fix(node): 资源池移除节点接口，显卡管理，任务管理

server/api/v1/card.proto

@@ -65,6 +65,7 @@ message GPUReply {
   string node_uid = 10;
   bool health = 11;
   string mode = 12;
+  repeated string resource_pools = 13;
 }
 
 message GPUsReply {

server/api/v1/container.proto

@@ -61,11 +61,16 @@ message ContainerReply {
   string end_time = 11;
   string pod_uid = 12;
   string node_uid = 13;
-  string resource_pool = 14;
+  repeated string resource_pool = 14;
   string flavor = 15;
   string priority = 16;
   string namespace = 17;
   repeated string device_ids = 18;
+  string pod_name = 19;
+  string task_type = 20;
+  string shixun_name = 21;
+  string role = 22;
+  string username = 23;
 }
 
 message ContainersReply {

server/api/v1/resource_pool.proto

@@ -41,6 +41,16 @@ service ResourcePool {
     };
   }
 
+  rpc RemoveNode (RemoveNodeRequest) returns (BaseResponse) {
+    option (google.api.http) = {
+      post: "/v1/resource/pool/removeNode",
+      body: "*"
+    };
+    option (grpc.gateway.protoc_gen_openapiv2.options.openapiv2_operation) = {
+      summary: "移除节点";
+    };
+  }
+
   rpc List (ResourcePoolListRequest) returns (ResourcePoolListResponse) {
     option (google.api.http) = {
       get: "/v1/resource/pool/list"
@@ -52,7 +62,7 @@ service ResourcePool {
 
   rpc GetDetail (ResourcePoolDetailRequest) returns (ResourcePoolDetailResponse) {
     option (google.api.http) = {
-      get: "/v1/resource/pool/detail"
+      post: "/v1/resource/pool/detail"
     };
     option (grpc.gateway.protoc_gen_openapiv2.options.openapiv2_operation) = {
       summary: "资源池详情";
@@ -67,6 +77,7 @@ service ResourcePool {
       summary: "可用节点列表";
     };
   }
+
 }
 
 message BaseResponse {
@@ -118,6 +129,7 @@ message PoolNodeReply {
   string architecture = 20;
   string creation_timestamp = 21;
   int64 disk_size = 22;
+  int64 node_id = 23;
 }
 
 message ResourcePoolDetailRequest {
@@ -137,7 +149,7 @@ message ResourcePoolListData{
   int64 available_memory = 6;//kb
   int64 total_memory = 7; // kb
   int64 disk_size = 8;
-  repeated string node_list = 9;
+  repeated Nodes node_list = 9;
 }
 
 message ResourcePoolListRequest {
@@ -167,4 +179,8 @@ message AvailableNodesInfo{
 }
 
 
+message RemoveNodeRequest{
+  int64 node_id = 1;
+}
+
 

server/cmd/server/main.go

@@ -34,6 +34,7 @@ func main() {
 	flag.Parse()
 	var ctx = context.Background()
 
+	database.InitConfigPath(flagconf)
 	if err := initDatabase(); err != nil {
 		log.Errorf("数据库初始化失败: %v", err)
 		os.Exit(1)
@@ -71,13 +72,22 @@ func getNodeSelectors(c *conf.Bootstrap) map[string]string {
 }
 
 func initDatabase() error {
-	config, err := database.LoadConfig(flagconf)
-	log.Infof("config: %+v", config)
+	driver, err := database.Get("database.driver")
 	if err != nil {
 		log.Errorf("Failed to load config: %v", err)
 		return err
 	}
-	database.InitDB(&config.Database)
-	log.Infof("初始化%s成功", config.Database.Driver)
+
+	dataSourceName, err := database.Get("database.dataSourceName")
+	if err != nil {
+		log.Errorf("Failed to load config: %v", err)
+		return err
+	}
+
+	var config = &database.DatabaseConfig{}
+	config.Driver = driver.(string)
+	config.DataSourceName = dataSourceName.(string)
+	database.InitDB(config)
+	log.Infof("初始化%s成功", driver)
 	return nil
 }

server/config/config.yaml

@@ -15,4 +15,6 @@ node_selectors:
   MLU: mlu=on
 database:
   driver: mysql
-  dataSourceName: testeducoder:TEST@123@tcp(testeducoder-public.mysql.polardb.rds.aliyuncs.com:3306)/hami?parseTime=true&loc=Local
+  dataSourceName: testeducoder:TEST@123@tcp(testeducoder-public.mysql.polardb.rds.aliyuncs.com:3306)/hami?parseTime=true&loc=Local
+web_domain: http://172.16.100.14
+big_model_resource_pool_name: "大模型资源池"

server/internal/biz/pod.go

@@ -21,6 +21,7 @@ type Container struct {
 	Priority         string
 	NodeUID          string
 	Namespace        string
+	TpiID            string
 }
 
 type PodInfo struct {
@@ -31,6 +32,7 @@ type PodInfo struct {
 	Devices   PodDevices
 	CtrIDs    []string
 	Ctrs      []*Container
+	Labels    map[string]string
 }
 
 type PodRepo interface {

server/internal/data/pod.go

@@ -22,15 +22,17 @@ type podRepo struct {
 	data      *Data
 	podLister listerscorev1.PodLister
 	pods      map[k8stypes.UID]*biz.PodInfo
+	allPods   []*biz.PodInfo
 	mutex     sync.RWMutex
 	log       *log.Helper
 }
 
 func NewPodRepo(data *Data, logger log.Logger) biz.PodRepo {
 	repo := &podRepo{
-		data: data,
-		pods: make(map[k8stypes.UID]*biz.PodInfo),
-		log:  log.NewHelper(logger),
+		data:    data,
+		pods:    make(map[k8stypes.UID]*biz.PodInfo),
+		allPods: []*biz.PodInfo{},
+		log:     log.NewHelper(logger),
 	}
 	repo.init()
 	return repo
@@ -91,8 +93,9 @@ func (r *podRepo) addPod(pod *corev1.Pod, nodeID string, devices biz.PodDevices)
 	r.mutex.Lock()
 	defer r.mutex.Unlock()
 	ctrs := r.fetchContainerInfo(pod)
-	pi := &biz.PodInfo{Name: pod.Name, UID: pod.UID, Namespace: pod.Namespace, NodeID: nodeID, Devices: devices, Ctrs: ctrs}
+	pi := &biz.PodInfo{Name: pod.Name, UID: pod.UID, Namespace: pod.Namespace, NodeID: nodeID, Devices: devices, Ctrs: ctrs, Labels: pod.Labels}
 	r.pods[pod.UID] = pi
+	r.allPods = append(r.allPods, pi)
 	r.log.Infof("Pod added: Name: %s, UID: %s, Namespace: %s, NodeID: %s", pod.Name, pod.UID, pod.Namespace, nodeID)
 }
 
@@ -185,7 +188,11 @@ func (r *podRepo) GetStartTime(pod *corev1.Pod) time.Time {
 func (r *podRepo) ListAll(context.Context) ([]*biz.Container, error) {
 	var containerList []*biz.Container
 	for _, pod := range r.pods {
-		containerList = append(containerList, pod.Ctrs...)
+		TpiID := pod.Labels["tpi-id"]
+		for _, container := range pod.Ctrs {
+			container.TpiID = TpiID
+			containerList = append(containerList, container)
+		}
 	}
 	return containerList, nil
 }

server/internal/database/config.go

@@ -1,31 +1,97 @@
 package database
 
 import (
+	"encoding/json"
 	"fmt"
+	"github.com/go-kratos/kratos/v2/log"
 	"gopkg.in/yaml.v3"
 	"os"
+	"strings"
+	"sync"
 )
 
-type DatabaseConfig struct {
-	Driver         string `yaml:"driver"`
-	DataSourceName string `yaml:"dataSourceName"`
-}
+var (
+	configData map[string]interface{}
+	loadOnce   sync.Once
+	loadErr    error
+	configPath string
+)
 
-type Config struct {
-	Database DatabaseConfig `yaml:"database"`
+// InitConfigPath 设置配置路径（可选）
+func InitConfigPath(path string) {
+	configPath = path
 }
 
-func LoadConfig(filePath string) (*Config, error) {
+// loadYAML 加载 YAML 到 map
+func loadYAML(filePath string) (map[string]interface{}, error) {
 	yamlFile, err := os.ReadFile(filePath)
 	if err != nil {
 		return nil, fmt.Errorf("failed to read config file: %v", err)
 	}
 
-	var config Config
-	err = yaml.Unmarshal(yamlFile, &config)
+	var raw map[string]interface{}
+	err = yaml.Unmarshal(yamlFile, &raw)
+	if err != nil {
+		return nil, fmt.Errorf("failed to unmarshal yaml: %v", err)
+	}
+
+	b, _ := json.MarshalIndent(raw, "", "  ")
+	log.Info("loadYAML: ", string(b))
+	converted, ok := toStringKeyMap(raw).(map[string]interface{})
+	if !ok {
+		return nil, fmt.Errorf("failed to convert config to map[string]interface{}")
+	}
+
+	return converted, nil
+}
+
+// GetConfig 获取全局配置 map，只加载一次
+func GetConfig() (map[string]interface{}, error) {
+	loadOnce.Do(func() {
+		if configPath == "" {
+			configPath = "config.yaml"
+		}
+		configData, loadErr = loadYAML(configPath)
+	})
+	return configData, loadErr
+}
+
+// Get 获取嵌套配置，例如 Get("database.driver")
+func Get(key string) (interface{}, error) {
+	cfg, err := GetConfig()
 	if err != nil {
-		return nil, fmt.Errorf("failed to unmarshal config file: %v", err)
+		return nil, err
+	}
+
+	keys := strings.Split(key, ".")
+	var val interface{} = cfg
+
+	for _, k := range keys {
+		m, ok := val.(map[string]interface{})
+		if !ok {
+			return nil, fmt.Errorf("invalid path: %s", key)
+		}
+		val, ok = m[k]
+		if !ok {
+			return nil, fmt.Errorf("key not found: %s", key)
+		}
 	}
+	return val, nil
+}
 
-	return &config, nil
+func toStringKeyMap(i interface{}) interface{} {
+	switch x := i.(type) {
+	case map[interface{}]interface{}:
+		m2 := map[string]interface{}{}
+		for k, v := range x {
+			keyStr := fmt.Sprintf("%v", k)
+			m2[keyStr] = toStringKeyMap(v)
+		}
+		return m2
+	case []interface{}:
+		for i, v := range x {
+			x[i] = toStringKeyMap(v)
+		}
+	}
+	return i
 }

server/internal/database/init_db.go

@@ -8,6 +8,15 @@ import (
 
 var db *sql.DB
 
+type DatabaseConfig struct {
+	Driver         string `yaml:"driver"`
+	DataSourceName string `yaml:"dataSourceName"`
+}
+
+type Config struct {
+	Database DatabaseConfig `yaml:"database"`
+}
+
 func InitDB(config *DatabaseConfig) {
 	var err error
 	switch config.Driver {

server/internal/database/resource_pool_db.go

@@ -180,6 +180,37 @@ func QueryResourceNamesByIp(nodeIp string) ([]string, error) {
 	return resourcePoolNames, nil
 }
 
+func QueryResourceNamesByNodeName(nodeName string) ([]string, error) {
+	// 执行查询
+	rows, err := db.Query("select pool_name from resource_pool where id in (select distinct pool_id from nodes where node_name=?)", nodeName)
+	if err != nil {
+		log.Infof("Query failed: %v", err)
+		return nil, err
+	}
+	defer rows.Close()
+
+	// 存放结果的切片
+	resourcePoolNames := make([]string, 0)
+
+	// 遍历每一行
+	for rows.Next() {
+		var name string
+		err := rows.Scan(&name)
+		if err != nil {
+			log.Infof("Scan failed: %v", err)
+			return nil, err
+		}
+		resourcePoolNames = append(resourcePoolNames, name)
+	}
+
+	// 检查 rows 是否遍历中出错
+	if err := rows.Err(); err != nil {
+		return nil, err
+	}
+
+	return resourcePoolNames, nil
+}
+
 func InsertResourcePool(poolName string) (int64, error) {
 	querySql := "INSERT INTO resource_pool(pool_name) VALUES (?)"
 
