Use Dask instead of multiprocessing module #530
Conversation
andrii-i
force-pushed
the
dask
branch
6 times, most recently
from
027c090
to
ece6ccc
Compare
|
Kicking CI |
andrii-i
force-pushed
the
dask
branch
3 times, most recently
from
12993c1
to
d753850
Compare
| """ | ||
| Placeholder method for a cleanup code to run when the server is stopping. | ||
| """ | ||
| pass |
There was a problem hiding this comment.
Do we need to keep this method if it does nothing?
There was a problem hiding this comment.
Yes. BaseScheduler is a base class / klass for schedulers, essentially an interface for all schedulers instantiated and used by SchedulerApp. BaseScheduler.stop_extension creates a placeholder for scheduler cleanup / stop_extension logic that is called from SchedulerApp.stop_extension.
SchedulerApp.stop_extension is a public method inherited from ExtensionApp called by Jupyter Server when the server is stopping.
I implemented stop_extension in default Scheduler in Scheduler.stop_extension to correctly terminate Dask and avoid JupyterLab process not shutting down immediately on interrupt as for example we've seen in Jupyter AI.
andrii-i
force-pushed
the
dask
branch
2 times, most recently
from
12dc364
to
768753f
Compare
…ction into scheduler class at extension launch
Use Dask instead of multiprocessing module