You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Originally reported by: Martin Uhrin (Bitbucket: muhrin, GitHub: muhrin)
If you, say, call run(...) in a workflow to launch a remote calculation this call will block until complete. In the meantime the daemon will tick the workflow engine again and perform the first step, relaunching the calculation. This should no happen.
In general we need a blocking mechanism to deal with workflows that are currently being executed so they aren't ran twice.
Even more generally we should probably move to a work queue model (instead of the polling model) which would more cleanly sort out issues like this one.
This is fixed both in the released 0.11.0 as well as in the current workflows branch which contains the new workflow engine, which will be merged into develop soon for 0.12.0
Originally reported by: Martin Uhrin (Bitbucket: muhrin, GitHub: muhrin)
If you, say, call run(...) in a workflow to launch a remote calculation this call will block until complete. In the meantime the daemon will tick the workflow engine again and perform the first step, relaunching the calculation. This should no happen.
In general we need a blocking mechanism to deal with workflows that are currently being executed so they aren't ran twice.
Even more generally we should probably move to a work queue model (instead of the polling model) which would more cleanly sort out issues like this one.
The text was updated successfully, but these errors were encountered: