EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

feat(hicache-3fs): 3FS-SGLang Hierarchical Cache Deployment Guide​ (#9213)

Browse Source
This commit is contained in:
hzh0425
2025-08-15 14:54:31 +08:00
committed by GitHub
parent 9d54c6e6dd
commit f7dd651dbd
2 changed files with 19 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
python/sglang/srt/mem_cache/storage/hf3fs/docs/README.md
View File

@@ -1,19 +1,26 @@
# HF3FS as L3 KV Cache
# Using HF3FS as L3 Global KV Cache
This document describes how to use deepseek-hf3fs as the L3 KV cache for SGLang.
This document provides step-by-step instructions for setting up a k8s + 3FS + SGLang runtime environment from scratch, describing how to utilize deepseek-hf3fs as the L3 KV cache for SGLang.
The process consists of five main steps:
## Step1: Install deepseek-3fs by 3fs-Operator (Coming Soon)
## Step 1: Install deepseek-3fs via 3fs-Operator
Refer to the [3fs-operator documentation](https://github.com/aliyun/kvc-3fs-operator/blob/main/README_en.md) to deploy 3FS components in your Kubernetes environment using the Operator with one-click deployment.
## Step2: Setup usrbio client
## Step 2: Launch SGLang Pod
Start your SGLang Pod while specifying 3FS-related labels in the YAML configuration. Follow the [fuse-client-creation guide](https://github.com/aliyun/kvc-3fs-operator/blob/main/README_en.md#fuse-client-creation).
Please follow the document [setup_usrbio_client.md](setup_usrbio_client.md) to setup usrbio client.
## Step 3: Configure Usrbio Client in SGLang Pod
The Usrbio client is required for accessing 3FS. Install it in your SGLang Pod using either method below:
## Step3: Deployment
**Alternative 1 (Recommend):** Build from source (refer to [setup_usrbio_client.md](setup_usrbio_client.md))
### Single node deployment
**Alternative 2:** Run `pip3 install hf3fs-py-usrbio` (Follow https://pypi.org/project/hf3fs-py-usrbio/#files)
## Step 4: Deploy Model Serving
### Single Node Deployment
```bash
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib/python3.10/dist-packages
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib/python3.12/dist-packages
python3 -m sglang.launch_server \
--model-path /code/models/Qwen3-32B/ \
--host 0.0.0.0 --port 10000 \
@@ -24,6 +31,5 @@ python3 -m sglang.launch_server \
--hicache-storage-backend hf3fs
```
### Multi nodes deployment to share KV cache
Please follow the document [deploy_sglang_3fs_multinode.md](deploy_sglang_3fs_multinode.md) to deploy SGLang with 3FS on multiple nodes to share KV cache.
### Multi-Node Deployment (Shared KV Cache)
Follow the [deploy_sglang_3fs_multinode.md](deploy_sglang_3fs_multinode.md) guide to deploy SGLang with 3FS across multiple nodes for shared KV caching.

4
python/sglang/srt/mem_cache/storage/hf3fs/docs/deploy_sglang_3fs_multinode.md
View File

@@ -20,7 +20,7 @@ vim /sgl-workspace/sglang/benchmark/hf3fs/hf3fs_config.json
## node1
```bash
export SGLANG_HICACHE_HF3FS_CONFIG_PATH=/sgl-workspace/sglang/benchmark/hf3fs/hf3fs_config.json
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib/python3.10/dist-packages
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib/python3.12/dist-packages
rm -rf instance1.out && \
nohup python3 -m sglang.launch_server \
--model-path /code/models/Qwen3-32B/ \
@@ -35,7 +35,7 @@ nohup python3 -m sglang.launch_server \
## node2
```bash
export SGLANG_HICACHE_HF3FS_CONFIG_PATH=/sgl-workspace/sglang/benchmark/hf3fs/hf3fs_config.json
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib/python3.10/dist-packages
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib/python3.12/dist-packages
rm -rf instance2.out && \
nohup python3 -m sglang.launch_server \
--model-path /code/models/Qwen3-32B/ \