算力API

API 凭证

获取 API 凭证

在平台右上角的下拉菜单中选择 API 凭证，将跳转至 API 凭证页面。

点击新增 Token，输入自定义的 Token 名称，默认有效期为 3650 天，可根据需要进行调整。

点击确定后，您将看到新生成的 Token 显示在列表中。请注意，Token 仅会生成一次，请务必及时复制并妥善保存！如果您遗失了 Token，可以选择新增 Token 或 重新生成。

获取 region id 和 clusterName

region_id：地区 id
clusterName：集群名称

请求

GET https://demo.cdc.datenfab.com/mdc/api/v1/regions

响应

python代码：

import requests
headers = {
    "namespace" "$userid",
    "Authorization": "bearer $token"
}

url = "https://demo.cdc.datenfab.com/mdc/api/v1/regions"

response = requests.get(url, headers=headers)
print(response.content.decode())

返回：

[{"id":"14c5a9e3-0d09-4aa0-99d3-c096b80cceaa","region":"北京一区","clusterName":"host-cluster","harborAddr":"images.neolink-ai.com","howlerID":"69e377a7-99d1-cca0-ef58-cfce14a336b2","createdAt":"2024-07-15T12:00:23.227+08:00","updatedAt":"2024-07-15T12:00:23.227+08:00"}]

查看 GPU 规格信息

请求

GET https://demo.cdc.datenfab.com/mdc/api/v1/skus/computings?region_id=$region_id

响应

python代码：

import requests
headers = {
     "namespace" "$userid",
    "Authorization": "bearer $token"
}

url = "https://demo.cdc.datenfab.com/mdc/api/v1/skus/computings?region_id=14c5a9e3-0d09-4aa0-99d3-c096b80cceaa"
response = requests.get(url, headers=headers)

print(response.content.decode())

返回：

[{"name":"H20","regionID":"14c5a9e3-0d09-4aa0-99d3-c096b80cceaa","cpuArch":"x86-64","cpuSeries":"Inter Xeon","cpucores":24,"memoryCount":192,"cpuVendor":"Inter","gpuSeries":"H20","gpuProduct":"H20","gpuResourceKey":"nvidia.com/gpu-h20","gpuCounts":1,"gpuMemory":96,"systemstorage":30,"datastorage":50,"skusID":"2","regionName":"北京一区","clusterName":"host-cluster","sellingType":"1","ratio":0.48,"discountID":"10000","discountStartTime":"2024-08-15T15:26:38+08:00","discountEndTime":"2024-10-01T00:05:00+08:00","price":10,"remainGPUCount":8},...]

必要参数说明：

参数	说明
name	GPU 型号
regionID	region id
gpuCounts	GPU 数量
sellingType	消费类型，1为按量，2为包月，3为包周，4为包日

创建算力实例

请求

POST https://demo.cdc.datenfab.com/mdc/api/v1/clusters/$clusteName/instances

Body 请求参数：

{
  "Cost": {
    "CostType": "",
    "CouponID": "",
    "DiskSKUID": "",
    "GPUSKUID": "",
    "CostTime": ""
  },
  "VolumeMount": [
    {
      "PathInContainer": "",
      "Size": ,
      "Type": ""
    }
  ],
  "namespace": "",
  "displayName": "",
  "image": "",
  "Region": "",
}

body 参数说明：

名称	类型	必选	说明
Cost	object	是	GPU 规格信息
CostType	string	是	消费类型，1为按量，2为包月，3为包周，4为包日
CouponID	string	否	优惠券 id，算力券只能用于按量的消费类型
DiskSKUID	string	是	磁盘 id ，目前只有 1
GPUSKUID	string	是	GPU 规则 id，根据章节查看GPU规格信息获取
CostTime	string	否	付费时长：D，W，M, 例如3天：3D，当更新付费时长时，请在前面加+，例如，新增3天：+3D
VolumeMount	object	是	磁盘挂载信息
PathInContainer	string	否	挂载路径，默认是/root/data
Size	integer	否	磁盘大小，单位G
Type	string	否	磁盘类型，目前仅支持 shared-filesystem
namespace	string	是	命名空间，和用户 id 相同
displayName	string	是	实例名称
image	string	是	镜像名，包括官方镜像和私有镜像
Region	string	否	地区，默认北京一区