@@ -272,3 +303,18 @@ func DeleteNodesByPoolId(poolId int64) (int64, error) {
 
 	return rowsAffected, nil
 }
+
+func DeleteNodeById(nodeId int64) (int64, error) {
+	result, err := db.Exec("DELETE FROM nodes WHERE id = ?", nodeId)
+	if err != nil {
+		return 0, fmt.Errorf("delete failed: %w", err)
+	}
+
+	// 返回影响的行数（0 表示未删除任何数据）
+	rowsAffected, err := result.RowsAffected()
+	if err != nil {
+		return 0, fmt.Errorf("get rows affected failed: %w", err)
+	}
+
+	return rowsAffected, nil
+}

server/internal/service/card.go

@@ -7,6 +7,7 @@ import (
 	"strings"
 	pb "vgpu/api/v1"
 	"vgpu/internal/biz"
+	"vgpu/internal/database"
 )
 
 type CardService struct {
@@ -51,6 +52,11 @@ func (s *CardService) GetAllGPUs(ctx context.Context, req *pb.GetAllGpusReq) (*p
 		gpu.NodeUid = device.NodeUid
 		gpu.Health = device.Health
 		gpu.Mode = device.Mode
+		resourcePoolNames, err := database.QueryResourceNamesByNodeName(device.NodeName)
+		if err != nil {
+			return nil, err
+		}
+		gpu.ResourcePools = resourcePoolNames
 
 		vGPU, core, memory, err := s.pod.StatisticsByDeviceId(ctx, device.AliasId)
 		if err == nil {

server/internal/service/container.go

@@ -2,11 +2,16 @@ package service
 
 import (
 	"context"
+	"encoding/json"
+	"github.com/go-kratos/kratos/v2/log"
+	"slices"
 	"sort"
 	"strings"
 	"time"
 	pb "vgpu/api/v1"
 	"vgpu/internal/biz"
+	"vgpu/internal/database"
+	"vgpu/internal/utils"
 )
 
 var statusOrder = map[string]int{
@@ -86,6 +91,32 @@ func (s *ContainerService) GetAllContainers(ctx context.Context, req *pb.GetAllC
 		if containerReply.DeviceIds == nil {
 			continue
 		}
+
+		resourcePoolNames, err := database.QueryResourceNamesByNodeName(container.NodeName)
+		if err != nil {
+			return nil, err
+		}
+
+		containerReply.ResourcePool = resourcePoolNames
+		resourcePoolName, err := database.Get("big_model_resource_pool_name")
+		if err != nil {
+			return nil, err
+		}
+
+		if slices.Contains(resourcePoolNames, resourcePoolName.(string)) {
+			containerReply.TaskType = "big_model"
+		} else {
+			containerReply.TaskType = "shixun"
+		}
+
+		if len(container.TpiID) > 0 {
+			err := s.setShixunData(ctx, containerReply, container.TpiID)
+			if err != nil {
+				return nil, err
+			}
+		}
+
+		containerReply.PodName = container.PodName
 		containerReply.CreateTime = container.CreateTime.Format(time.RFC3339)
 		res.Items = append(res.Items, containerReply)
 	}
@@ -127,3 +158,34 @@ func (s *ContainerService) GetContainer(ctx context.Context, req *pb.GetContaine
 	ctrReply.CreateTime = container.CreateTime.Format(time.RFC3339)
 	return ctrReply, nil
 }
+
+func (s *ContainerService) setShixunData(ctx context.Context, containerReply *pb.ContainerReply, tpiId string) error {
+	webDomain, err := database.Get("web_domain")
+	if err != nil {
+		return err
+	}
+
+	client := utils.GetDefaultClient()
+	url := webDomain.(string) + "/api/myshixuns/get_shixun_info.json"
+	log.Info("Get shixun info url: ", url, " tpiId: ", tpiId)
+	jsonData := map[string]interface{}{
+		"tpiID": tpiId,
+	}
+	body, status, err := client.PostJSON(ctx, url, jsonData, nil)
+	if err != nil {
+		return err
+	}
+	log.Infof("Get shixun info: %s, status: %d", string(body), status)
+
+	var respMap map[string]interface{}
+	err = json.Unmarshal(body, &respMap)
+	log.Info("Get shixun info: ", respMap, "----", respMap["status"])
+	if respMap["status"].(float64) == 0 {
+		data := respMap["data"].(map[string]interface{})
+		containerReply.ShixunName = data["shixun_name"].(string)
+		containerReply.Role = data["user_identity"].(string)
+		containerReply.Username = data["user_name"].(string)
+	}
+
+	return nil
+}

server/internal/service/resource_pool.go

@@ -109,6 +109,17 @@ func (s *ResourcePoolService) Delete(ctx context.Context, req *pb.ResourcePoolDe
 	return &pb.BaseResponse{Code: 200, Message: "成功"}, nil
 }
 
