0

We have a workflow A that we run on self hosted mac. Now this workflow has a step a that requires access to some system resource 1

Now when we run multiple instances of workflow A at the same time, all of them try to access the system resource 1 at the same time and eventually all except 1 that has access and lock on the resource fail.

When we run the failed jobs again, one of the again succeeds.

Now the only way I can think of solving this problem is calling the github api and check whether any of the instances of workflow A are on step a and then wait until that step completes.

I wonder if there is a better way to solve this problem.

DockerNaughty
  • 163
  • 1
  • 5
  • Are you running multiple runners on the same physical machine? It's for reasons like this the guidance is to *not* run multiple runner agents on the same machine. Resource contention and deadlocks will arise and be extremely difficult to troubleshoot. – tj-cappelletti Oct 07 '22 at 01:52

0 Answers0