响应

Python 代码：

import requests
import json
headers = {
    "namespace" "$userid",
    "Authorization": "bearer $token"
}

url = "https://demo.cdc.datenfab.com/mdc/api/v1/clusters/host-cluster/instances"

body={
  "Cost": {
    "CostType": "1",
    "CouponID": "",
    "DiskSKUID": "1",
    "GPUSKUID": "66",
    "CostTime": ""
  },
  "VolumeMount": [
    {
      "PathInContainer": "/root/data",
      "Size": 50,
      "Type": "shared-filesystem"
    }
  ],
  "namespace": "xxx",
  "displayName": "test1",
  "image": "images.neolink-ai.com/matrixdc/tensorflow:2.15.0-python3.10-cuda12.1.0-cudnn8-devel-ubuntu22.04",
  "Region": ""
}
response = requests.post(url, json=body, headers=headers)

print(json.dumps(response.json(), indent=4))

返回：

{
    "name": "instance-ccffxxxx",
    "namespace": "xxx",
    "Region": "host-cluster",
    "displayName": "test1",
    "description": "",
    "annotations": {
        "demo.cdc.datenfab.com/cost-coupon-id": "",
        "demo.cdc.datenfab.com/cost-discount-id": "",
        "demo.cdc.datenfab.com/cost-disk-sku-id": "1",
        "demo.cdc.datenfab.com/cost-gpu-sku-id": "66",
        "demo.cdc.datenfab.com/cost-order-id": "94b9e467-9d34-45bf-8e8f-e04115cf58b4",
        "demo.cdc.datenfab.com/cost-time": "",
        "demo.cdc.datenfab.com/cost-type": "1",
        "demo.cdc.datenfab.com/description": "",
        "demo.cdc.datenfab.com/display-name": "test1",
        "demo.cdc.datenfab.com/gpu.machine": "4090",
        "demo.cdc.datenfab.com/gpu.resource.key": "nvidia.com/gpu-4090",
        "demo.cdc.datenfab.com/region": "host-cluster",
        "demo.cdc.datenfab.com/saveDataEnabled": "false",
        "demo.cdc.datenfab.com/unit-price": "250"
    },
    "uid": "8d40c2e6-a0b0-45c2-9b40-a0de065104f5",
    "image": "images.neolink-ai.com/matrixdc/tensorflow:2.15.0-python3.10-cuda12.1.0-cudnn8-devel-ubuntu22.04",
    "replicas": 1,
    "resources": {
        "cpu": "16",
        "memory": "96Gi",
        "GPUNum": "1",
        "GPUType": "4090"
    },
    "volumeMounts": [
        {
            "PathInContainer": "/",
            "Type": "local-filesystem",
            "Size": 30,
            "PVCName": "",
            "displayName": "\u7cfb\u7edf\u76d8",
            "filesystemType": "system"
        },
        {
            "PathInContainer": "/root/data",
            "Type": "shared-filesystem",
            "Size": 0,
            "PVCName": "volume-920020f7",
            "displayName": "\u78c1\u76d8-volume-920020f7",
            "filesystemType": "data"
        }
    ],
    "creationTimestamp": "2024-09-02T18:34:33+08:00",
    "updateTimestamp": "0001-01-01T00:00:00Z",
    "ScheduleStopTimestamp": "0001-01-01T00:00:00Z",
    "status": "Creating",
    "cost": {
        "CostType": "1",
        "CostTime": "",
        "DiscountID": "",
        "GPUSKUID": "66",
        "DiskSKUID": "1",
        "CouponID": ""
    }
}

获取实例列表

请求

GET https://demo.cdc.datenfab.com/mdc/api/v1/clusters/$clusterName/instances

响应

Python 代码：

import requests
import json
headers = {
    "namespace" "$userid",
    "Authorization": "bearer $token"
}

url = "https://demo.cdc.datenfab.com/mdc/api/v1/clusters/host-cluster/instances"
response = requests.get(url, headers=headers)

data = response.json()

print(json.dumps(data, indent=4))

