Any idea on how this git fetch can be optimized?

Question

git fetch origin %Branch_SourcePath% --depth 100 --no-tags --progress --verbose --recurse-submodules 2>&1

The command above takes about 1 hour to complete which is too long. Using --depth 1 completes the fetch in under 4 minutes however I do need the commit information for at least the last 100 commits.

Majority of the time is spent on download/receiving objects.

Receiving objects: 99% (764574/764593), 25.39 GiB | 13.32 MiB/s

Answer 1

option #1: sparse checkout

If you only need a subtree of the original directory tree, one thing you can try is the so-called sparse checkout.

For example, lets say we are interested only in /foo subdirectory.

git clone --sparse --depth 10 --branch master https://github.com/Acme/helloworld

Change directory into the newly-cloned repository and enable sparse-checkout mechanism (to fetch parts of the subtree in steps of increasing depth):

cd helloworld
git sparse-checkout init --cone

Checkout a subtree, say at /foo:

git sparse-checkout set foo

option #2: filtering BLOBs

Another thing you can try is to filter BLOB objects (--filter blob:none in git-clone command).