Understanding Tracking Branches in Git

Here’s a topic that took me a while to understand in git, and now (I think!) I do, I thought I’d write it all down while I can remember!

Some branches in git (such as your origin/master branch) will usually track the remote branch that they are related to. But what if you want to create a relationship between local and remote branches? Or stop them from tracking? Here’s some pointers

What is a tracking branch?

This is a branch which knows which remote branch it is related to, and making this link allows us to take advantage of some neat git tricks. In particular:

  • You can git push or git pull without specifying remote or branch and it will use its tracking sister branch by default
  • The git status command will include information about how far behind your tracking branch you are – useful to remind you that you haven’t pushed your changes yet! It looks like this:
$ git status
# On branch branch1
# Your branch is ahead of 'origin/branch1' by 1 commit.
#   (use "git push" to publish your local commits)
#
nothing to commit, working directory clean

Which branches are tracking what?

Let’s start with getting a handle on what branches are currently tracking other branches, using git branch -vv

$ git branch -vv
  branch1    c98bacf [origin/branch1] A random change of 24459 to ideas2.txt
  experiment 09a0eb9 A random change of 16603 to list2.txt
* master     09a0eb9 [origin/master] A random change of 16603 to list2.txt

This shows three branches, two of them are tracking branches on another remote. It’s common for the branches to have the same names on the various remotes but it’s only a convention; you can actually call them anything you like (but beware of getting confused!)

Remove a current tracking relationship

This bit is simple, but rarely documented: git branch --unset-upstream

Link a branch to an upstream branch

If you’re pushing a branch (even if it exists in both places) to a remote location and you want to create the link:

$ git push -u origin feature
Counting objects: 9, done.
Delta compression using up to 4 threads.
Compressing objects: 100% (5/5), done.
Writing objects: 100% (9/9), 746 bytes | 0 bytes/s, done.
Total 9 (delta 2), reused 0 (delta 0)
To /home/lorna/.../scripts/origin.git
 * [new branch]      feature -> feature
Branch feature set up to track remote branch feature from origin.

If you’re creating a local version of an existing remote branch:

$ git fetch origin
remote: Counting objects: 9, done.
remote: Compressing objects: 100% (5/5), done.
remote: Total 9 (delta 2), reused 0 (delta 0)
Unpacking objects: 100% (9/9), done.
From /home/lorna/.../scripts/origin
 * [new branch]      feature    -> origin/feature

Then:

$ git checkout feature
Branch feature set up to track remote branch feature from origin.
Switched to a new branch 'feature'

Which branches get pushed/pulled?

This is an area of confusion because the defaults changed between versions of git. Take a look at your config (use git config --list) and find a setting called push.default. The usual default is simple which will only push/pull the branch to/from the branch you’re currently on and the one it tracks. You can optionally set this to be matching (this was once the default) which will then push/pull all changes between all branches and their remote tracking sisters. Sometimes this second option is cumbersome if you have a branch that is not the one you’re working on get into a state where it won’t merge!

Tracking Branches

Tracking branches are a timesaver once you have them set up correctly, so I hope that this collection of notes helps you to do just that :) Did I miss anything? Leave me a comment!

9 thoughts on “Understanding Tracking Branches in Git

  1. Pingback: Video: Git Remotes and Tracking Branches | LornaJane

  2. Pingback: Supporting a Github-Flow workflow – raw commands » Tech Spin

  3. I though the tracking is only work with pull not push? correct me if I’m wrong.

    When you track a remote branch, you can run command “git pull” and it knows which remote branch to pull from.

    When it comes to push, I don’t think you can do this “git push” or “git push origin” or “git push origin/trackbranch”

    • It’s the same settings for push and for pull, the branch is configured to “track” a branch on the remote and that means that as well as “git pull” to get new commits from the remote, you can also “git pull” without additional arguments to push your changes up to that branch (if you have permission)

      • Hi,
        In my case I have a remote tracking branch (tracking my upstream’s master ) so if I do any changes from that tracking branch and do a git push (I have set the git config –global push.default matching) it says “Everything up-to-date” and my changes are not getting pushed.
        I have to use the following command to push the changes directly to my upstream master
        git push upstream HEAD:master
        If I am missing anything please let me know

  4. Great blog!
    I would add that setting upstream of existing branch without pushing can be done with the following command:

    git branch –set-upstream-to=origin/

  5. Pingback: Understanding Tracking Branches in Git - Jacob is studying on web programming

Leave a Reply

Please use [code] and [/code] around any source code you wish to share.

This site uses Akismet to reduce spam. Learn how your comment data is processed.