【发布时间】：2020-03-20 21:54:46
【问题描述】：
我为 kubernetes windows pod 部署了三个后端服务，以确保它们相互通信。但是，我发现尽管容器、服务、dns 和端点都可用并且正在运行，但是当我尝试从一个容器访问任何服务（内部或外部）到另一个容器时，它仍然无法解析 dns 并收到“可以无法解析主机”或“无法连接网络无法访问”
在尝试通过 (Resources-service) 访问人员配备或聚合器服务时，收到以下错误：
C:\WINDOWS\system32>kubectl exec -it resources-deployment-6b98d47b5b-ts674 cmd
Microsoft Windows [版本 10.0.17763.864] (c) 2018 年微软公司。保留所有权利。
C:\app\resources>卷曲http://172.20.101.110:8081/swagger/index.html curl: (7) 连接172.20.101.110 8081端口失败：网络不可达
C:\app\resources>卷曲http://aggregator-service:8081/swagger/index.html curl：（6）无法解析主机：聚合器服务
C:\app\resources>卷曲http://staffing-service:8080/swagger/index.html curl：（6）无法解析主机：staffing-service
C:\app\resources>卷曲http://172.20.5.156:8080/swagger/index.html curl: (7) 连接172.20.5.156 8080端口失败：网络不可达
C:\app\resources>ping 10.91.24.250
使用 32 字节数据 Ping 10.91.24.250： PING：传输失败。一般失败。
代码很简单，这里是完整的细节。
它是一个混合环境，有两个 windows 节点和两个 linux 节点。但是，目前的要求是仅在 windows 节点上部署以下应用程序。基础设施部署在 AWS 提供的 EKS（弹性 Kubernetes 服务）上。
三个后端（aspnet）服务是：-
LoadBalancer 类型的 Staffing.yaml 文件。
Resources.Yaml 类型为 NodePort
Aggregator.Yaml 具有 NodePort 类型
要求 Resources API 应该能够连接到 Aggregator 和 Staffing 服务。而且 Staffing API 也可以连接到任何服务（资源和聚合器）
部署和服务 YAML 文件是：- 人员配备.yaml 文件
apiVersion: apps/v1
kind: Deployment
metadata:
name: staffing-deployment
#namespace: staffing-kube-system
labels:
app: staffing-app
spec:
replicas: 2
selector:
matchLabels:
app: staffing-app
template:
metadata:
labels:
app: staffing-app
spec:
nodeSelector:
beta.kubernetes.io/os: windows
#failure-domain.beta.kubernetes.io/zone: us-east-1a
containers:
- name: staffing-app
ports:
- name: http
containerPort: 8080
image: test/staffing:v1
##########################
#Staffing.service.yaml file
---
apiVersion: v1
kind: Service
metadata:
labels:
app: staffing-app
name: staffing-service
annotations:
service.beta.kubernetes.io/aws-load-balancer-internal: 0.0.0.0/0
spec:
ports:
-
name: http
#nodePort: 30475
port: 8080
protocol: TCP
targetPort: 8080
-
name: https
#nodePort: 30476
port: 443
protocol: TCP
targetPort: 44346
selector:
app: staffing-app
type: LoadBalancer
resources.service.deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: resources-deployment
#namespace: staffing-kube-system
labels:
app: resources-app
spec:
replicas: 2
selector:
matchLabels:
app: resources-app
template:
metadata:
labels:
app: resources-app
spec:
nodeSelector:
beta.kubernetes.io/os: windows
#failure-domain.beta.kubernetes.io/zone: us-east-1a
containers:
- name: resources-app
ports:
- name: http
containerPort: 8001
image: test/resources:v1
################
#resources.service.yaml
apiVersion: v1
kind: Service
metadata:
labels:
app: resources-app
name: resources-service
spec:
ports:
-
name: http
#nodePort: 30475
port: 8001
protocol: TCP
targetPort: 8001
-
name: https
#nodePort: 30476
port: 443
protocol: TCP
targetPort: 44345
selector:
app: resources-app
type: NodePort
aggregator.service.deplpoyment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: aggegator-deployment
#namespace: staffing-kube-system
labels:
app: aggegator-app
spec:
replicas: 2
selector:
matchLabels:
app: aggregator-app
template:
metadata:
labels:
app: aggregator-app
spec:
nodeSelector:
beta.kubernetes.io/os: windows
#failure-domain.beta.kubernetes.io/zone: us-east-1a
containers:
- name: aggregator-app
ports:
- name: http
containerPort: 8001
image: test/aggregator:v1
################
#aggregator.service.yaml
apiVersion: v1
kind: Service
metadata:
labels:
app: resources-app
name: resources-service
spec:
ports:
-
name: http
#nodePort: 30475
port: 8001
protocol: TCP
targetPort: 8001
-
name: https
#nodePort: 30476
port: 443
