check_link_status

Applicability

Note: For the Atlas A2 training product/Atlas A2 inference product, only the Atlas 800I A2 inference server and A200I A2 Box heterogeneous subrack are supported.

Function Description

Quickly checks whether the link status is normal.

Prototype

1

check_link_status(remote_cluster_id: int)

Parameters

Example

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

from llm_datadist import LLMDataDist, LLMRole, LLMStatusCode, LLMClusterInfo
...
try:
    data_dist.check_link_status(remote_cluster_id=0)
except LLMException as ex:
    print(f"check_link_status exception:{ex.status_code}")
    raise ex
kv_cache_manager = data_dist.kv_cache_manager
...
kv_cache_manager.pull_cache(prompt_cache_key, local_kv_cache, batch_index=0)

Returns

In normal cases, no value is returned.

If the execution fails, an LLMException is thrown.

If a parameter is incorrect, a TypeError or ValueError may be thrown.

Constraints

This API must be called by the client.

If the call fails and an unrecoverable error code is reported, the link needs to be re-established.

If the call to this API fails, you must keep calling it until it succeeds before calling APIs such as pull_cache and pull_blocks.

If this API is called concurrently with APIs such as pull_cache and pull_blocks, an LLMException may be thrown with the error code LLM_LINK_BUSY.

The timeout period is specified by llm.SyncKvCacheWaitTime.