Gitlfs push is stuck on "Uploading LFS objects:100%"

[Mahesh Khambadkone]

Following these instructions to convert a repository to LFS using BFG (https://confluence.atlassian.com/bitbucket/use-bfg-to-migrate-a-repo-to-git-lfs-834233484.html), git is stuck at 100% during the push.

    git push --force

    Uploading LFS objects: 100% (7574/7574), 662 MB | 0 B/s  

Repeated attempts to push the content results in the same outcome.

FYI The push previously failed with multiple errors 

    Remote "origin" does not support the LFS locking API. Consider disabling it with:
$ git config lfs.https://bitbucket.org/xxxxx/repo.git/info/lfs.locksverify false

and this error was resolved by executing :

    git config lfs.https://bitbucket.org/XXXX/repo.git/info/lfs.locksverify false

[Ben_Dol]

Was this resolved?

[Scott Moore]

Unfortunately, this step is not very communicative about what is happening.  Once you see the "Uploading LFS objects: 100%" message, the system begins scanning ALL blobs that are under 1024 bytes to determine if they are LFS objects. Because of some system limitations, this operation is performed 100 items at a time, which can not be adjusted.  With many branches/objects, this can be excruciatingly slow!  Most likely this is not actually hung, just in the many-hours slow category.   I would expect if you let this run for enough time it would complete.  I confirmed this on my most recent transfer, which took many hours at this step with no feedback until complete.  

It would be nice if Atlassian added this warning to the documentation for converting BitBucket repositories to LFS mentioned in the original post. 

[Elliott Chenger]

Thanks for this response @Scott Moore this is what I assumed based on looking at the Process tab in the Activity Monitor. What is the best way to tell the progress of this step or is it literally just a sit and wait until it is done? 

I am curious also how you came to know that this was happening under the hood?

[Ana Retamal]

Hi Mahesh! We'll need some more info to help you troubleshoot this issue. I have created a Support ticket for you and you'll receive an email notification shortly. Please follow up in the ticket, send us your logs, repository URL and LFS size and we'll assist you until this problem is solved.

Let us know if you need anything else!

Best regards,

Ana

[Markus_Schicker]

We encounter the exact same issue regarding the 100% and then everything is stalled. GIT_TRACE=1 git push --force shows repeatedly the following:

13:33:14.947974 trace git-lfs: tq: starting transfer adapter "basic"
13:33:14.948037 trace git-lfs: tq: running as batched queue, batch size of 100
13:33:14.948167 trace git-lfs: run_command: git rev-list --stdin --objects --not --remotes=origin --
13:33:14.953508 trace git-lfs: run_command: git cat-file --batch
13:33:16.637977 trace git-lfs: tq: sending batch of size 0
13:33:16.638006 trace git-lfs: tq: running as batched queue, batch size of 100
13:33:16.638096 trace git-lfs: run_command: git rev-list --stdin --objects --not --remotes=origin --
13:33:16.643522 trace git-lfs: run_command: git cat-file --batch
13:33:18.640716 trace git-lfs: tq: sending batch of size 0
13:33:18.640753 trace git-lfs: tq: running as batched queue, batch size of 100
13:33:18.640860 trace git-lfs: run_command: git rev-list --stdin --objects --not --remotes=origin --
13:33:18.647560 trace git-lfs: run_command: git cat-file --batch
13:33:20.849759 trace git-lfs: tq: sending batch of size 0
13:33:20.849787 trace git-lfs: tq: running as batched queue, batch size of 100
13:33:20.849903 trace git-lfs: run_command: git rev-list --stdin --objects --not --remotes=origin --
13:33:20.854547 trace git-lfs: run_command: git cat-file --batch
13:33:22.861807 trace git-lfs: tq: sending batch of size 0
13:33:22.861836 trace git-lfs: tq: running as batched queue, batch size of 100
13:33:22.861948 trace git-lfs: run_command: git rev-list --stdin --objects --not --remotes=origin --
13:33:22.862468 trace git-lfs: run_command: git cat-file --batch
13:33:24.989772 trace git-lfs: tq: sending batch of size 0
13:33:24.989800 trace git-lfs: tq: running as batched queue, batch size of 100
13:33:24.989897 trace git-lfs: run_command: git rev-list --stdin --objects --not --remotes=origin --
13:33:24.991228 trace git-lfs: run_command: git cat-file --batch
13:33:27.010523 trace git-lfs: tq: sending batch of size 0
13:33:27.010551 trace git-lfs: tq: running as batched queue, batch size of 100
13:33:27.010644 trace git-lfs: run_command: git rev-list --stdin --objects --not --remotes=origin --
13:33:27.011967 trace git-lfs: run_command: git cat-file --batch
13:33:29.128833 trace git-lfs: tq: sending batch of size 0
13:33:29.128864 trace git-lfs: tq: running as batched queue, batch size of 100
13:33:29.129429 trace git-lfs: run_command: git rev-list --stdin --objects --not --remotes=origin --
13:33:29.133546 trace git-lfs: run_command: git cat-file --batch
13:33:31.273945 trace git-lfs: tq: sending batch of size 0
13:33:31.273977 trace git-lfs: tq: running as batched queue, batch size of 100
13:33:31.274115 trace git-lfs: run_command: git rev-list --stdin --objects --not --remotes=origin --
13:33:31.282657 trace git-lfs: run_command: git cat-file --batch

[Zongkun Dou]

I am having exactly the same issue right now when I do lfs migration. Could someone give some help on how to resolve this issue?

[Christian]

Hi,

Got the issue as well. Believe it has to do with credentials. Solved by completely wiping my git and source tree installations. When I set it up again it worked. In my previous installation I also had credentials set for gitlab (don't know if that was the problem). When asked by credential helper in current installation I used manager option.