返回：

{
    "hasNextPage": false,
    "nodes": [
        {
            "name": "instance-ccffxxxx",
            "namespace": "xxx",
            "Region": "host-cluster",
            "displayName": "test1",
            "description": "",
            "annotations": {
                "demo.cdc.datenfab.com/cost-coupon-id": "",
                "demo.cdc.datenfab.com/cost-discount-id": "",
                "demo.cdc.datenfab.com/cost-disk-sku-id": "1",
                "demo.cdc.datenfab.com/cost-gpu-sku-id": "66",
                "demo.cdc.datenfab.com/cost-order-id": "94b9e467-9d34-45bf-8e8f-e04115cf58b4",
                "demo.cdc.datenfab.com/cost-time": "",
                "demo.cdc.datenfab.com/cost-type": "1",
                "demo.cdc.datenfab.com/description": "",
                "demo.cdc.datenfab.com/display-name": "test1",
                "demo.cdc.datenfab.com/gpu.machine": "4090",
                "demo.cdc.datenfab.com/gpu.resource.key": "nvidia.com/gpu-4090",
                "demo.cdc.datenfab.com/region": "host-cluster",
                "demo.cdc.datenfab.com/saveDataEnabled": "false",
                "demo.cdc.datenfab.com/unit-price": "250"
            },
            "uid": "8d40c2e6-a0b0-45c2-9b40-a0de065104f5",
            "servers": {
                "Jupyterlab": "http://8d40c2e6-a0b0-45c2-9b40-a0de065104f5.gw.neolink-ai.com/jupyter?token=7d27443c5718e75bee2da9f92fe6a50c",
                "Tensorboard": "http://8d40c2e6-a0b0-45c2-9b40-a0de065104f5.gw.neolink-ai.com/monitor",
                "SSH": {
                    "sshAddr": "ssh -p30099 root@8d40c2e6-a0b0-45c2-9b40-a0de065104f5.gws.demo.cdc.datenfab.com",
                    "sshPassword": "kEyIX5R2dqJD"
                }
            },
            "image": "images.neolink-ai.com/matrixdc/tensorflow:2.15.0-python3.10-cuda12.1.0-cudnn8-devel-ubuntu22.04",
            "replicas": 1,
            "resources": {
                "cpu": "16",
                "memory": "96Gi",
                "GPUNum": "1",
                "GPUType": "4090"
            },
            "volumeMounts": [
                {
                    "PathInContainer": "/",
                    "Type": "local-filesystem",
                    "Size": 30,
                    "PVCName": "",
                    "displayName": "\u7cfb\u7edf\u76d8",
                    "filesystemType": "system"
                },
                {
                    "PathInContainer": "/root/data",
                    "Type": "shared-filesystem",
                    "Size": 50,
                    "PVCName": "volume-920020f7",
                    "displayName": "\u78c1\u76d8-volume-920020f7",
                    "filesystemType": "data"
                }
            ],
            "creationTimestamp": "2024-09-02T18:34:33+08:00",
            "updateTimestamp": "0001-01-01T00:00:00Z",
            "ScheduleStopTimestamp": "0001-01-01T00:00:00Z",
            "status": "Running",
            "cost": {
                "CostType": "1",
                "CostTime": "",
                "DiscountID": "",
                "GPUSKUID": "66",
                "DiskSKUID": "1",
                "CouponID": ""
            },
            "Pods": [
                {
                    "Name": "instance-ccff84c2-dd6nm",
                    "Containers": [
                        {
                            "Name": "worker"
                        }
                    ],
                    "nodeScheduled": "gpu-4090-03",
                    "ip": "172.31.213.79"
                }
            ]
        }
    ],
    "totalCount": 1
}

获取实例详情

请求

GET https://demo.cdc.datenfab.com/mdc/api/v1/clusters/$clusterName/instances/$instant_id

响应

Python 代码：

import requests
import json
headers = {
    "namespace" :"$userid",
    "Authorization": "bearer $token"
}

url = "https://demo.cdc.datenfab.com/mdc/api/v1/clusters/host-cluster/instances/instance-4ca9xxxx"
response = requests.get(url, headers=headers)

data = response.json()

print(json.dumps(data, indent=4))

