From d762691ecce7c879afaf74b32a0bf5cac05f5fea Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E4=B9=8B=E6=B1=9F=E5=AE=9E=E9=AA=8C=E5=AE=A4?= Date: Fri, 15 Jan 2021 13:40:32 +0800 Subject: [PATCH] fix config --- .../resources/config/application-prod.yml | 71 ++++++++++++------ .../resources/config/application-prod.yml | 75 +++++++++++++------ 2 files changed, 102 insertions(+), 44 deletions(-) diff --git a/dubhe-server/dubhe-admin/src/main/resources/config/application-prod.yml b/dubhe-server/dubhe-admin/src/main/resources/config/application-prod.yml index 05431f6..739a15a 100644 --- a/dubhe-server/dubhe-admin/src/main/resources/config/application-prod.yml +++ b/dubhe-server/dubhe-admin/src/main/resources/config/application-prod.yml @@ -96,6 +96,16 @@ k8s: host: notebook.test.com # k8s ingress-controller 对外port，获取方式：部署 ingress-controller 后，在k8s master节点执行 kubectl get svc -A | grep 'ingress-nginx-controller' 获取80对应的外部端口 port: 33334 + # k8s ingress-controller 对外grpc port + https-port: 31365 + # k8s 模型部署配置 + serving: + # k8s 模型部署域名 如需测试需修改为合适的域名 + host: serving.test.com + # tls 证书 crt + tls-crt: + # tls 证书 key + tls-key: # elasticsearch暴露的服务地址，获取方式 部署 管理集群日志 后，在k8s msater节点执行 kubectl get svc -A | grep 'elasticsearch' 获取9200对应的外部端口 elasticsearch: hostlist: ${eshostlist:127.0.0.1:33333} @@ -117,6 +127,9 @@ k8s: pod: metrics: grafanaUrl: http://127.0.0.1:30006/d/job/monitor?orgId=1&refresh=5s&kiosk&var-pod= + prometheus: + query-url: http://127.0.0.1:30003/api/v1/query + gpu-query-param: sum(container_accelerator_duty_cycle{pod="{}"})by(pod,acc_id) nfs-storage-class-name: zjlab-nfs-storage #配置harbor harbor: @@ -128,32 +141,14 @@ harbor: # data模块配置 data: annotation: - # 自动标注服务url,最后一位需加'/' - endpoint: http://127.0.0.1/annotation-dev/ task: # 自动标注任务分割的文件split size splitSize: 16 - # 单位ms - retryInterval: 5000 - # 任务不更新置为失败的时间单位秒 - failTime: 43200 - track: - # 自动追踪服务url - endpoint: http://127.0.0.1:7091/ server: # 文件存储服务器用户名 userName: root - enhance: - # 数据增强 - endpoint: http://127.0.0.1:8070/ - imageNet: - # imageNet - endpoint: http://127.0.0.1:8092/ - ofrecord: - # 二进制转换 - endpoint: http://127.0.0.1:8093/ - #数据集训练配置 - ptversion: http://127.0.0.1:8000/ + # 数据集训练配置 + ptversion: localhost:${server.port}/ # minio配置 minio: @@ -166,6 +161,40 @@ minio: presignedUrlExpiryTime: 300 annotation: /annotation/ +# 数据处理医学影像数据集dcm服务器配置，查看影像功能需要使用该服务 +dcm: + host: 10.5.29.100 + port: 11112 + +#镜像脚本路径 +image: + shell-path: /data/prod/dubhe-prod/ #logback logging.config: - classpath:logback-spring-dev.xml \ No newline at end of file + classpath:logback-spring-dev.xml +#是否开启 swagger-ui +swagger: + enabled: false + +docker: + host: 127.0.0.1 + port: 2375 + +# 配置转发 +dubhe-proxy: + visual: + keyword: visual + server: localhost + port: 9898 + refine: + keyword: refine + server: localhost + port: 9797 + +# serving模块配置 +serving: + sourcePath: "/serving/TS_Serving" + gateway-uri-postfix: .api.dubhe.ai + rootPath: "/serving/deployment/" + batchRootPath: "/serving/batch/" + group: serving_prod_group \ No newline at end of file diff --git a/dubhe-server/dubhe-task/src/main/resources/config/application-prod.yml b/dubhe-server/dubhe-task/src/main/resources/config/application-prod.yml index bbfd110..739a15a 100644 --- a/dubhe-server/dubhe-task/src/main/resources/config/application-prod.yml +++ b/dubhe-server/dubhe-task/src/main/resources/config/application-prod.yml @@ -1,5 +1,5 @@ server: - port: 9527 + port: 8000 # rest API 版本号 rest-version: v1 @@ -83,7 +83,7 @@ spring: show: true k8s: - # k8s集群配置文件,将k8s集群master下$HOME/.kube/config文件 复制到dubhe-task/src/main/resources/下 重命名为 kubeconfig + # k8s集群配置文件 kubeconfig: kubeconfig # nfs服务暴露的IP地址 如需测试需修改为合适的地址 nfs: 127.0.0.1 @@ -96,6 +96,16 @@ k8s: host: notebook.test.com # k8s ingress-controller 对外port，获取方式：部署 ingress-controller 后，在k8s master节点执行 kubectl get svc -A | grep 'ingress-nginx-controller' 获取80对应的外部端口 port: 33334 + # k8s ingress-controller 对外grpc port + https-port: 31365 + # k8s 模型部署配置 + serving: + # k8s 模型部署域名 如需测试需修改为合适的域名 + host: serving.test.com + # tls 证书 crt + tls-crt: + # tls 证书 key + tls-key: # elasticsearch暴露的服务地址，获取方式 部署 管理集群日志 后，在k8s msater节点执行 kubectl get svc -A | grep 'elasticsearch' 获取9200对应的外部端口 elasticsearch: hostlist: ${eshostlist:127.0.0.1:33333} @@ -117,6 +127,9 @@ k8s: pod: metrics: grafanaUrl: http://127.0.0.1:30006/d/job/monitor?orgId=1&refresh=5s&kiosk&var-pod= + prometheus: + query-url: http://127.0.0.1:30003/api/v1/query + gpu-query-param: sum(container_accelerator_duty_cycle{pod="{}"})by(pod,acc_id) nfs-storage-class-name: zjlab-nfs-storage #配置harbor harbor: @@ -128,32 +141,14 @@ harbor: # data模块配置 data: annotation: - # 自动标注服务url,最后一位需加'/' - endpoint: http://127.0.0.1/annotation-dev/ task: # 自动标注任务分割的文件split size splitSize: 16 - # 单位ms - retryInterval: 5000 - # 任务不更新置为失败的时间单位秒 - failTime: 43200 - track: - # 自动追踪服务url - endpoint: http://127.0.0.1:7091/ server: # 文件存储服务器用户名 userName: root - enhance: - # 数据增强 - endpoint: http://127.0.0.1:8070/ - imageNet: - # imageNet - endpoint: http://127.0.0.1:8092/ - ofrecord: - # 二进制转换 - endpoint: http://127.0.0.1:8093/ - #数据集训练配置 - ptversion: http://127.0.0.1:8000/ + # 数据集训练配置 + ptversion: localhost:${server.port}/ # minio配置 minio: @@ -166,6 +161,40 @@ minio: presignedUrlExpiryTime: 300 annotation: /annotation/ +# 数据处理医学影像数据集dcm服务器配置，查看影像功能需要使用该服务 +dcm: + host: 10.5.29.100 + port: 11112 + +#镜像脚本路径 +image: + shell-path: /data/prod/dubhe-prod/ #logback logging.config: - classpath:logback-spring-dev.xml \ No newline at end of file + classpath:logback-spring-dev.xml +#是否开启 swagger-ui +swagger: + enabled: false + +docker: + host: 127.0.0.1 + port: 2375 + +# 配置转发 +dubhe-proxy: + visual: + keyword: visual + server: localhost + port: 9898 + refine: + keyword: refine + server: localhost + port: 9797 + +# serving模块配置 +serving: + sourcePath: "/serving/TS_Serving" + gateway-uri-postfix: .api.dubhe.ai + rootPath: "/serving/deployment/" + batchRootPath: "/serving/batch/" + group: serving_prod_group \ No newline at end of file