software-engineering/spark-vllm-docker
2
0
Fork 0
Code Issues Actions Packages Projects Releases Wiki Activity

Updated README to reflect --master-port parameter

Browse Source
This commit is contained in:
Eugene Rakhmatulin
2026-03-18 21:23:28 -07:00
parent 7e4150feed
commit 15d295887c
1 changed files with 18 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
README.md
View File

@@ -149,6 +149,22 @@ Don't do it every time you rebuild, because it will slow down compilation times.
For periodic maintenance, I recommend using a filter: `docker builder prune --filter until=72h` For periodic maintenance, I recommend using a filter: `docker builder prune --filter until=72h`
### 2026-03-18
#### `--master-port` / `--head-port` Parameter
Added `--master-port` (synonym: `--head-port`) to both `launch-cluster.sh` and `run-recipe.sh` to configure the port used for cluster coordination:
- In **Ray mode**: sets the Ray head node port (previously hardcoded to 6379)
- In **No-Ray mode**: sets the PyTorch distributed `--master-port` passed to vLLM
Default is `29501`.
```bash
./launch-cluster.sh --master-port 29501 --no-ray exec vllm serve ...
./run-recipe.sh qwen3.5-122b-fp8 --no-ray --master-port 29501
```
### 2026-03-17 ### 2026-03-17
#### EXPERIMENTAL Intel/Qwen3.5-397B-A17B-int4-AutoRound Recipe #### EXPERIMENTAL Intel/Qwen3.5-397B-A17B-int4-AutoRound Recipe
@@ -200,7 +216,7 @@ Added `--no-ray` flag to `launch-cluster.sh` to run multi-node vLLM clusters wit
``` ```
The following `launch-cluster.sh` flags are now also passed through from `run-recipe.sh`: The following `launch-cluster.sh` flags are now also passed through from `run-recipe.sh`:
`--name`, `--eth-if`, `--ib-if`, `-j`, `--no-cache-dirs`, `--non-privileged`, `--mem-limit-gb`, `--mem-swap-limit-gb`, `--pids-limit`, `--shm-size-gb`. `--master-port`, `--name`, `--eth-if`, `--ib-if`, `-j`, `--no-cache-dirs`, `--non-privileged`, `--mem-limit-gb`, `--mem-swap-limit-gb`, `--pids-limit`, `--shm-size-gb`.
#### Nemotron-3-Nano-NVFP4 Switched to Marlin Backend #### Nemotron-3-Nano-NVFP4 Switched to Marlin Backend
@@ -976,6 +992,7 @@ You can override the auto-detected values if needed:
| `--check-config` | Check configuration and auto-detection without launching. | | `--check-config` | Check configuration and auto-detection without launching. |
| `--solo` | Solo mode: skip autodetection, launch only on current node, do not launch Ray cluster | | `--solo` | Solo mode: skip autodetection, launch only on current node, do not launch Ray cluster |
| `--no-ray` | No-Ray mode: run multi-node vLLM without Ray (uses PyTorch distributed backend). | | `--no-ray` | No-Ray mode: run multi-node vLLM without Ray (uses PyTorch distributed backend). |
| `--master-port` / `--head-port` | Port for cluster coordination: Ray head port or PyTorch distributed master port (default: 29501). |
| `--no-cache-dirs` | Do not mount default cache directories (~/.cache/vllm, ~/.cache/flashinfer, ~/.triton). | | `--no-cache-dirs` | Do not mount default cache directories (~/.cache/vllm, ~/.cache/flashinfer, ~/.triton). |
| `--launch-script` | Path to bash script to execute in the container (from examples/ directory or absolute path). If launch script is specified, action should be omitted. | | `--launch-script` | Path to bash script to execute in the container (from examples/ directory or absolute path). If launch script is specified, action should be omitted. |
| `-d` | Run in daemon mode (detached). | | `-d` | Run in daemon mode (detached). |