[NIXL] Increase default KV block eviction timeout on P (vllm-project#25897)

Signed-off-by: NickLucche <[email protected]>

Author: NickLucche

docs/features/nixl_connector_usage.md

@@ -84,7 +84,7 @@ python tests/v1/kv_connector/nixl_integration/toy_proxy_server.py \
     - Connection info is passed via KVTransferParams from prefiller to decoder for handshake
 
 - `VLLM_NIXL_ABORT_REQUEST_TIMEOUT`: Timeout (in seconds) for automatically releasing the prefiller’s KV cache for a particular request. (Optional)
-    - Default: 120
+    - Default: 480
     - If a request is aborted and the decoder has not yet read the KV-cache blocks through the nixl channel, the prefill instance will release its KV-cache blocks after this timeout to avoid holding them indefinitely.
 
 ## Multi-Instance Setup

vllm/envs.py

@@ -174,7 +174,7 @@
                                                  "NONE"] = "NONE"
     VLLM_ROCM_QUICK_REDUCE_CAST_BF16_TO_FP16: bool = True
     VLLM_ROCM_QUICK_REDUCE_MAX_SIZE_BYTES_MB: Optional[int] = None
-    VLLM_NIXL_ABORT_REQUEST_TIMEOUT: int = 120
+    VLLM_NIXL_ABORT_REQUEST_TIMEOUT: int = 480
     VLLM_USE_CUDNN_PREFILL: bool = False
     VLLM_ENABLE_CUDAGRAPH_GC: bool = False
     VLLM_LOOPBACK_IP: str = ""
@@ -1330,7 +1330,7 @@ def get_vllm_port() -> Optional[int]:
     # consumer. This is only applicable when using NixlConnector in a
     # disaggregated decode-prefill setup.
     "VLLM_NIXL_ABORT_REQUEST_TIMEOUT":
-    lambda: int(os.getenv("VLLM_NIXL_ABORT_REQUEST_TIMEOUT", "120")),
+    lambda: int(os.getenv("VLLM_NIXL_ABORT_REQUEST_TIMEOUT", "480")),
 
     # Controls whether or not to use cudnn prefill
     "VLLM_USE_CUDNN_PREFILL":