[Weight Cache] Enhance test and support EPLB

python/sglang/srt/server_args.py

@@ -8216,11 +8216,6 @@ def _handle_load_format(self):
                 "(--weight-cache-mode off) for this configuration."
             )
 
-        if self.weight_cache_mode != "off" and self.enable_eplb:
-            raise ValueError(
-                "--weight-cache-mode is not supported together with --enable-eplb."
-            )
-
     def _is_mistral_native_format(self) -> bool:
         """True iff the checkpoint requires load_format=mistral.
 

python/sglang/srt/weight_cache/daemon.py

@@ -314,6 +314,7 @@ def load(self):
         # The initialized groups are the authority for rank identity. This
         # avoids maintaining a second copy of the model-parallel hierarchy.
         self._init_distributed(server_args, model_config)
+        self._initialize_eplb_expert_location_metadata(model_config)
         moe_dp_rank = get_parallel().moe_dp_rank
         moe_ep_rank = get_parallel().moe_ep_rank
         self.config = CacheConfig(
@@ -455,6 +456,24 @@ def _export_state(self):
             f"metadata size ~{total_bytes / 1024 / 1024:.1f} MB"
         )
 
+    def _initialize_eplb_expert_location_metadata(self, model_config) -> None:
+        """Build the same initial physical expert layout as the engine."""
+        if not self.server_args.enable_eplb:
+            return
+
+        from sglang.srt.eplb.expert_location import (
+            compute_initial_expert_location_metadata,
+            set_global_expert_location_metadata,
+        )
+
+        set_global_expert_location_metadata(
+            compute_initial_expert_location_metadata(
+                server_args=self.server_args,
+                model_config=model_config,
+                moe_ep_rank=get_parallel().moe_ep_rank,
+            )
+        )
+
     def serve(self):
         """Block and serve IPC handles over Unix socket."""
         # Do NOT unlink an existing socket here: stale-file cleanup is the launch

test/registered/model_loading/test_weight_cache_daemon.py

@@ -10,6 +10,7 @@
 from sglang.srt.utils import kill_process_tree
 from sglang.test.ci.ci_register import register_cuda_ci
 from sglang.test.test_utils import (
+    DEFAULT_TARGET_MODEL_EAGLE_DP_ATTN,
     DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
     DEFAULT_URL_FOR_TEST,
     CustomTestCase,
@@ -26,7 +27,7 @@
 # IPC handoff is exercised on every PR. Since the CI runner executes the whole
 # file per suite, TestWeightCacheDaemonTP2 self-skips when fewer than 2 GPUs are
 # visible (i.e. on the 1-gpu runner).
-register_cuda_ci(est_time=100, stage="extra-a", runner_config="2-gpu-large")
+register_cuda_ci(est_time=280, stage="extra-a", runner_config="2-gpu-large")
 register_cuda_ci(est_time=45, stage="base-b", runner_config="1-gpu-small")
 
 # Capture the client server's logs so test_loaded_via_ipc can assert the IPC
@@ -50,9 +51,13 @@
 class TestWeightCacheDaemonTP2(CustomTestCase):
     """E2E test: start weight cache daemons, then launch server in client mode with TP2."""
 
+    model_override = None
+    daemon_args = []
+    server_args = []
+
     @classmethod
     def setUpClass(cls):
-        cls.model = DEFAULT_MODEL
+        cls.model = cls.model_override or DEFAULT_MODEL
         cls.base_url = DEFAULT_URL_FOR_TEST
         cls.tp_size = 2
 
@@ -74,6 +79,7 @@ def setUpClass(cls):
                 cls.model,
                 "--tp-size",
                 str(cls.tp_size),
+                *cls.daemon_args,
             ]
         )
 
@@ -107,6 +113,7 @@ def setUpClass(cls):
                 str(cls.tp_size),
                 "--weight-cache-mode",
                 "client",
+                *cls.server_args,
             ],
             return_stdout_stderr=(cls.stdout, cls.stderr),
         )
@@ -334,5 +341,39 @@ def test_loaded_via_ipc(self):
         )
 
 
+class TestWeightCacheDaemonQwen3MoeDP(TestWeightCacheDaemonTP2):
+    """Qwen3 with static attention DP through the existing IPC fixture."""
+
+    daemon_args = [
+        "--dp",
+        "2",
+        "--ep-size",
+        "1",
+        "--enable-dp-attention",
+        "--enable-dp-lm-head",
+        "--random-seed",
+        "42",
+    ]
+    server_args = daemon_args[:]
+
+
+class TestWeightCacheDaemonQwen3MoeEP(TestWeightCacheDaemonTP2):
+    """Qwen3 MoE with static expert parallelism through the existing IPC fixture."""
+
+    model_override = DEFAULT_TARGET_MODEL_EAGLE_DP_ATTN
+    daemon_args = [
+        "--dp",
+        "1",
+        "--ep-size",
+        "2",
+        "--enable-eplb",
+        "--ep-num-redundant-experts",
+        "2",
+        "--random-seed",
+        "42",
+    ]
+    server_args = daemon_args[:]
+
+
 if __name__ == "__main__":
     unittest.main()

test/registered/unit/model_loader/test_weight_cache_protocol.py

@@ -269,8 +269,9 @@ def test_spawn_forwards_complete_server_args_without_projection(self):
         from sglang.srt.weight_cache import daemon
 
         # The spawn helper receives Engine's already-resolved ServerArgs. A
-        # minimal namespace keeps this projection test CPU-only and
-        # model-independent; importantly, no EPLB configuration is involved.
+        # minimal namespace keeps this forwarding test CPU-only and
+        # model-independent while covering the static DP/EP and EPLB fields
+        # consumed by the daemon.
         server_args = SimpleNamespace(
             model_path="/models/demo",
             tp_size=8,
@@ -284,6 +285,8 @@ def test_spawn_forwards_complete_server_args_without_projection(self):
             moe_dense_tp_size=1,
             moe_a2a_backend="mooncake",
             deepep_mode="low_latency",
+            enable_eplb=True,
+            ep_num_redundant_experts=72,
             load_format="safetensors",
             dtype="bfloat16",
             quantization="fp8",

test/registered/unit/server_args/test_server_args.py

@@ -43,6 +43,17 @@
 
 
 class TestPrepareServerArgs(CustomTestCase):
+    def test_weight_cache_daemon_allows_static_eplb(self):
+        args = ServerArgs(
+            model_path="dummy",
+            weight_cache_mode="daemon",
+            enable_eplb=True,
+        )
+
+        # This validation runs before model construction and should allow the
+        # daemon to build the same static EPLB layout as the engine.
+        args._handle_load_format()
+
     def test_enable_w4a4_mxfp4_megamoe_sets_deepgemm_env(self):
         deepgemm_env = {
             "DG_USE_FP4_ACTS": "0",