elastic_ep: Fix stateless group port races (#36330)
Signed-off-by: Itay Alroy <ialroy@nvidia.com>
This commit is contained in:
@@ -162,10 +162,8 @@ class ElasticEPScalingExecutor:
|
||||
new_dp_size=new_dp_size,
|
||||
new_world_size_across_dp=new_world_size_across_dp,
|
||||
master_ip=reconfig_request.new_data_parallel_master_ip,
|
||||
world_group_ports=reconfig_request.new_stateless_world_group_port_list,
|
||||
dp_group_ports=reconfig_request.new_stateless_dp_group_port_list,
|
||||
ep_group_ports=reconfig_request.new_stateless_ep_group_port_list,
|
||||
eplb_group_ports=reconfig_request.new_stateless_eplb_group_port_list,
|
||||
coord_store_port=reconfig_request.coord_store_port,
|
||||
enable_eplb=updated_config.parallel_config.enable_eplb,
|
||||
)
|
||||
self.worker.model_runner.eep_eplb_suppressed = True
|
||||
standby_ep_group = get_standby_ep_group()
|
||||
|
||||
Reference in New Issue
Block a user