I have a git repository that was converted from a subversion repository that had lots of binaries checked in. I thought I had gotten all of them before doing the initial push to stash but it turns out I was wrong.
I have followed the guide here: https://confluence.atlassian.com/display/BITBUCKET/Maintaining+a+Git+Repositoryto remove the binary files which lowered the repository size from ~2.7gb to ~340mb. I then force pushed the repository up to stash.
However, when I do an initial clone from stash, it still downloads ~2.5gb of files. Once the clone is finished, I can then run "git gc --prune=now" to get the filesize back to ~340mb. This leads me to believe there are some references to the large files which need to be removed on the server, which do not come with a "git clone". My initial thought is it could be related to our merged pull requests.
I did also try manually running the "git gc --prune=now" on the server, but this did not seem have any affect.