+func (s *ResourcePoolService) RemoveNode(ctx context.Context, req *pb.RemoveNodeRequest) (*pb.BaseResponse, error) {
+	log.Info("RemoveNode called", req)
+	nodeId := req.NodeId
+	num, err := database.DeleteNodeById(nodeId)
+	if err != nil {
+		return &pb.BaseResponse{Code: 500, Message: "移除节点失败"}, nil
+	}
+	log.Infof("RemoveNode success poolId: %d, 影响行数: %d", nodeId, num)
+	return &pb.BaseResponse{Code: 200, Message: "成功"}, nil
+}
+
 func (s *ResourcePoolService) List(ctx context.Context, req *pb.ResourcePoolListRequest) (*pb.ResourcePoolListResponse, error) {
 	log.Info("GetResourcePoolList", req)
 
@@ -137,7 +148,10 @@ func (s *ResourcePoolService) List(ctx context.Context, req *pb.ResourcePoolList
 			poolData.TotalMemory = poolData.TotalMemory + node.TotalMemory
 			poolData.AvailableMemory = poolData.AvailableMemory + node.AvailableMemory
 			poolData.DiskSize = poolData.DiskSize + node.DiskTotal
-			poolData.NodeList = append(poolData.NodeList, n.NodeIp)
+			poolData.NodeList = append(poolData.NodeList, &pb.Nodes{
+				NodeIp:   n.NodeIp,
+				NodeName: n.NodeName,
+			})
 		}
 		data = append(data, &poolData)
 	}
@@ -165,6 +179,7 @@ func (s *ResourcePoolService) GetDetail(ctx context.Context, req *pb.ResourcePoo
 			continue
 		}
 		nodeReply, err := s.buildNodeReply(ctx, node)
+		nodeReply.NodeId = poolNode.Id
 		if err != nil {
 			return nil, err
 		}

server/internal/utils/httpclient.go

@@ -0,0 +1,110 @@
+package utils
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"io"
+	"net/http"
+	"net/url"
+	"strings"
+	"sync"
+	"time"
+)
+
+type HttpClient struct {
+	client *http.Client
+}
+
+// NewHttpClient 创建带超时的 HTTP 客户端
+func NewHttpClient(timeout time.Duration) *HttpClient {
+	return &HttpClient{
+		client: &http.Client{
+			Timeout: timeout,
+		},
+	}
+}
+
+// Get 发送 GET 请求
+func (hc *HttpClient) Get(ctx context.Context, rawUrl string, headers map[string]string) ([]byte, int, error) {
+	req, err := http.NewRequestWithContext(ctx, http.MethodGet, rawUrl, nil)
+	if err != nil {
+		return nil, 0, err
+	}
+	for k, v := range headers {
+		req.Header.Set(k, v)
+	}
+
+	resp, err := hc.client.Do(req)
+	if err != nil {
+		return nil, 0, err
+	}
+	defer resp.Body.Close()
+
+	body, err := io.ReadAll(resp.Body)
+	return body, resp.StatusCode, err
+}
+
+// PostJSON 发送 POST 请求，Body 是 JSON
+func (hc *HttpClient) PostJSON(ctx context.Context, rawUrl string, data interface{}, headers map[string]string) ([]byte, int, error) {
+	bodyBytes, err := json.Marshal(data)
+	if err != nil {
+		return nil, 0, err
+	}
+
+	req, err := http.NewRequestWithContext(ctx, http.MethodPost, rawUrl, bytes.NewReader(bodyBytes))
+	if err != nil {
+		return nil, 0, err
+	}
+	req.Header.Set("Content-Type", "application/json")
+	for k, v := range headers {
+		req.Header.Set(k, v)
+	}
+
+	resp, err := hc.client.Do(req)
+	if err != nil {
+		return nil, 0, err
+	}
+	defer resp.Body.Close()
+
+	body, err := io.ReadAll(resp.Body)
+	return body, resp.StatusCode, err
+}
+
+// PostForm 发送 POST 表单请求
+func (hc *HttpClient) PostForm(ctx context.Context, rawUrl string, formData map[string]string, headers map[string]string) ([]byte, int, error) {
+	data := url.Values{}
+	for k, v := range formData {
+		data.Set(k, v)
+	}
+
+	req, err := http.NewRequestWithContext(ctx, http.MethodPost, rawUrl, strings.NewReader(data.Encode()))
+	if err != nil {
+		return nil, 0, err
+	}
+	req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+	for k, v := range headers {
+		req.Header.Set(k, v)
+	}
+
+	resp, err := hc.client.Do(req)
+	if err != nil {
+		return nil, 0, err
+	}
+	defer resp.Body.Close()
+
+	body, err := io.ReadAll(resp.Body)
+	return body, resp.StatusCode, err
+}
+
+var (
+	defaultClient *HttpClient
+	once          sync.Once
+)
+
+func GetDefaultClient() *HttpClient {
+	once.Do(func() {
+		defaultClient = NewHttpClient(10 * time.Second)
+	})
+	return defaultClient
+}

server/openapi.yaml

@@ -3,42 +3,37 @@
 
 openapi: 3.0.3
 info:
-    title: Node API
+    title: ResourcePool API
     version: 0.0.1
 paths:
-    /v1/node:
+    /v1/available/nodes:
         get:
             tags:
-                - Node
-            operationId: Node_GetNode
-            parameters:
-                - name: uid
-                  in: query
-                  schema:
-                    type: string
+                - ResourcePool
+            operationId: ResourcePool_GetAvailableNodes
             responses:
                 "200":
                     description: OK
                     content:
                         application/json:
                             schema:
-                                $ref: '#/components/schemas/NodeReply'
+                                $ref: '#/components/schemas/AvailableNodesResponse'
                 default:
                     description: Default error response
                     content:
                         application/json:
                             schema:
                                 $ref: '#/components/schemas/Status'
-    /v1/node/discovered:
+    /v1/resource/pool/create:
         post:
             tags:
-                - Node
-            operationId: Node_DiscoveredNode
+                - ResourcePool
+            operationId: ResourcePool_Create
             requestBody:
                 content:
                     application/json:
                         schema:
-                            $ref: '#/components/schemas/DiscoveredNodeRequest'
+                            $ref: '#/components/schemas/ResourcePoolCreateRequest'
                 required: true
             responses:
                 "200":
@@ -46,23 +41,23 @@ paths:
                     content:
                         application/json:
                             schema:
-                                $ref: '#/components/schemas/DiscoveredNodeResponse'
+                                $ref: '#/components/schemas/BaseResponse'
                 default:
                     description: Default error response
                     content:
                         application/json:
                             schema:
                                 $ref: '#/components/schemas/Status'
-    /v1/node/join:
+    /v1/resource/pool/delete:
         post:
             tags:
-                - Node
-            operationId: Node_JoinNode
+                - ResourcePool
+            operationId: ResourcePool_Delete
             requestBody:
                 content:
                     application/json:
                         schema:
-                            $ref: '#/components/schemas/JoinNodeRequest'
+                            $ref: '#/components/schemas/ResourcePoolDeleteRequest'
                 required: true
             responses:
                 "200":
@@ -70,47 +65,64 @@ paths:
                     content:
                         application/json:
                             schema:
-                                $ref: '#/components/schemas/JoinNodeResponse'
+                                $ref: '#/components/schemas/BaseResponse'
                 default:
                     description: Default error response
                     content:
                         application/json:
                             schema:
                                 $ref: '#/components/schemas/Status'
-    /v1/node/status/update:
+    /v1/resource/pool/detail:
         post:
             tags:
-                - Node
-            operationId: Node_UpdateNodeStatus
-            requestBody:
-                content:
-                    application/json:
-                        schema:
-                            $ref: '#/components/schemas/UpdateNodeStatusRequest'
-                required: true
+                - ResourcePool
+            operationId: ResourcePool_GetDetail
+            parameters:
+                - name: poolId
+                  in: query
+                  schema:
+                    type: string
+            responses:
+                "200":
+                    description: OK
+                    content:
+                        application/json:
+                            schema:
+                                $ref: '#/components/schemas/ResourcePoolDetailResponse'
+                default:
+                    description: Default error response
+                    content:
+                        application/json:
+                            schema:
+                                $ref: '#/components/schemas/Status'
+    /v1/resource/pool/list:
+        get:
+            tags:
+                - ResourcePool
+            operationId: ResourcePool_List
             responses:
                 "200":
                     description: OK
                     content:
                         application/json:
                             schema:
-                                $ref: '#/components/schemas/UpdateNodeStatusResponse'
+                                $ref: '#/components/schemas/ResourcePoolListResponse'
                 default:
                     description: Default error response
                     content:
                         application/json:
                             schema:
                                 $ref: '#/components/schemas/Status'
-    /v1/nodes:
+    /v1/resource/pool/removeNode:
         post:
             tags:
-                - Node
-            operationId: Node_GetAllNodes
+                - ResourcePool
+            operationId: ResourcePool_RemoveNode
             requestBody:
                 content:
                     application/json:
