Pipeline started to fail all the time from 6 days ago

[Xingyu Wu]

Our pipeline has started to fail all the time from 6 days ago.

We are running a Rails application in the pipeline and we have discovered that when our file uploader (CarrierWave) tries to copy the original asset file to project_root/tmp folder before processing it, the generated file in the project_root/tmp folder is empty.

As far as I know, no related code towards file upload has changed since out last successful running. I have tried to re-run a previously successful pipeline run but it again failed.

Our issue might be related to this post here since indeed different docker engine version yielded different results: https://community.atlassian.com/t5/Bitbucket-Pipelines-questions/Has-the-docker-version-of-bitbucket-pipeline-been-upgraded-to/qaq-p/1685651

Any help would be greatly appreciated. Thanks.

Kyle

[Xingyu Wu]

Reason and solution (monkey-patching FileUtils) can be found here (Anthony's reply): https://community.atlassian.com/t5/Bitbucket-Pipelines-questions/Has-the-docker-version-of-bitbucket-pipeline-been-upgraded-to/qaq-p/1685651