kubernetes 集群运行在两个节点上。一位师傅，一位 worker …… weave net 是 pod 网络。
[root@irf-centos1 ~]# kubectl cluster-info
Kubernetes master is running at https://10.8.156.184:6443
KubeDNS is running at https://10.8.156.184:6443/api/v1/namespaces/kube-system/services/kube-dns:dns/proxy
To further debug and diagnose cluster problems, use 'kubectl cluster-info dump'.
已将rabbit docker 镜像部署为kubernetes pod 中的容器。
[root@irf-centos1 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
rabbitmq-86bd97fd9d-8h444 1/1 Running 0 51m
rabbitmq-86bd97fd9d-n2kgk 1/1 Running 0 51m
以下是服务和部署yaml文件
部署文件
---
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
name: rabbitmq
spec:
replicas: 2
template:
metadata:
labels:
app: rabbitmqapp
spec:
containers:
-
image: "docker.io/rabbitmq:latest"
imagePullPolicy: Always
name: rabbitmq
ports:
-
containerPort: 5672
name: http-port
volumeMounts:
-
mountPath: /var/rabbitmqapp_home
name: rabbitmqapp-home
volumes:
-
emptyDir: {}
name: rabbitmqapp-home
服务文件
---
apiVersion: v1
kind: Service
metadata:
name: rabbitmq
namespace: default
spec:
ports:
-
port: 5672
targetPort: 5672
protocol: TCP
nodePort: 31111
selector:
app: rabbitmqapp
type: NodePort
这里是服务和部署详细信息
[root@irf-centos1 ~]# kubectl get deployments
NAME READY UP-TO-DATE AVAILABLE AGE
rabbitmq 2/2 2 2 55m
[root@irf-centos1 ~]# kubectl get services
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kubernetes ClusterIP 10.96.0.1 <none> 443/TCP 3d
rabbitmq NodePort 10.98.204.64 <none> 5672:31111/TCP 55m
现在，当我尝试点击节点端口上的rabbitmq仪表板/UI时。无法访问
[root@irf-centos1 ~]# curl http://10.8.156.187:31111
curl: (56) Recv failure: Connection reset by peer
AMQP [root@irf-centos1 ~]#
当我从网络浏览器点击相同的 URL 时，没有任何反应
请提出建议
注意:此集群是在 AZure VM 上使用 kubeadm 部署的。为了进行故障排除，我已经打开了这些虚拟机上的所有入站/出站端口，因此，这不应该是防火墙、端口阻塞问题。
编辑 1:
我按如下方式修改了服务文件并重新部署了它。公共(public)安全局
---
apiVersion: v1
kind: Service
metadata:
name: rabbitmq
namespace: default
spec:
ports:
- name: ui
protocol: TCP
port: 15672
targetPort: 15672
nodePort: 31112
- name: service
port: 5672
targetPort: 5672
protocol: TCP
nodePort: 31111
selector:
app: rabbitmq
type: NodePort
仍然遇到同样的错误
[root@irf-centos1 ~]# curl -I http://guest:<a href="https://stackoverflow.com/cdn-cgi/l/email-protection" class="__cf_email__" data-cfemail="7d1a08180e093d4c4d5345534c484b534c454a" rel="noreferrer noopener nofollow">[email protected]</a>:31111/api/users
curl: (56) Recv failure: Connection reset by peer
AMQP [root@irf-centos1 ~]# curl -I http://guest:<a href="https://stackoverflow.com/cdn-cgi/l/email-protection" class="__cf_email__" data-cfemail="8fe8faeafcfbcfbebfa1b7a1bebab9a1beb7b8" rel="noreferrer noopener nofollow">[email protected]</a>:31112/api/users
curl: (7) Failed connect to 10.8.156.187:31112; Connection refused
最佳答案
对于rabbitmq仪表板/UI，它运行在:15672
因此服务文件中的端口号应包含:15672
然后访问仪表板/UI 为应用程序创建用户。然后使用该用户curl。
---
apiVersion: v1
kind: Service
metadata:
name: rabbitmq
namespace: default
spec:
ports:
-
port: 5672
targetPort: 5672
protocol: TCP
nodePort: 31111
-
protocol: TCP
port: 15672
targetPort: 15672
nodePort: 31112
selector:
app: rabbitmqapp
type: NodePort
或者使用rabbitmq的默认用户名和密码guest:
curl -I http://guest:<a href="https://stackoverflow.com/cdn-cgi/l/email-protection" class="__cf_email__" data-cfemail="3e594b5b4d4a7e0f0e1006100f0b08100f0609" rel="noreferrer noopener nofollow">[email protected]</a>:31112/api/users
部署文件添加containerPort:
image: "docker.io/rabbitmq:latest"
imagePullPolicy: Always
name: rabbitmq
ports:
-
containerPort: 5672
name: http-port
containerPort: 15672
name: ui-port
关于docker - 无法从 kubernetes 集群访问rabbitmq仪表板，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/56233490/