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
Rarely I get this error when terminating a parallel processing run. A race condition or deadlock in the distributed gc / teardown logic?
ERROR (unhandled task failure): ArgumentError: stream is closed or unusable
in check_open at stream.jl:299in uv_write at stream.jl:736in flush at stream.jl:790in anonymous at multi.jl:179in lock at stream.jl:812in send_msg_ at ./multi.jl:170in remote_do at multi.jl:702in flush_gc_msgs at ./multi.jl:163in flush_gc_msgs at ./multi.jl:189in anonymous at task.jl:478
Is there a way to know that the process is shutting down? In normal circumstances, I would like the task failure to be printed, just needs to be ignored during exit.
@andreasnoack , yes, now that we are printing unhandled errors in tasks, they will show up in logs now and will need to be handled. I think it is better to create separate issues for each different error.
Rarely I get this error when terminating a parallel processing run. A race condition or deadlock in the distributed gc / teardown logic?
cc @amitmurthy
The text was updated successfully, but these errors were encountered: