disagg.sh 631 Bytes
Newer Older
Alec's avatar
Alec committed
1
2
3
4
5
6
7
#!/bin/bash
# SPDX-FileCopyrightText: Copyright (c) 2025 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: Apache-2.0
set -e
trap 'echo Cleaning up...; kill 0' EXIT

# run ingress
8
python -m dynamo.frontend --http-port=8000 &
Alec's avatar
Alec committed
9

10
# --enforce-eager is added for quick deployment. for production use, need to remove this flag
Alec's avatar
Alec committed
11
CUDA_VISIBLE_DEVICES=0 python3 -m dynamo.vllm --model Qwen/Qwen3-0.6B --enforce-eager &
Alec's avatar
Alec committed
12

13
14
DYN_VLLM_KV_EVENT_PORT=20081 \
VLLM_NIXL_SIDE_CHANNEL_PORT=20097 \
Alec's avatar
Alec committed
15
CUDA_VISIBLE_DEVICES=1 python3 -m dynamo.vllm \
Alec's avatar
Alec committed
16
17
18
    --model Qwen/Qwen3-0.6B \
    --enforce-eager \
    --is-prefill-worker