protocol: TCP
targetPort: 44345
selector:
app: resources-app
type: NodePort
Pod 和服务的输出
C:\WINDOWS\system32>kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
aggegator-deployment-64497699d-792dd 1/1 Running 0 2d12h 10.91.25.210 ip-10-91-25-125.ec2.internal <none> <none>
aggegator-deployment-64497699d-tfsww 1/1 Running 0 2d12h 10.91.24.250 ip-10-91-24-148.ec2.internal <none> <none>
resources-deployment-6b98d47b5b-7x8fv 1/1 Running 0 2d13h 10.91.25.197 ip-10-91-25-125.ec2.internal <none> <none>
resources-deployment-6b98d47b5b-ts674 1/1 Running 0 2d13h 10.91.24.122 ip-10-91-24-148.ec2.internal <none> <none>
staffing-deployment-797464966d-4j4nc 1/1 Running 0 2d16h 10.91.24.224 ip-10-91-24-148.ec2.internal <none> <none>
staffing-deployment-797464966d-mrw7x 1/1 Running 0 2d17h 10.91.24.113 ip-10-91-24-148.ec2.internal <none> <none>
C:\WINDOWS\system32>kubectl get services -o wide
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE SELECTOR
aggregator-service NodePort 172.20.101.110 <none> 8081:32246/TCP,443:30240/TCP 2d12h app=aggregator-app
kubernetes ClusterIP 172.20.0.1 <none> 443/TCP 21d <none>
resources-service NodePort 172.20.228.37 <none> 8001:31365/TCP,443:31686/TCP 2d13h app=resources-app
staffing-service LoadBalancer 172.20.5.156 internal-af561cd570d0d11eab0c00a891642f9b-874414568.us-east-1.elb.amazonaws.com 8080:30010/TCP,443:31327/TCP 2d17h app=staffing-app
C:\WINDOWS\system32>kubectl get pods --selector="app=resources-app"
NAME READY STATUS RESTARTS AGE
resources-deployment-6b98d47b5b-7x8fv 1/1 Running 0 2d21h
resources-deployment-6b98d47b5b-ts674 1/1 Running 0 2d21h
C:\WINDOWS\system32>kubectl get pods --selector="app=aggregator-app"
NAME READY STATUS RESTARTS AGE
aggegator-deployment-64497699d-792dd 1/1 Running 0 2d20h
aggegator-deployment-64497699d-tfsww 1/1 Running 0 2d20h
C:\WINDOWS\system32>kubectl get pods --selector="app=staffing-app"
NAME READY STATUS RESTARTS AGE
staffing-deployment-797464966d-4j4nc 1/1 Running 0 3d
staffing-deployment-797464966d-mrw7x 1/1 Running 0 3d1h
C:\WINDOWS\system32>kubectl get endpoints
NAME ENDPOINTS AGE
aggregator-service 10.91.24.250:8081,10.91.25.210:8081,10.91.24.250:44347 + 1 more... 2d20h
kubernetes 10.91.24.167:443,10.91.25.120:443 22d
resources-service 10.91.24.122:8001,10.91.25.197:8001,10.91.24.122:44345 + 1 more... 2d21h
staffing-service 10.91.24.113:44346,10.91.24.224:44346,10.91.24.113:8080 + 1 more... 3d2h
> Kubectl logs
C:\WINDOWS\system32>kubectl get pods --namespace=kube-system -l k8s-app=kube-dns
NAME READY STATUS RESTARTS AGE
coredns-8455f84f99-d6v75 1/1 Running 0 21d
coredns-8455f84f99-jf7tz 1/1 Running 0 21d
C:\WINDOWS\system32>kubectl logs coredns-8455f84f99-jf7tz -n kube-system
.:53
2019-11-03T14:27:30.347Z [INFO] CoreDNS-1.3.1
2019-11-03T14:27:30.347Z [INFO] linux/amd64, go1.10.8, 6b56a9c9
CoreDNS-1.3.1
linux/amd64, go1.10.8, 6b56a9c9
2019-11-03T14:27:30.347Z [INFO] plugin/reload: Running configuration MD5 = 86a350d0c3832da65c10531b445c45d5
[INFO] Reloading
2019-11-25T07:00:09.649Z [INFO] plugin/reload: Running configuration MD5 = ebe4e401326d5dac294cde91064e2e76
[INFO] Reloading complete
2019-11-25T11:38:44.107Z [INFO] 10.91.24.148:57354 - 8232 "A IN . udp 17 false 512" NOERROR qr,rd,ra 92 0.000084547s
2019-11-25T11:38:44.113Z [INFO] 10.91.24.148:57364 - 33842 "A IN . udp 17 false 512" NOERROR qr,rd,ra 92 0.000059248s
2019-11-25T11:39:15.287Z [INFO] 10.91.24.148:57366 - 23694 "A IN v10.events.data.microsoft.com. udp 47 false 512" NOERROR qr,rd,ra 284 0.003154278s
2019-11-25T11:39:15.331Z [INFO] 10.91.24.148:57366 - 23694 "A IN v10.events.data.microsoft.com. udp 47 false 512" NOERROR qr,rd,ra 284 0.069842673s
