[Doc] Clearer corresponding relationship between configurations for multi-node guides (#3441)

Optimize multi-node guide: more clearer corresponding relationship
between configuration items and nodes

### What this PR does / why we need it?
Some issues caused by misunderstandings due to unclear guidance content,
for example: #3367

### Does this PR introduce _any_ user-facing change?
NA
### How was this patch tested?
NA

- vLLM version: v0.11.0rc3
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0

Signed-off-by: leo-pony <nengjunma@outlook.com>
This commit is contained in:
leo-pony
2025-10-16 08:54:03 +08:00
committed by GitHub
parent aa6154703a
commit ff91904ee2
4 changed files with 32 additions and 8 deletions

View File

@ -192,7 +192,7 @@ Before launch the inference server, ensure the following environment variables a
#!/bin/sh
# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxxx"
local_ip="xxxx"
@ -230,9 +230,14 @@ vllm serve /root/.cache/Modelers_Park/DeepSeek-V3.2-Exp \
```shell
#!/bin/sh
# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxx"
local_ip="xxx"
# The value of node0_ip must be consistent with the value of local_ip set in node0 (master node)
node0_ip="xxxx"
export VLLM_USE_MODELSCOPE=True
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
@ -249,7 +254,7 @@ vllm serve /root/.cache/Modelers_Park/DeepSeek-V3.2-Exp \
--data-parallel-size 2 \
--data-parallel-size-local 1 \
--data-parallel-start-rank 1 \
--data-parallel-address <node0_ip> \
--data-parallel-address $node0_ip \
--data-parallel-rpc-port 13389 \
--tensor-parallel-size 16 \
--seed 1024 \
@ -300,7 +305,7 @@ Run the following scripts on two nodes respectively
#!/bin/sh
# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxxx"
local_ip="xxxx"
@ -341,9 +346,14 @@ vllm serve vllm-ascend/DeepSeek-V3.2-Exp-W8A8 \
```shell
#!/bin/sh
# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxx"
local_ip="xxx"
# The value of node0_ip must be consistent with the value of local_ip set in node0 (master node)
node0_ip="xxxx"
export VLLM_USE_MODELSCOPE=True
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
@ -362,7 +372,7 @@ vllm serve vllm-ascend/DeepSeek-V3.2-Exp-W8A8 \
--data-parallel-size 2 \
--data-parallel-size-local 1 \
--data-parallel-start-rank 1 \
--data-parallel-address <node0_ip> \
--data-parallel-address $node0_ip \
--data-parallel-rpc-port 13389 \
--tensor-parallel-size 8 \
--seed 1024 \

View File

@ -103,7 +103,7 @@ Before launch the inference server, ensure the following environment variables a
#!/bin/sh
# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxxx"
local_ip="xxxx"
@ -144,9 +144,14 @@ vllm serve vllm-ascend/DeepSeek-V3.1-W8A8 \
```shell
#!/bin/sh
# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxx"
local_ip="xxx"
# The value of node0_ip must be consistent with the value of local_ip set in node0 (master node)
node0_ip="xxxx"
export VLLM_USE_MODELSCOPE=True
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
@ -164,7 +169,7 @@ vllm serve vllm-ascend/DeepSeek-V3.1-W8A8 \
--data-parallel-size 4 \
--data-parallel-size-local 2 \
--data-parallel-start-rank 2 \
--data-parallel-address { node0 ip } \
--data-parallel-address $node0_ip \
--data-parallel-rpc-port 13389 \
--tensor-parallel-size 4 \
--seed 1024 \

View File

@ -59,7 +59,7 @@ Before launch the inference server, ensure the following environment variables a
#!/bin/sh
# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxxx"
local_ip="xxxx"
@ -101,9 +101,14 @@ vllm serve /home/cache/weights/Kimi-K2-Instruct-W8A8 \
```shell
#!/bin/sh
# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxxx"
local_ip="xxxx"
# The value of node0_ip must be consistent with the value of local_ip set in node0 (master node)
node0_ip="xxxx"
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name

View File

@ -59,7 +59,7 @@ node0
```shell
#!/bin/sh
# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxxx"
local_ip="xxxx"
@ -97,8 +97,12 @@ node1
```shell
#!/bin/sh
# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip of the current node
nic_name="xxxx"
local_ip="xxxx"
# The value of node0_ip must be consistent with the value of local_ip set in node0 (master node)
node0_ip="xxxx"
export HCCL_IF_IP=$local_ip