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
It would make sense to always use the termination detection if the Distributed Clock is used from mango/util/distributed_clock.py.
The distributed clock requires, that one agent per container exists to keep track of the time.
After the fix on single container termination detection is merged !69, the distributed termination is reduced to one per container so that we can use this in the same way - and add the termination detection directly to DistributedClockManager (NegotiationTerminationDetectorRole) and DistributedClockAgent (NegotiationTerminationParticipantRole).
The text was updated successfully, but these errors were encountered:
In GitLab by @maurerle on Apr 19, 2023, 17:20
Running mango in a distributed way is part of the target functionality.
Therefore a reliable termination detection in distributed containers is needed too.
Currently this exists in the mango-library:
https://gitlab.com/mango-agents/mango-library/-/blob/96955e59b7f44295fa828752bbf5b4a17131baec/mango_library/negotiation/termination.py
Here, we need to add the role to every single agent in each container.
It would make sense to always use the termination detection if the Distributed Clock is used from
mango/util/distributed_clock.py
.The distributed clock requires, that one agent per container exists to keep track of the time.
After the fix on single container termination detection is merged !69, the distributed termination is reduced to one per container so that we can use this in the same way - and add the termination detection directly to
DistributedClockManager
(NegotiationTerminationDetectorRole
) andDistributedClockAgent
(NegotiationTerminationParticipantRole
).The text was updated successfully, but these errors were encountered: