From 7be66cec75904a040f5b753f1f4bf679531e51a0 Mon Sep 17 00:00:00 2001
From: wangyu <53896905+yenuo26@users.noreply.github.com>
Date: Fri, 20 Mar 2026 11:33:48 +0800
Subject: [PATCH] [Test] Add the always_check_nodes parameter to the
 _wait_for_multiple_servers function in conftest.py for the EPD test case.
 (#7410)

### What this PR does / why we need it?
This PR add the always_check_nodes parameter to the
_wait_for_multiple_servers function in conftest.py for the EPD test
case.

### Does this PR introduce _any_ user-facing change?
no

### How was this patch tested?
1.by running the test
`pytest -sv test_disaggregated_encoder.py`

2.by running ci

- vLLM version: v0.17.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/4497431df654e46fb1fb5e64bf8611e762ae5d87

---------

Signed-off-by: yenuo26 <410167048@qq.com>
---
 tests/e2e/conftest.py | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/tests/e2e/conftest.py b/tests/e2e/conftest.py
index d124b9ce..b1c47086 100644
--- a/tests/e2e/conftest.py
+++ b/tests/e2e/conftest.py
@@ -349,7 +349,9 @@ class RemoteOpenAIServer:
         # Then wait for all api_server nodes
         self._wait_for_multiple_servers(targets=targets, timeout=timeout)
 
-    def _wait_for_multiple_servers(self, targets, timeout: float, log_interval: float = 30.0):
+    def _wait_for_multiple_servers(
+        self, targets, timeout: float, log_interval: float = 30.0, always_check_nodes: bool = False
+    ):
         """
         targets: List[(node_ip, url)]
         log_interval
@@ -367,14 +369,14 @@ class RemoteOpenAIServer:
             should_log = (now - last_log_time) >= log_interval
 
             for node_ip, url in targets:
-                if ready[node_ip]:
+                if ready[node_ip] and not always_check_nodes:
                     continue
 
                 try:
                     resp = client.get(url)
                     if resp.status_code == 200:
                         ready[node_ip] = True
-                        logger.info(f"[READY] Node {node_ip} is ready.")
+                        logger.info(f"[READY] Node {node_ip}: {url} is ready.")
                 except RequestException:
                     all_ready = False
                     if should_log:
@@ -498,7 +500,9 @@ class RemoteEPDServer(RemoteOpenAIServer):
             self._proc_list.append(proc)
 
         timeout_value = float(max_wait_seconds) if max_wait_seconds is not None else 2800.0
-        super()._wait_for_multiple_servers([(self.host, url) for url in self.health_url_list], timeout=timeout_value)
+        super()._wait_for_multiple_servers(
+            [(self.host, url) for url in self.health_url_list], timeout=timeout_value, always_check_nodes=True
+        )
 
     def _poll(self) -> int | None:
         return None