changing the default behavior of push to automatically set the upstream branch to the branch with the same name on the remote.

You can do this by changing the git push.default option to current.

git config --global push.default current

See also Working with Remotes: