Pushing large # of commits fails with " remote unpack failed: index-pack abnormal exit"

John Lindwall January 22, 2020

We have been using bitbucket for a year or so. The issue we have is when pushing a large-ish number of commits (say 40-60).  When we push, the remote returns this error (edited a bit to hide internal details):

C:\mycode>git push --set-upstream origin feature/mybranch

Enumerating objects: 1387, done.
Counting objects: 100% (1351/1351), done.
Compressing objects: 100% (356/356), done.
remote: usage: git index-pack [-v] [-o <index-file>] [--keep | --keep=<msg>] [--verify] [--strict] (<pack-file> | --stdin [--fix-thin] [<pack-file>])
fatal: sha1 file '<stdout>' write error: Broken pipe
error: remote unpack failed: index-pack abnormal exit
To ssh://mybitbucket:2222/mycode.git
! [remote rejected] feature/mybranch -> feature/mybranch (unpacker error)
error: failed to push some refs to 'ssh://mybitbucket:2222/mycode.git'

If I painstakingly/manually push the commits one by one, it will work.  However this is time consuming; I should be able to push all of my local commits in one push.

I've played with some client-side settings regarding pack limits to no avail.  I feel like something is wrong on the server because of that "remote: usage: git index-pack" output?

Any ideas? Thanks!

1 answer

0 votes
Mikael Sandberg
Community Leader
Community Leader
Community Leaders are connectors, ambassadors, and mentors. On the online community, they serve as thought leaders, product experts, and moderators.
January 24, 2020

Hi @John Lindwall,

Welcome to Atlassian Community.

Have you checked the disk space on your Bitbucket server? Does this only happen for one user or all your users? If only for one it could indicate that that clone is corrupted. Have you tried cloning the repository again and then push in the new one?   

John Lindwall January 29, 2020

Thanks for the response.  I have opened a support case for this issue with Atlassian and they are working through various investigations with me.

Disk space on the server is fine. Has happened to several different users using several different repos.

Suggest an answer

Log in or Sign up to answer
TAGS
AUG Leaders

Atlassian Community Events