2019-11-25T11:39:15.615Z [INFO] 10.91.25.125:61221 - 24722 "A IN v10.events.data.microsoft.com. udp 47 false 512" NOERROR qr,rd,ra 284 0.000102425s
2019-11-25T11:39:17.575Z [INFO] 10.91.25.125:61223 - 8379 "A IN v10.events.data.microsoft.com. udp 47 false 512" NOERROR qr,rd,ra 284 0.000096859s
2019-11-25T11:39:18.925Z [INFO] 10.91.25.125:61239 - 49217 "A IN settings-win.data.microsoft.com. udp 49 false 512" NOERROR qr,rd,ra 176 0.030653295s
2019-11-25T11:39:19.006Z [INFO] 10.91.25.125:61239 - 49217 "A IN settings-win.data.microsoft.com. udp 49 false 512" NOERROR qr,rd,ra 176 0.142995164s
2019-11-25T11:39:19.205Z [INFO] 10.91.25.125:61241 - 25186 "A IN settings-win.data.microsoft.com. udp 49 false 512" NOERROR qr,rd,ra 176 0.000088335s
2019-11-25T11:39:43.422Z [INFO] 10.91.25.125:61257 - 26043 "A IN . udp 17 false 512" SERVFAIL qr,rd,ra 17 0.013510742s
2019-11-25T11:39:44.129Z [INFO] 10.91.24.148:57368 - 382 "A IN . udp 17 false 512" SERVFAIL qr,rd,ra 17 0.013958603s
2019-11-25T11:39:44.270Z [INFO] 10.91.25.125:61261 - 53992 "A IN . udp 17 false 512" SERVFAIL qr,rd,ra 17 0.014368787s
2019-11-25T11:39:44.981Z [INFO] 10.91.24.148:57370 - 41136 "A IN . udp 17 false 512" SERVFAIL qr,rd,ra 17 0.014143535s
2019-11-25T11:40:15.649Z [INFO] 10.91.25.125:61231 - 53428 "A IN ctldl.windowsupdate.com. udp 41 false 512" NOERROR qr,rd,ra 355 0.013625947s
2019-11-25T11:40:45.838Z [INFO] 10.91.24.148:57372 - 36065 "A IN . udp 17 false 512" SERVFAIL qr,rd,ra 17 0.013761424s
2019-11-25T11:40:46.688Z [INFO] 10.91.24.148:57382 - 26755 "A IN . udp 17 false 512" SERVFAIL qr,rd,ra 17 0.01357897s
2019-11-25T11:40:51.896Z [INFO] 10.91.24.148:57388 - 607 "A IN v10.events.data.microsoft.com. udp 47 false 512" NOERROR qr,rd,ra 284 0.033610722s
2019-11-25T11:40:52.006Z [INFO] 10.91.24.148:57388 - 607 "A IN v10.events.data.microsoft.com. udp 47 false 512" NOERROR qr,rd,ra 284 0.11348618s
2019-11-25T11:41:46.839Z [INFO] 10.91.25.125:61227 - 9202 "A IN . udp 17 false 512" SERVFAIL qr,rd,ra 17 0.013585559s
2019-11-25T11:41:47.553Z [INFO] 10.91.24.148:57339 - 12070 "A IN . udp 17 false 512" NOERROR qr,rd,ra 92 0.001922224s
2019-11-25T11:41:48.403Z [INFO] 10.91.24.148:57341 - 52311 "A IN . udp 17 false 512" NOERROR qr,rd,ra 92 0.000091287s
2019-11-25T11:42:04.827Z [INFO] 10.91.24.148:57343 - 23836 "A IN v10.events.data.microsoft.com. udp 47 false 512" NOERROR qr,rd,ra 284 0.028061963s
2019-11-25T11:42:04.827Z [INFO] 10.91.24.148:57343 - 23836 "A IN v10.events.data.microsoft.com. udp 47 false 512" NOERROR qr,rd,ra 284 0.004729734s
2019-11-25T11:42:13.855Z [hINFO] 10.91.24.148:57344 - 62298 "A IN ctldl.windowsupdate.com. udp 41 false 512" NOERROR qr,rd,ra 540 0.012435227s
对此的任何帮助将不胜感激。谢谢
【问题讨论】：
-
您在尝试访问该服务的 pod 所在的同一命名空间中是否有一个名为
aggregator-service的服务？ -
描述每个服务以检查每个 pod 的端点是否被映射
-
@suren 是的，它们都在同一个 default 命名空间中。
[NAMESPACE NAME default aggregator-service-ip default kubernetes default resources-service default resources-service-ip default staffing-service default windows-server-iis-service kube-system kube-controller-manager kube-system kube-dns kube-system kube-scheduler kube-system vpc-admission-webhook ] -
@paltaa 是的，每个服务都映射到一个 pod，请在上面提供的 pod 和服务的输出中查看相同的内容。
C:\WINDOWS\system32>kubectl get pods --selector="app=aggregator-app" NAME READY STATUS RESTARTS AGE aggegator-deployment-64497699d-792dd 1/1 Running 0 2d20h aggegator-deployment-64497699d-tfsww 1/1 Running 0 2d20h -
如果你 kubectl exec -ti
bash，然后 telnet 端口，对于你试图达到的集群 ip 是否有效？你确定每个 pod 都打开了相关的 pod 吗？
标签： kubernetes kubernetes-ingress kubernetes-pod amazon-eks windows-container