Hi,
That’s a plausible scenario, altho TBH, race condition didn’t occur to me. I just assumed the job that started first finished first. But yes, in the scenario you describe (B finishes first) that’s exactly the behavior you would get.
Coming back to the initial branch conflict, in very quickly reviewing the thread I notice that you’re on 6.7.3. The current version of the LTS is 6.7.6, altho looking over the release notes (6.7.4, 6.7.5, 6.7.6) I don’t anything that seems immediately relevant to the question.
Before you go to a project-per-branch strategy (because that will explode your instance LOC & therefore increase your bill) let’s take another look at this. You say the error disappeared when you disabled the SCM integration. Can you turn it back on and see if the problem comes back? And if it does, could you bump up your server log level to TRACE just log enough to reproduce this error (Administration > System > Logs level) and post any relevant log excerpts? You’ll want to turn the logging level back down as soon as you’ve reproduced the error, BTW, since TRACE logging impacts performance and produces lots of logs.
Ann