# tensorplay.distributed.elastic.rendezvous API

Source: https://www.tensorplay.cn/docs/api/tensorplay.distributed.elastic.rendezvous.html

## Functions 1

[#](#api-tensorplay.distributed.elastic.rendezvous.create_handler)

### create_handler

function[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.create_handler.html)

```python
tensorplay.distributed.elastic.rendezvous.create_handler(params: RendezvousParameters) → RendezvousHandler
```

Create a handler from the process-wide registry.

## Classes 9

[#](#api-tensorplay.distributed.elastic.rendezvous.DynamicRendezvousHandler)

### DynamicRendezvousHandler

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.DynamicRendezvousHandler.html)

```python
class tensorplay.distributed.elastic.rendezvous.DynamicRendezvousHandler(backend: RendezvousBackend | _NodeDesc | None = None, settings: RendezvousSettings | None = None, local_addr: str | None = None, node_rank: int = 0, run_id: str = '', store: Store | None = None, backend_name: str = 'core', state_holder: _RendezvousStateHolder | None = None, node: _NodeDesc | None = None)
```

Coordinate membership, ranks, heartbeats, and round transitions.

```python
pre_shutdown() → None
```

Optional hook invoked before shutdown() during agent teardown.

[#](#api-tensorplay.distributed.elastic.rendezvous.P10dRendezvousBackend)

### P10dRendezvousBackend

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.P10dRendezvousBackend.html)

```python
class tensorplay.distributed.elastic.rendezvous.P10dRendezvousBackend(store: Store, run_id: str)
```

Rendezvous state backend backed by a TP key/value Store.

[#](#api-tensorplay.distributed.elastic.rendezvous.RendezvousHandler)

### RendezvousHandler

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.RendezvousHandler.html)

```python
class tensorplay.distributed.elastic.rendezvous.RendezvousHandler
```

Algorithmic interface of one rendezvous backend.

```python
abstractmethod get_backend() → str
```

Backend name (e.g. static or core).

```python
abstractmethod get_run_id() → str
```

Run identifier of this rendezvous.

```python
abstractmethod is_closed() → bool
```

Whether the rendezvous has been marked closed.

```python
abstractmethod next_rendezvous() → RendezvousInfo
```

Join or re-join the rendezvous; blocks until this round completes.

```python
abstractmethod num_nodes_waiting() → int
```

Number of nodes queued for the next round (pending scale-up).

```python
pre_shutdown() → None
```

Optional hook invoked before [shutdown()](#tensorplay.distributed.elastic.rendezvous.RendezvousHandler.shutdown) during agent teardown.

```python
abstractmethod set_closed() → None
```

Mark the rendezvous closed; new joiners are rejected.

```python
abstractmethod shutdown() → bool
```

Release rendezvous resources; returns True when the backend shut down.

```python
property use_agent_store: bool
```

Whether workers can reuse the agent-held store as their bootstrap store.

[#](#api-tensorplay.distributed.elastic.rendezvous.RendezvousInfo)

### RendezvousInfo

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.RendezvousInfo.html)

```python
class tensorplay.distributed.elastic.rendezvous.RendezvousInfo(store: Store, rank: int, world_size: int, bootstrap_store_info: RendezvousStoreInfo | None = None, participants: list | None = None, wait_list: list | None = None)
```

Outcome of a successful rendezvous for one agent.

```python
property bootstrap_store_info: RendezvousStoreInfo | None
```

Connection info for the worker-facing bootstrap store.

```python
property participants: list
```

Nodes participating in this round.

```python
property rank: int
```

This agent’s rank within the rendezvous.

```python
property store: Store
```

Shared store for this rendezvous round.

```python
property wait_list: list
```

Nodes waiting for a future round.

```python
property world_size: int
```

Number of participating agents.

[#](#api-tensorplay.distributed.elastic.rendezvous.RendezvousParameters)

### RendezvousParameters

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.RendezvousParameters.html)

```python
class tensorplay.distributed.elastic.rendezvous.RendezvousParameters(backend: str, endpoint: str, run_id: str, *args, min_nodes: int | None = None, max_nodes: int | None = None, local_addr: str | None = None, node_rank: int = 0, local_world_size: int = 1, config: dict | None = None, **kwargs)
```

Parameters describing one rendezvous request.

endpoint is host:port of the rendezvous store, local_addr the
address of this node (autodetected when empty), and config carries
backend-specific options accessible through [get()](#tensorplay.distributed.elastic.rendezvous.RendezvousParameters.get).

```python
get(key: str, default=None)
```

Return backend parameter key or default.

```python
get_as_bool(key: str, default: bool | None = None) → bool | None
```

Return key coerced to bool (accepts 1/0/true/false).

```python
get_as_int(key: str, default: int | None = None) → int | None
```

Return key coerced to int.

```python
property timeout: RendezvousTimeout
```

Timeouts assembled from join/last_call/close/heartbeat params (seconds).