                         schema:
-                            $ref: '#/components/schemas/GetAllNodesReq'
+                            $ref: '#/components/schemas/RemoveNodeRequest'
                 required: true
             responses:
                 "200":
@@ -118,23 +130,23 @@ paths:
                     content:
                         application/json:
                             schema:
-                                $ref: '#/components/schemas/NodesReply'
+                                $ref: '#/components/schemas/BaseResponse'
                 default:
                     description: Default error response
                     content:
                         application/json:
                             schema:
                                 $ref: '#/components/schemas/Status'
-    /v1/summary:
+    /v1/resource/pool/update:
         post:
             tags:
-                - Node
-            operationId: Node_GetSummary
+                - ResourcePool
+            operationId: ResourcePool_Update
             requestBody:
                 content:
                     application/json:
                         schema:
-                            $ref: '#/components/schemas/GetSummaryReq'
+                            $ref: '#/components/schemas/ResourcePoolUpdateRequest'
                 required: true
             responses:
                 "200":
@@ -142,7 +154,7 @@ paths:
                     content:
                         application/json:
                             schema:
-                                $ref: '#/components/schemas/DeviceSummaryReply'
+                                $ref: '#/components/schemas/BaseResponse'
                 default:
                     description: Default error response
                     content:
@@ -151,78 +163,40 @@ paths:
                                 $ref: '#/components/schemas/Status'
 components:
     schemas:
-        DeviceSummaryReply:
+        AvailableNodesInfo:
             type: object
             properties:
-                vgpuUsed:
-                    type: integer
-                    format: int32
-                vgpuTotal:
-                    type: integer
-                    format: int32
-                coreUsed:
-                    type: integer
-                    format: int32
-                coreTotal:
-                    type: integer
-                    format: int32
-                memoryUsed:
-                    type: integer
-                    format: int32
-                memoryTotal:
-                    type: integer
-                    format: int32
-                gpuCount:
-                    type: integer
-                    format: int32
-                nodeCount:
-                    type: integer
-                    format: int32
-        DiscoveredNodeInfo:
-            type: object
-            properties:
-                nodeIp:
-                    type: string
                 nodeName:
                     type: string
-        DiscoveredNodeRequest:
-            type: object
-            properties: {}
-        DiscoveredNodeResponse:
-            type: object
-            properties:
-                list:
-                    type: array
-                    items:
-                        $ref: '#/components/schemas/DiscoveredNodeInfo'
-        GetAllNodesReq:
-            type: object
-            properties:
-                filters:
-                    $ref: '#/components/schemas/GetAllNodesReq_Filters'
-        GetAllNodesReq_Filters:
-            type: object
-            properties:
-                ip:
+                cpuCores:
                     type: string
-                type:
+                gpuNum:
                     type: string
-                isSchedulable:
+                gpuMemory:
+                    type: string
+                totalMemory:
+                    type: string
+                diskSize:
+                    type: string
+                nodeIp:
                     type: string
-        GetSummaryReq:
+        AvailableNodesResponse:
             type: object
             properties:
-                filters:
-                    $ref: '#/components/schemas/GetSummaryReq_Filters'
-        GetSummaryReq_Filters:
+                data:
+                    type: array
+                    items:
+                        $ref: '#/components/schemas/AvailableNodesInfo'
+        BaseResponse:
             type: object
             properties:
-                type:
-                    type: string
-                nodeUid:
-                    type: string
-                deviceId:
+                code:
+                    type: integer
+                    format: int32
+                message:
                     type: string
+                data:
+                    type: object
         GoogleProtobufAny:
             type: object
             properties:
@@ -231,22 +205,14 @@ components:
                     description: The type of the serialized message.
             additionalProperties: true
             description: Contains an arbitrary serialized message along with a @type that describes the type of the serialized message.
-        JoinNodeRequest:
+        Nodes:
             type: object
             properties:
-                nodeNames:
-                    type: array
-                    items:
-                        type: string
-        JoinNodeResponse:
-            type: object
-            properties:
-                code:
-                    type: integer
-                    format: int32
-                message:
+                nodeIp:
+                    type: string
