pytorch/torch/distributed/elastic/rendezvous/registry.py
Sathyanarayanan Saravanamuthu 136b19b062 Adding entry-point based support for out-of-tree rendezvous plugins (#132633)
Fixes #127519

Currently in torchrun rendezvous, there are only two rendezvous backends supported out of the box: `C10d` and `Etcd`. The changes in this PR enables the distributed elastic users to bring their out-of-tree rendezvous backend implementations as Python packages.

#### AUTHORING NEW PLUGIN
Any new plugin will be a python package exposing entry-points. For example, the structure of redis plugin is as follows:

```
plugin_root
|_ pyproject.toml
|_ src
   |_ redis
      |_ __init__.py
      |_ redis_store.py
      |_ redis_backend.py
```

The contents of the `pyproject.toml` should indicate that this is exposes a torchrun entry-point by mentioning the group name `torchrun.plugins`. The `pyproject.toml` for redis plugin would be as follows:

```
[project]
name = "redis"
version = "0.0.1"

[project.entry-points.'torchrun.plugins']
redis = 'redis'
```

The `src/redis/__init__.py` file would contain functions that return the plugin name and plugin handler. The contents of `__init__.py` for redis would be as follows:

```
def getPluginHandler():
    def _create_redis_handler(params: RendezvousParameters):
        from redis_rendezvous_backend import create_backend
        backend, store = create_backend(params)
        return create_handler(store, backend, params)
    return _create_redis_handler
```

The files `redis_store` and `redis_backend` contain the implementation of [Store](41189b0da4/torch/_C/_distributed_c10d.pyi (L171)) and [RendezvousBackend](e782918b8e/torch/distributed/elastic/rendezvous/dynamic_rendezvous.py (L61)) respectively.

#### USER EXPERIENCE
Before using the plugin for the first time, the user has to install the plugin packages. For example, the published packages can be installed using `pip3 install <plugin-name>` and the plugin is in local file systemcan be installed using `pip3 install -e <plugin-location>`.

Once installed, the new backend can be used in torchrun as follows:

```
torchrun --rdzv-backend=redis --rdzv-endpoint=redis-container:6379 --nnodes=3 --nproc-per-node=1 --max-restarts=3 --rdzv-id=1 test.py
```

Pull Request resolved: https://github.com/pytorch/pytorch/pull/132633
Approved by: https://github.com/wconstab
2024-08-27 07:09:41 +00:00

97 lines
2.9 KiB
Python

# Copyright (c) Facebook, Inc. and its affiliates.
# All rights reserved.
#
# This source code is licensed under the BSD-style license found in the
# LICENSE file in the root directory of this source tree.
import logging
import sys
from .api import (
rendezvous_handler_registry as handler_registry,
RendezvousHandler,
RendezvousParameters,
)
from .dynamic_rendezvous import create_handler
if sys.version_info < (3, 10):
from importlib_metadata import entry_points
else:
from importlib.metadata import entry_points
log = logging.getLogger(__name__)
__all__ = ["get_rendezvous_handler"]
def _create_static_handler(params: RendezvousParameters) -> RendezvousHandler:
from . import static_tcp_rendezvous
return static_tcp_rendezvous.create_rdzv_handler(params)
def _create_etcd_handler(params: RendezvousParameters) -> RendezvousHandler:
from . import etcd_rendezvous
return etcd_rendezvous.create_rdzv_handler(params)
def _create_etcd_v2_handler(params: RendezvousParameters) -> RendezvousHandler:
from .etcd_rendezvous_backend import create_backend
backend, store = create_backend(params)
return create_handler(store, backend, params)
def _create_c10d_handler(params: RendezvousParameters) -> RendezvousHandler:
from .c10d_rendezvous_backend import create_backend
backend, store = create_backend(params)
return create_handler(store, backend, params)
def _register_default_handlers() -> None:
handler_registry.register("etcd", _create_etcd_handler)
handler_registry.register("etcd-v2", _create_etcd_v2_handler)
handler_registry.register("c10d", _create_c10d_handler)
handler_registry.register("static", _create_static_handler)
def _register_out_of_tree_handlers() -> None:
discovered_handler_generators = entry_points(group="torchrun.handlers")
for handler_generator in discovered_handler_generators:
try:
get_handler = discovered_handler_generators[handler_generator.name].load()
handler_registry.register(handler_generator.name, get_handler())
except Exception:
log.warning(
"Exception while registering out of tree plugin %s: ",
handler_generator.name,
exc_info=True,
)
def get_rendezvous_handler(params: RendezvousParameters) -> RendezvousHandler:
"""
Obtain a reference to a :py:class`RendezvousHandler`.
Custom rendezvous handlers can be registered by
::
from torch.distributed.elastic.rendezvous import rendezvous_handler_registry
from torch.distributed.elastic.rendezvous.registry import get_rendezvous_handler
def create_my_rdzv(params: RendezvousParameters):
return MyCustomRdzv(params)
rendezvous_handler_registry.register("my_rdzv_backend_name", create_my_rdzv)
my_rdzv_handler = get_rendezvous_handler("my_rdzv_backend_name", RendezvousParameters)
"""
return handler_registry.create_handler(params)