接口功能
指定实例身份。
接口格式
操作类型：POST
URL：https://{ip}:{port}/v2/role/{role}
- {ip}字段优先读取环境变量值MIES_CONTAINER_MANAGEMENT_IP；如果没有该环境变量，则取配置文件的“managementIpAddress”参数；如果配置文件中没有“managementIpAddress”参数，则取配置文件的“ipAddress”参数。
- {port}字段优先读取配置文件的“managementPort”参数；如果配置文件中没有“managementPort”参数，则取配置文件的“port”参数。
请求参数
参数 |
是否必选 |
类型 |
描述 |
|---|---|---|---|
local |
必选 |
结构体 |
实例信息。 |
local[].dp_inst_list |
必选 |
结构体数组 |
当前Prefill/Decode实例下的DP组信息列表。 |
local[].dp_inst_list[].device |
必选 |
结构体数组 |
当前DP组下NPU卡相关信息，可能存在多卡。 |
local[].dp_inst_list[].device[].device_id |
必选 |
整型字符串 |
NPU卡的物理ID。 |
local[].dp_inst_list[].device[].device_ip |
必选 |
字符串 |
NPU卡的IP地址。 |
local[].dp_inst_list[].device[].rank_id |
必选 |
整型字符串 |
NPU卡的Rank ID。 |
local[].dp_inst_list[].device[].device_logical_id |
必选 |
整型字符串 |
NPU卡的逻辑ID。 |
local[].dp_inst_list[].dp_inst_id |
必选 |
整型字符串 |
当前Prefill/Decode实例下的DP组ID。 |
local[].host_ip |
必选 |
字符串 |
当前节点的Host IP地址。 |
local[].server_ip |
必选 |
字符串 |
当前节点的Server IP地址。 |
peers |
必选 |
结构体数组 |
表示需要建连的对象列表。每个实例只能看到与自己有关的实例。
|
peers[].dp_inst_list |
必选 |
结构体数组 |
对端Prefill/Decode实例下的DP组信息列表。 |
peers[].dp_inst_list[].device |
必选 |
结构体数组 |
对端DP组下NPU卡相关信息，可能存在多卡。 |
peers[].dp_inst_list[].device[].device_id |
必选 |
整型字符串 |
对端DP组下NPU卡的物理ID。 |
peers[].dp_inst_list[].device[].device_ip |
必选 |
字符串 |
对端DP组下NPU卡的IP地址。 |
peers[].dp_inst_list[].device[].rank_id |
必选 |
整型字符串 |
对端DP组下NPU卡的Rank ID。 |
peers[].dp_inst_list[].device[].device_logical_id |
必选 |
整型字符串 |
对端DP组下NPU卡的逻辑ID。 |
peers[].dp_inst_list[].dp_inst_id |
必选 |
整型字符串 |
对端Prefill/Decode实例下的DP组ID。 |
peers[].host_ip |
必选 |
字符串 |
对端节点的Host IP地址。 |
peers[].server_ip |
必选 |
字符串 |
对端节点的Server IP地址。 |
使用样例
请求样例：
POST https://{ip}:{port}/v2/role/prefill
请求消息体：
{
"local": [
{
"dp_inst_list": [
{
"device": [
{
"device_id": "0",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "0",
"device_logical_id": "0"
},
{
"device_id": "1",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "1"
},
{
"device_id": "2",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "2",
"device_logical_id": "2"
},
{
"device_id": "3",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "3",
"device_logical_id": "3"
}
],
"dp_inst_id": 1111110000
},
{
"device": [
{
"device_id": "4",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "4",
"device_logical_id": "4"
},
{
"device_id": "5",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "5",
"device_logical_id": "5"
},
{
"device_id": "6",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "6",
"device_logical_id": "6"
},
{
"device_id": "7",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "7",
"device_logical_id": "7"
}
],
"dp_inst_id": 1111110001
}
],
"host_ip": "xxx.xxx.xxx.xxx",
"server_ip": "xxx.xxx.xxx.xxx"
},
{
"dp_inst_list": [
{
"device": [
{
"device_id": "0",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "0"
},
{
"device_id": "1",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "1"
},
{
"device_id": "2",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "2"
},
{
"device_id": "3",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "3"
}
],
"dp_inst_id": 1111110002
},
{
"device": [
{
"device_id": "4",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "4"
},
{
"device_id": "5",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "5"
},
{
"device_id": "6",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "6"
},
{
"device_id": "7",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "7"
}
],
"dp_inst_id": 1111110003
}
],
"host_ip": "141.61.33.76",
"server_ip": "192.168.165.113"
}
],
"peers": [
[
{
"dp_inst_list": [
{
"device": [
{
"device_id": "0",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "0"
},
{
"device_id": "1",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "1"
},
{
"device_id": "2",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "2"
},
{
"device_id": "3",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "3"
}
],
"dp_inst_id": 2222220000
},
{
"device": [
{
"device_id": "4",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "4"
},
{
"device_id": "5",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "5"
},
{
"device_id": "6",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "6"
},
{
"device_id": "7",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "7"
}
],
"dp_inst_id": 2222220001
}
],
"host_ip": "141.61.33.78",
"server_ip": "192.168.195.39"
},
{
"dp_inst_list": [
{
"device": [
{
"device_id": "0",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "0",
"device_logical_id": "0"
},
{
"device_id": "1",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "1",
"device_logical_id": "1"
},
{
"device_id": "2",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "2",
"device_logical_id": "2"
},
{
"device_id": "3",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "3",
"device_logical_id": "3"
}
],
"dp_inst_id": 2222220002
},
{
"device": [
{
"device_id": "4",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "4",
"device_logical_id": "4"
},
{
"device_id": "5",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "5",
"device_logical_id": "5"
},
{
"device_id": "6",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "6",
"device_logical_id": "6"
},
{
"device_id": "7",
"device_ip": "xxx.xxx.xxx.xxx",
"rank_id": "7",
"device_logical_id": "7"
}
],
"dp_inst_id": 2222220003
}
],
"host_ip": "xxx.xxx.xxx.xxx",
"server_ip": "xxx.xxx.xxx.xxx"
}
]
]
}
响应样例：
- 成功：
{ "result": "OK", } - 失败：
{ "error": "xxx", "error_type": "xxxxx" }
响应状态码：
- 200：成功。
- 400：错误请求。
- 422：请求体非法。
输出说明
参数 |
类型 |
描述 |
|---|---|---|
result |
string |
成功。 |
error |
string |
错误描述。 |
error_type |
string |
错误类型。 |