You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
#When using in a production environment, it is recommended to set domainID to be globally unique. It is recommended to use: "company name"-"department name"-"node name", such as:
18
18
# domainID: mycompany-secretflow-trainlite
19
19
domainID: alice
20
-
#节点私钥配置, 用于节点间的通信认证, 节点应用的证书签发
21
-
#执行命令 "docker run -it --rm secretflow-registry.cn-hangzhou.cr.aliyuncs.com/secretflow/kuscia scripts/deploy/generate_rsa_key.sh" 生成私钥
20
+
#Node private key configuration, used for communication authentication between nodes and certificate issuance for node applications.
21
+
#Run the command "docker run -it --rm secretflow-registry.cn-hangzhou.cr.aliyuncs.com/secretflow/kuscia scripts/deploy/generate_rsa_key.sh" to generate the private key.
#When the node first deploys and connects to the Master, the Master uses this Token to verify the node's identity (Token issued by the Master). For security reasons, this Token is invalidated immediately after the node is successfully deployed.
41
+
#When deploying on multiple machines, keep this Token unchanged.
42
+
#If the node's private key is lost, delete the node's public key on the Master and reapply for a Token for deployment.
#Task scheduling to the specified organization's K8s namespace
55
55
namespace: ""
56
-
#机构 K8s 集群的 pod dns 配置,用于解析节点的应用域名,runk 拉起 pod 所使用的 dns 地址,应配置为 kuscia service 的 clusterIP
56
+
#pod dns configuration of the organization's K8s cluster, used to resolve the application domain name of the node. The dns address used by runk to pull up the pod should be configured as the clusterIP of the kuscia service.
#kubeconfig of the organization's K8s cluster, default is serviceaccount if not filled; currently, do not fill, use serviceaccount by default
59
59
kubeconfigFile:
60
-
#是否开启 kuscia pod 日志记录,默认为 false (不开启),当开启时需要在rbac.yaml (示例:https://github.com/secretflow/kuscia/blob/main/hack/k8s/autonomy/rbac.yaml) 里开通pods/log权限
60
+
#Whether to enable kuscia pod log recording, default is false (disabled). When enabled, pods/log permissions need to be enabled in rbac.yaml (example: https://github.com/secretflow/kuscia/blob/main/hack/k8s/autonomy/rbac.yaml)
#The capacity that the node can use for scheduling applications. runc/runp automatically obtains the current container's system resources if not filled, manual configuration is required in runk mode
64
64
capacity:
65
65
cpu: #4
66
66
memory: #8Gi
67
67
pods: #500
68
68
storage: #100Gi
69
69
ephemeralStorage: #100Gi
70
70
71
-
# agent 镜像配置
71
+
# agent image configuration
72
72
image:
73
-
pullPolicy: #是否允许拉取远程镜像(remote)|仅使用本地已导入镜像(local)
73
+
pullPolicy: #Whether to allow pulling remote images (remote) | only use locally imported images (local)
74
74
defaultRegistry: ""
75
-
#拉取镜像的代理地址,如:http://127.0.0.1:8080|不填则不使用代理
75
+
#Proxy address for pulling images, such as: http://127.0.0.1:8080 | do not use proxy if not filled
#Workload approval configuration, note: this configuration only takes effect in P2P networking. KusciaJob execution does not require approval in centralized networking.
90
+
#By default, workload approval configuration is disabled. If approval configuration is enabled, all Jobs need to call KusciaAPI for job approval when the local party is a participant. Production environment is recommended to enable approval.
# It is recommended to check the generated file in advance after the command is executed to avoid deployment startup problems caused by configuration file errors
0 commit comments