+                nodeName:
                     type: string
-        NodeReply:
+        PoolNodeReply:
             type: object
             properties:
                 ip:
@@ -271,8 +237,7 @@ components:
                 coreTotal:
                     type: string
                 memoryUsed:
-                    type: integer
-                    format: int32
+                    type: string
                 memoryTotal:
                     type: string
                 uid:
@@ -300,47 +265,90 @@ components:
                     type: string
                 diskSize:
                     type: string
-                resourcePools:
+                nodeId:
+                    type: string
+        RemoveNodeRequest:
+            type: object
+            properties:
+                nodeId:
+                    type: string
+        ResourcePoolCreateRequest:
+            type: object
+            properties:
+                poolName:
+                    type: string
+                nodes:
                     type: array
                     items:
-                        type: string
-        NodesReply:
+                        $ref: '#/components/schemas/Nodes'
+        ResourcePoolDeleteRequest:
+            type: object
+            properties:
+                poolId:
+                    type: string
+        ResourcePoolDetailResponse:
             type: object
             properties:
                 list:
                     type: array
                     items:
-                        $ref: '#/components/schemas/NodeReply'
-        Status:
+                        $ref: '#/components/schemas/PoolNodeReply'
+        ResourcePoolListData:
             type: object
             properties:
-                code:
-                    type: integer
-                    description: The status code, which should be an enum value of [google.rpc.Code][google.rpc.Code].
-                    format: int32
-                message:
+                poolId:
                     type: string
-                    description: A developer-facing error message, which should be in English. Any user-facing error message should be localized and sent in the [google.rpc.Status.details][google.rpc.Status.details] field, or localized by the client.
-                details:
+                poolName:
+                    type: string
+                cpuCores:
+                    type: string
+                nodeNum:
+                    type: string
+                gpuNum:
+                    type: string
+                availableMemory:
+                    type: string
+                totalMemory:
+                    type: string
+                diskSize:
+                    type: string
+                nodeList:
                     type: array
                     items:
-                        $ref: '#/components/schemas/GoogleProtobufAny'
-                    description: A list of messages that carry the error details.  There is a common set of message types for APIs to use.
-            description: 'The `Status` type defines a logical error model that is suitable for different programming environments, including REST APIs and RPC APIs. It is used by [gRPC](https://github.com/grpc). Each `Status` message contains three pieces of data: error code, error message, and error details. You can find out more about this error model and how to work with it in the [API Design Guide](https://cloud.google.com/apis/design/errors).'
-        UpdateNodeStatusRequest:
+                        $ref: '#/components/schemas/Nodes'
+        ResourcePoolListResponse:
             type: object
             properties:
-                nodeName:
+                data:
+                    type: array
+                    items:
+                        $ref: '#/components/schemas/ResourcePoolListData'
+        ResourcePoolUpdateRequest:
+            type: object
+            properties:
+                poolId:
                     type: string
-                status:
+                poolName:
                     type: string
-        UpdateNodeStatusResponse:
+                nodes:
+                    type: array
+                    items:
+                        $ref: '#/components/schemas/Nodes'
+        Status:
             type: object
             properties:
                 code:
                     type: integer
+                    description: The status code, which should be an enum value of [google.rpc.Code][google.rpc.Code].
                     format: int32
                 message:
                     type: string
+                    description: A developer-facing error message, which should be in English. Any user-facing error message should be localized and sent in the [google.rpc.Status.details][google.rpc.Status.details] field, or localized by the client.
+                details:
+                    type: array
+                    items:
+                        $ref: '#/components/schemas/GoogleProtobufAny'
+                    description: A list of messages that carry the error details.  There is a common set of message types for APIs to use.
+            description: 'The `Status` type defines a logical error model that is suitable for different programming environments, including REST APIs and RPC APIs. It is used by [gRPC](https://github.com/grpc). Each `Status` message contains three pieces of data: error code, error message, and error details. You can find out more about this error model and how to work with it in the [API Design Guide](https://cloud.google.com/apis/design/errors).'
 tags:
-    - name: Node
+    - name: ResourcePool