[#](#api-tensorplay.distributed.elastic.rendezvous.RendezvousSettings)

### RendezvousSettings

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.RendezvousSettings.html)

```python
class tensorplay.distributed.elastic.rendezvous.RendezvousSettings(*args: Any, join: timedelta | None = None, last_call: timedelta | None = None, close: timedelta | None = None, heartbeat: timedelta | None = None, min_nodes: int = 1, max_nodes: int = 1, run_id: str = '', timeout: RendezvousTimeout | None = None, keep_alive_interval: timedelta | None = None, keep_alive_max_attempt: int = 3)
```

Configuration shared by the rendezvous state machine.

[#](#api-tensorplay.distributed.elastic.rendezvous.RendezvousStoreInfo)

### RendezvousStoreInfo

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.RendezvousStoreInfo.html)

```python
class tensorplay.distributed.elastic.rendezvous.RendezvousStoreInfo(master_addr: str, port: int)
```

Connection information for the bootstrap store handed to workers.

```python
property addr: str
```

Alias of master_addr.

```python
classmethod build(rank: int, store: Store, local_addr: str | None = None, server_port: int | None = None)
```

Derive store info for rank from the rendezvous store.

Rank 0 owns the address/port pair: it uses the store’s own listening
endpoint when available and otherwise publishes one through the
store so every other rank can read it.

[#](#api-tensorplay.distributed.elastic.rendezvous.StaticTCPRendezvous)

### StaticTCPRendezvous

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.StaticTCPRendezvous.html)

```python
class tensorplay.distributed.elastic.rendezvous.StaticTCPRendezvous(params: RendezvousParameters | str, master_port: int | None = None, rank: int = 0, world_size: int = 1, run_id: str = '', timeout: int = 600)
```

One-shot rendezvous over a static TCPStore endpoint.

```python
pre_shutdown() → None
```

Optional hook invoked before shutdown() during agent teardown.

[#](#api-tensorplay.distributed.elastic.rendezvous.TpRendezvousBackend)

### TpRendezvousBackend

class[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.TpRendezvousBackend.html)

```python
class tensorplay.distributed.elastic.rendezvous.TpRendezvousBackend(store: Store, run_id: str)
```

Store rendezvous state behind compare-and-set operations.

## Exceptions 7

[#](#api-tensorplay.distributed.elastic.rendezvous.RendezvousClosedError)

### RendezvousClosedError

exception[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.RendezvousClosedError.html)

```python
exception tensorplay.distributed.elastic.rendezvous.RendezvousClosedError
```

The rendezvous has been closed and accepts no more participants.

[#](#api-tensorplay.distributed.elastic.rendezvous.RendezvousConnectionError)

### RendezvousConnectionError

exception[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.RendezvousConnectionError.html)

```python
exception tensorplay.distributed.elastic.rendezvous.RendezvousConnectionError
```

The rendezvous backend store could not be reached.

[#](#api-tensorplay.distributed.elastic.rendezvous.RendezvousError)

### RendezvousError

exception[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.RendezvousError.html)

```python
exception tensorplay.distributed.elastic.rendezvous.RendezvousError
```

Base class for all rendezvous failures.

[#](#api-tensorplay.distributed.elastic.rendezvous.RendezvousExhaustedError)

### RendezvousExhaustedError

exception[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.RendezvousExhaustedError.html)

```python
exception tensorplay.distributed.elastic.rendezvous.RendezvousExhaustedError(config, timeout)
```

The rendezvous join window elapsed without reaching min nodes.

[#](#api-tensorplay.distributed.elastic.rendezvous.RendezvousGracefulExitError)

### RendezvousGracefulExitError

exception[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.RendezvousGracefulExitError.html)

```python
exception tensorplay.distributed.elastic.rendezvous.RendezvousGracefulExitError
```

Raised to unwind an agent when it is the last node leaving.

[#](#api-tensorplay.distributed.elastic.rendezvous.RendezvousStateError)

### RendezvousStateError

exception[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.RendezvousStateError.html)

```python
exception tensorplay.distributed.elastic.rendezvous.RendezvousStateError
```

The rendezvous state is corrupt or the backend rejected an update.

[#](#api-tensorplay.distributed.elastic.rendezvous.RendezvousTimeoutError)

### RendezvousTimeoutError

exception[Full reference ↗](/docs/generated/tensorplay.distributed.elastic.rendezvous.RendezvousTimeoutError.html)

```python
exception tensorplay.distributed.elastic.rendezvous.RendezvousTimeoutError
```

A rendezvous phase exceeded its time budget.