[PD] Publish prefill DP rank whenever forced lookup is enabled (#41786)
This commit is contained in:
@@ -1799,27 +1799,27 @@ class CommonKVSender(BaseKVSender):
|
||||
|
||||
self.kv_mgr.update_status(self.bootstrap_room, KVPoll.Bootstrapping)
|
||||
if get_parallel().num_dp_ranks > 1 and not req_has_disagg_prefill_dp_rank:
|
||||
if get_parallel().load_balance_method != "follow_bootstrap_room":
|
||||
if (
|
||||
get_parallel().load_balance_method != "follow_bootstrap_room"
|
||||
or envs.SGLANG_DISAGGREGATION_FORCE_QUERY_PREFILL_DP_RANK.get()
|
||||
):
|
||||
self._register_prefill_dp_rank()
|
||||
elif (
|
||||
self.kv_mgr.attn_dp_rank
|
||||
!= self.bootstrap_room % get_parallel().num_dp_ranks
|
||||
):
|
||||
# follow_bootstrap_room was overridden by external routed_dp_rank
|
||||
if envs.SGLANG_DISAGGREGATION_FORCE_QUERY_PREFILL_DP_RANK.get():
|
||||
self._register_prefill_dp_rank()
|
||||
else:
|
||||
self.kv_mgr.record_failure(
|
||||
self.bootstrap_room,
|
||||
f"follow_bootstrap_room conflict: dispatched to dp_rank "
|
||||
f"{self.kv_mgr.attn_dp_rank} but bootstrap_room "
|
||||
f"{self.bootstrap_room} implies dp_rank "
|
||||
f"{self.bootstrap_room % get_parallel().num_dp_ranks}. "
|
||||
f"Set SGLANG_DISAGGREGATION_FORCE_QUERY_PREFILL_DP_RANK=1 "
|
||||
f"to allow mixed routing.",
|
||||
)
|
||||
self.kv_mgr.update_status(self.bootstrap_room, KVPoll.Failed)
|
||||
return
|
||||
self.kv_mgr.record_failure(
|
||||
self.bootstrap_room,
|
||||
f"follow_bootstrap_room conflict: dispatched to dp_rank "
|
||||
f"{self.kv_mgr.attn_dp_rank} but bootstrap_room "
|
||||
f"{self.bootstrap_room} implies dp_rank "
|
||||
f"{self.bootstrap_room % get_parallel().num_dp_ranks}. "
|
||||
f"Set SGLANG_DISAGGREGATION_FORCE_QUERY_PREFILL_DP_RANK=1 "
|
||||
f"to allow mixed routing.",
|
||||
)
|
||||
self.kv_mgr.update_status(self.bootstrap_room, KVPoll.Failed)
|
||||
return
|
||||
|
||||
def _register_prefill_dp_rank(self):
|
||||
"""Register this request's prefill dp_rank to the bootstrap server."""
|
||||
|
||||
Reference in New Issue
Block a user