report_recover_strategy(fault_ranks: dict, strategy_list: list) -> int
The client reports the recovery policies supported by itself to the server so that the server can select the optimal policy. The server then delivers the policy to the client through the persistent connection constructed by start_subscribe.
Parameter |
Type |
Description |
|---|---|---|
fault_ranks |
Dictionary |
Rank of the faulty process |
strategy_list |
List |
Recovery policy list |
Type |
Description |
|---|---|
Integer |
Returns 0 if the operation is successful; returns a non-0 value if the operation fails. For details, see Return Codes. |
Parent topic: Elastic Agent (APIs Related to Resumable Training)