|
Description:
|
Assume you have a pipeline of jobs, A, which triggers B, which in turn triggers C.
Job A runs (successfully), and triggers B, which then fails (say, it wasn't able
to clean the workspace due to a locked file on the Windows slave). After resolving
the reason for the failure, I don't want to restart the whole pipeline (job A might
take hours to run), and just want to pick up from the point of failure (job B).
Currently, when I do a rebuild on the failed job B, the pipeline chain is broken –
job B's trigger is now "triggered by user", as opposed to "triggered by upstream build".
When job B depends on knowing about the upstream build (to get archived artifacts, for
example), a rebuild of job B can never succeed, because there is no longer an upstream
job.
Rebuilding something that was triggered by an upstream build needs to retain the
information of any upstream build that initially triggered it.
|
