wassname/ray
1
0
Fork 0
mirror of https://github.com/wassname/ray.git synced 2026-07-04 15:05:49 +08:00
Code Issues Packages Projects Releases Wiki Activity

Actor fault tolerance using object lineage reconstruction (#902)

Browse Source 
* Revert Python actor reconstruction

* Actor reconstruction using object lineage

* Add dummy arguments and return values for actor tasks

* Pin dummy outputs for actor tasks

* Skip checkpointing test for now

* TODOs

* minor edits

* Generate dummy object dependencies in Python, not C

* Fix linting.

* Move actor counter and dummy objects inside of the actor handle

* Refactor Worker._process_task, suppress exception propagation for
sequential actor tasks
This commit is contained in:
Stephanie Wang
2017-09-10 19:29:28 -07:00
committed by Robert Nishihara
parent d8aa826e63
commit 99c8b1f38c
7 changed files with 207 additions and 258 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/ray/workers/default_worker.py
-11
View File
@@ -79,17 +79,6 @@ if __name__ == "__main__":
ray.worker.connect(info, mode=ray.WORKER_MODE, actor_id=actor_id)
# If this is an actor started in reconstruct mode, rerun tasks to
# reconstruct its state.
if args.reconstruct:
try:
ray.actor.reconstruct_actor_state(actor_id,
ray.worker.global_worker)
except Exception as e:
redis_client = create_redis_client(args.redis_address)
push_error_to_all_drivers(redis_client, traceback.format_exc())
raise e
error_explanation = """
This error is unexpected and should not have happened. Somehow a worker
crashed in an unanticipated way causing the main_loop to throw an exception,