hazelcast-kubernetes 无法启动嵌入式服务器

Question

我正在尝试在 kubernetes 上的 spring 引导应用程序中启动一个简单的嵌入式 hazelcast 服务器。我正在使用 hazelcast-kubernetes 插件并遵循文档 here。我尝试了 LoadBalancer 和 NodePort 两种服务类型。

我还尝试使用 this example 部署示例应用程序（使用相同的 docker 图像）并尝试了 LoadBalancer 和 NodePort 服务类型。

在这两种情况下，我在重试 10 次连接到 Kubernetes Master 后都收到以下错误。其中a.b.c.d是kubernetes主节点的IP地址。

2020-03-09 18:20:57.322  INFO 1 --- [           main] c.h.c.LifecycleService                   : [a.b.c.d]:5701 [dev] [3.12] [a.b.c.d]:5701 is STARTING
2020-03-09 18:20:57.820  WARN 1 --- [           main] c.h.k.RetryUtils                         : Couldn't connect to the Kubernetes master, [1] retrying in 1 seconds...
2020-03-09 18:20:59.339  WARN 1 --- [           main] c.h.k.RetryUtils                         : Couldn't connect to the Kubernetes master, [2] retrying in 2 seconds...
2020-03-09 18:21:01.713  WARN 1 --- [           main] c.h.k.RetryUtils                         : Couldn't connect to the Kubernetes master, [3] retrying in 3 seconds...
2020-03-09 18:21:05.113  WARN 1 --- [           main] c.h.k.RetryUtils                         : Couldn't connect to the Kubernetes master, [4] retrying in 5 seconds...
2020-03-09 18:21:10.198  WARN 1 --- [           main] c.h.k.RetryUtils                         : Couldn't connect to the Kubernetes master, [5] retrying in 7 seconds...
2020-03-09 18:21:17.807  WARN 1 --- [           main] c.h.k.RetryUtils                         : Couldn't connect to the Kubernetes master, [6] retrying in 11 seconds...
2020-03-09 18:21:49.216  WARN 1 --- [           main] c.h.k.RetryUtils                         : Couldn't connect to the Kubernetes master, [7] retrying in 17 seconds...
2020-03-09 18:22:06.319  WARN 1 --- [           main] c.h.k.RetryUtils                         : Couldn't connect to the Kubernetes master, [8] retrying in 25 seconds...
2020-03-09 18:22:31.963  WARN 1 --- [           main] c.h.k.RetryUtils                         : Couldn't connect to the Kubernetes master, [9] retrying in 38 seconds...
2020-03-09 18:23:15.423  WARN 1 --- [           main] c.h.k.RetryUtils                         : Couldn't connect to the Kubernetes master, [10] retrying in 57 seconds...
2020-03-09 18:24:23.099 ERROR 1 --- [           main] c.h.i.c.i.DiscoveryJoiner                : [a.b.c.d]:5701 [dev] [3.12] Failure in KubernetesClient
com.hazelcast.kubernetes.KubernetesClientException: Failure in KubernetesClient
        at com.hazelcast.kubernetes.DefaultKubernetesClient.callGet(DefaultKubernetesClient.java:109) ~[hazelcast-kubernetes-1.3.1.jar!/:?]
        at com.hazelcast.kubernetes.DefaultKubernetesClient.endpointsByName(DefaultKubernetesClient.java:88) ~[hazelcast-kubernetes-1.3.1.jar!/:?]
        .
        .
        .       
        at org.springframework.boot.loader.JarLauncher.main(JarLauncher.java:51) [main.jar:?]
Caused by: java.net.UnknownHostException: kubernetes.default.svc
        at java.net.AbstractPlainSocketImpl.connect(AbstractPlainSocketImpl.java:184) ~[?:1.8.0_111]
        at java.net.SocksSocketImpl.connect(SocksSocketImpl.java:392) ~[?:1.8.0_111]

在第二种情况下，我使用的是 leszko 的 sample，它尝试连接的 IP 地址是 POD 的 IP。

[190.168.1.194]:5701 [dev] [3.11] Failure in KubernetesClient

190.168.1.194是POD的IP地址

我将下面的环境添加到部署 yaml 中，但看起来它并没有覆盖这个变量。

env:
           - name: service-name
             value: "hazelcast-embedded"

Kubernetes Discovery properties: { service-dns: null, service-dns-timeout: 5, service-name: null, service-port: 0, service-label: null, service-label-value: true, namespace: default, resolve-not-ready-addresses: false, kubernetes-master: https://kubernetes.default.svc}

Answer 1

您的 Kubernetes 集群本身似乎有问题。

首先您尝试使用 Hazelcast Kubernetes 发现策略 Kubernetes API，它对 Kubernetes master 进行 REST 调用。 Kubernetes master 应该始终可以通过主机名 kubernetes.default.svc 从您的 POD 访问。在你的情况下，它没有找到。
然后，您尝试使用 Hazelcast Kubernetes 发现策略 DNS Lookup，它使用服务 DNS 来解析 Hazelcast 成员。在您的情况下，正如@Mesut 所提到的，它没有正确解析服务的 DNS。

hazelcast-kubernetes 无法启动嵌入式服务器

hazelcast-kubernetes Unable to start Embedded server

hazelcast

spring-boot

kubernetes