返回：

{
    "name": "instance-4ca9xxxx",
    "namespace": "xxx",
    "Region": "host-cluster",
    "displayName": "test1",
    "description": "",
    "annotations": {
        "demo.cdc.datenfab.com/cost-coupon-id": "",
        "demo.cdc.datenfab.com/cost-discount-id": "",
        "demo.cdc.datenfab.com/cost-disk-sku-id": "1",
        "demo.cdc.datenfab.com/cost-gpu-sku-id": "66",
        "demo.cdc.datenfab.com/cost-order-id": "cf4c99ff-cf7f-4c86-8f15-87cf5fb2ba14",
        "demo.cdc.datenfab.com/cost-time": "",
        "demo.cdc.datenfab.com/cost-type": "1",
        "demo.cdc.datenfab.com/description": "",
        "demo.cdc.datenfab.com/display-name": "test1",
        "demo.cdc.datenfab.com/gpu.machine": "4090",
        "demo.cdc.datenfab.com/gpu.resource.key": "nvidia.com/gpu-4090",
        "demo.cdc.datenfab.com/region": "host-cluster",
        "demo.cdc.datenfab.com/unit-price": "250"
    },
    "uid": "a7d2cbab-591d-4482-95ea-2df994ed1b22",
    "servers": {
        "Jupyterlab": "http://a7d2cbab-591d-4482-95ea-2df994ed1b22.gw.neolink-ai.com/jupyter?token=23a7676a0d18451dee9c08c3649a8288",
        "Tensorboard": "http://a7d2cbab-591d-4482-95ea-2df994ed1b22.gw.neolink-ai.com/monitor",
        "SSH": {
            "sshAddr": "ssh -p30379 root@a7d2cbab-591d-4482-95ea-2df994ed1b22.gws.demo.cdc.datenfab.com",
            "sshPassword": "RuRbjZLWklIC"
        }
    },
    "image": "images.neolink-ai.com/matrixdc/tensorflow:2.15.0-python3.10-cuda12.1.0-cudnn8-devel-ubuntu22.04",
    "replicas": 1,
    "resources": {
        "cpu": "16",
        "memory": "96Gi",
        "GPUNum": "1",
        "GPUType": "4090"
    },
    "volumeMounts": [
        {
            "PathInContainer": "/",
            "Type": "local-filesystem",
            "Size": 30,
            "PVCName": "",
            "displayName": "\u7cfb\u7edf\u76d8",
            "filesystemType": "system"
        },
        {
            "PathInContainer": "/root/data",
            "Type": "shared-filesystem",
            "Size": 50,
            "PVCName": "volume-ed087526",
            "displayName": "\u78c1\u76d8-volume-ed087526",
            "filesystemType": "data"
        }
    ],
    "creationTimestamp": "2024-09-02T22:15:13+08:00",
    "updateTimestamp": "0001-01-01T00:00:00Z",
    "ScheduleStopTimestamp": "0001-01-01T00:00:00Z",
    "status": "Running",
    "cost": {
        "CostType": "1",
        "CostTime": "",
        "DiscountID": "",
        "GPUSKUID": "66",
        "DiskSKUID": "1",
        "CouponID": ""
    },
    "Pods": [
        {
            "Name": "instance-4ca97f58-6m9nv",
            "Containers": [
                {
                    "Name": "worker"
                }
            ],
            "nodeScheduled": "gpu-4090-03",
            "ip": "172.31.213.103"
        }
    ]
}

停止实例

请求

POST https://demo.cdc.datenfab.com/mdc/api/v1/clusters/$clusterName/instances/instance_name/power_off

body 请求参数：

body = {
    "gracePeriodSeconds": "",
    "needSnapshot": "",
    "imageName": "",
    "tag": ""
}

body 参数说明：

名称	类型	必选	说明
gracePeriodSeconds	string	否	Pod 体面终止限期，单位秒，系统首先发送停止系统指令，然后在强制删除之前等待容器内进程处理的时间, 默认为10秒
needSnapshot	string	否	是否需要再关机的时候保存运行 container 为 image，默认为 False
imageName	string	否	保存的镜像名称
tag	string	否	镜像版本

响应

Python 代码：

import requests
headers = {
    "namespace" :"$userid",
    "Authorization": "bearer $token"
}

url = "https://demo.cdc.datenfab.com/mdc/api/v1/clusters/host-cluster/instances/instance-3ba5xxxx/power_off"

body = {
    "gracePeriodSeconds": 10,
    "needSnapshot": True,
    "imageName": "image1",
    "tag": "v1.1.0"
}

response = requests.post(url,json=body,headers=headers)
print(response.content.decode())

返回：

"ok"

启动实例

请求

POST https://demo.cdc.datenfab.com/mdc/api/v1/clusters/$clusterName/instances/$instance_id/power_off

响应

import requests
headers = {
    "namespace" :"$userid",
    "Authorization": "bearer $token"
}

url = "https://demo.cdc.datenfab.com/mdc/api/v1/clusters/host-cluster/instances/instance-4ca9xxxx/power_on"

response = requests.post(url, headers=headers)
print(response.content.decode())

删除实例

请求

DELETE https://demo.cdc.datenfab.com/mdc/api/v1/clusters/$clusterName/instances?name=$instance_id

响应

Python 代码：

import requests
headers = {
    "namespace" :"$userid",
    "Authorization": "bearer $token"
}

url = "https://demo.cdc.datenfab.com/mdc/api/v1/clusters/host-cluster/instances?name=instance-ccffxxxx"

response = requests.delete(url, headers=headers)

print(response.content.decode())

返回：

"ok"

API 凭证​

获取 API 凭证​

获取 region id 和 clusterName​

请求​

响应​

查看 GPU 规格信息​

请求​

响应​

创建算力实例​

请求​

响应​

获取实例列表​

请求​

响应​

获取实例详情​

请求​

响应​

停止实例​

请求​

响应​

启动实例​

请求​

响应​

删除实例​

请求​

响应​

API 凭证

获取 API 凭证

获取 region id 和 clusterName

请求

响应

查看 GPU 规格信息

请求

响应

创建算力实例

请求

响应

获取实例列表

请求

响应

获取实例详情

请求

响应

停止实例

请求

响应

启动实例

请求

响应

删除实例

请求

响应