如何从集群外部访问我的Cassandra / Kubernetes集群？

• 由于您是Kubernetes的新手，因此您可能不熟悉StatefulSets：

StatefulSet是用于管理有状态应用程序的工作负载API对象。

管理一组的部署和缩放窗格，并提供有关的排序和唯一性保证这些荚。

Like a Deployment, a StatefulSet manages Pods that are based on an identical container spec. Unlike a Deployment, a StatefulSet maintains a sticky identity for each of their Pods. These pods are created from the same spec, but are not interchangeable: each has a persistent identifier that it maintains across any rescheduling.

• I recommend you to read these articles to learn more about it's mechanisms:
  • Kubernetes.io - Statefulsets
  • Megalix - Statefulsets 101
  • ITNext - Exposing Statefulsets in Kubernetes

How do I go about connecting the K8s/Cassandra cluster to outside work so that my web application can access it?

• I found out that datastax/cass-operator is still developing their documentation, I found this document that is not merged to master yet, but it explains very well about how to connect to Cassandra, I strongly recommend reading.
• There are several open issues for documenting methods for connection from outside the cluster.

I followed the guide in https://github.com/datastax/cass-operator to deploy the cass-operator + Cassandra Datacenter Example as from your images I believe you followed as well:

$ kubectl create -f https://raw.githubusercontent.com/datastax/cass-operator/v1.2.0/docs/user/cass-operator-manifests-v1.15.yaml
namespace/cass-operator created
serviceaccount/cass-operator created
secret/cass-operator-webhook-config created
customresourcedefinition.apiextensions.k8s.io/cassandradatacenters.cassandra.datastax.com created
clusterrole.rbac.authorization.k8s.io/cass-operator-cluster-role created
clusterrolebinding.rbac.authorization.k8s.io/cass-operator created
role.rbac.authorization.k8s.io/cass-operator created
rolebinding.rbac.authorization.k8s.io/cass-operator created
service/cassandradatacenter-webhook-service created
deployment.apps/cass-operator created
validatingwebhookconfiguration.admissionregistration.k8s.io/cassandradatacenter-webhook-registration created

$ kubectl create -f https://raw.githubusercontent.com/datastax/cass-operator/v1.2.0/operator/k8s-flavors/gke/storage.yaml
storageclass.storage.k8s.io/server-storage created

$ kubectl -n cass-operator create -f https://raw.githubusercontent.com/datastax/cass-operator/v1.2.0/operator/example-cassdc-yaml/cassandra-3.11.6/example-cassdc-minimal.yaml
cassandradatacenter.cassandra.datastax.com/dc1 created

$ kubectl get all -n cass-operator
NAME                                READY   STATUS    RESTARTS   AGE
pod/cass-operator-78c6469c6-6qhsb   1/1     Running   0          139m
pod/cluster1-dc1-default-sts-0      2/2     Running   0          138m
pod/cluster1-dc1-default-sts-1      2/2     Running   0          138m
pod/cluster1-dc1-default-sts-2      2/2     Running   0          138m

NAME                                          TYPE           CLUSTER-IP    EXTERNAL-IP    PORT(S)             AGE
service/cass-operator-metrics                 ClusterIP      10.21.5.65    <none>         8383/TCP,8686/TCP   138m
service/cassandradatacenter-webhook-service   ClusterIP      10.21.0.89    <none>         443/TCP             139m
service/cluster1-dc1-all-pods-service         ClusterIP      None          <none>         <none>              138m
service/cluster1-dc1-service                  ClusterIP      None          <none>         9042/TCP,8080/TCP   138m
service/cluster1-seed-service                 ClusterIP      None          <none>         <none>              138m

NAME                            READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/cass-operator   1/1     1            1           139m

NAME                                      DESIRED   CURRENT   READY   AGE
replicaset.apps/cass-operator-78c6469c6   1         1         1       139m

NAME                                        READY   AGE
statefulset.apps/cluster1-dc1-default-sts   3/3     138m

$ CASS_USER=$(kubectl -n cass-operator get secret cluster1-superuser -o json | jq -r '.data.username' | base64 --decode)
$ CASS_PASS=$(kubectl -n cass-operator get secret cluster1-superuser -o json | jq -r '.data.password' | base64 --decode)

$ echo $CASS_USER
cluster1-superuser

$ echo $CASS_PASS
_5ROwp851l0E_2CGuN_n753E-zvEmo5oy31i6C0DBcyIwH5vFjB8_g

• From the kubectl get all command above we can see there is an statefulset called statefulset.apps/cluster1-dc1-default-sts which controls the cassandra pods.
• In order to create a LoadBalancer service that makes available all the pods managed by this statefulset we need to use the same labels assigned to them:

$ kubectl describe statefulset cluster1-dc1-default-sts -n cass-operator
Name:               cluster1-dc1-default-sts
Namespace:          cass-operator
CreationTimestamp:  Tue, 30 Jun 2020 12:24:34 +0200
Selector:           cassandra.datastax.com/cluster=cluster1,cassandra.datastax.com/datacenter=dc1,cassandra.datastax.com/rack=default
Labels:             app.kubernetes.io/managed-by=cass-operator
                    cassandra.datastax.com/cluster=cluster1
                    cassandra.datastax.com/datacenter=dc1
                    cassandra.datastax.com/rack=default

• Now let's create the LoadBalancer service yaml and use the labels as selectors for the service:

apiVersion: v1
kind: Service
metadata:
  name: cassandra-loadbalancer
  namespace: cass-operator
  labels:
    cassandra.datastax.com/cluster: cluster1
    cassandra.datastax.com/datacenter: dc1
    cassandra.datastax.com/rack: default
spec:
  type: LoadBalancer
  ports:
  - port: 9042
    protocol: TCP
  selector:
    cassandra.datastax.com/cluster: cluster1
    cassandra.datastax.com/datacenter: dc1
    cassandra.datastax.com/rack: default

"My web application should be able to reach Cassandra on 9042. It seems load balancing is done for http/https. The Cassandra application is not a http/https request. So I don't need port 80 or 443."

• When you create a Service of type LoadBalancer, a Google Cloud controller wakes up and configures a network load balancer in your project. The load balancer has a stable IP address that is accessible from outside of your project.

• The network load balancer supports any and all ports. You can use Network Load Balancing to load balance TCP and UDP traffic. Because the load balancer is a pass-through load balancer, your backends terminate the load-balanced TCP connection or UDP packets themselves.

• Now let's apply the yaml and note the Endpoint IPs of the pods being listed:

$ kubectl apply -f cassandra-loadbalancer.yaml 
service/cassandra-loadbalancer created

$ kubectl get service cassandra-loadbalancer -n cass-operator 
NAME                     TYPE           CLUSTER-IP    EXTERNAL-IP    PORT(S)          AGE
cassandra-loadbalancer   LoadBalancer   10.21.4.253   146.148.89.7   9042:30786/TCP   5m13s

$ kubectl describe svc cassandra-loadbalancer -n cass-operator
Name:                     cassandra-loadbalancer
Namespace:                cass-operator
Labels:                   cassandra.datastax.com/cluster=cluster1
                          cassandra.datastax.com/datacenter=dc1
                          cassandra.datastax.com/rack=default
Annotations:              Selector:  cassandra.datastax.com/cluster=cluster1,cassandra.datastax.com/datacenter=dc1,cassandra.datastax.com/rack=default
Type:                     LoadBalancer
IP:                       10.21.4.253
LoadBalancer Ingress:     146.148.89.7
Port:                     <unset>  9042/TCP
TargetPort:               9042/TCP
NodePort:                 <unset>  30786/TCP
Endpoints:                10.24.0.7:9042,10.24.2.7:9042,10.24.3.9:9042
Session Affinity:         None
External Traffic Policy:  Cluster
Events:                   <none>

• To test it, I'll use my cloud shell with a cassandra container to emulate your notebook using the LoadBalancer IP provided above:

$ docker run -it cassandra /bin/sh

# cqlsh -u cluster1-superuser -p _5ROwp851l0E_2CGuN_n753E-zvEmo5oy31i6C0DBcyIwH5vFjB8_g 146.148.89.7 9042                

Connected to cluster1 at 146.148.89.7:9042.
[cqlsh 5.0.1 | Cassandra 3.11.6 | CQL spec 3.4.4 | Native protocol v4]
Use HELP for help.
cluster1-superuser@cqlsh> select * from system.peers;

 peer      | data_center | host_id                              | preferred_ip | rack    | release_version | rpc_address | schema_version                       | tokens
-----------+-------------+--------------------------------------+--------------+---------+-----------------+-------------+--------------------------------------+--------------------------
 10.24.3.9 |         dc1 | bcec6c12-49a1-41d5-be58-5150e99f5dfb |         null | default |          3.11.6 |   10.24.3.9 | e84b6a60-24cf-30ca-9b58-452d92911703 |  {'2248175870989649036'}
 10.24.0.7 |         dc1 | 68409f08-9d6e-4e40-91ff-f43581c8b6f3 |         null | default |          3.11.6 |   10.24.0.7 | e84b6a60-24cf-30ca-9b58-452d92911703 | {'-1105923522927946373'}

(2 rows)

"have a url so that my web application uses that URL to connect to the cassandra/K8s cluster instead of IP address. So I need a dns. Does it come by default in K8S? Would would be the url? Would K8s managing the dns mapping for me in some nodes get restarted?"

• 关于cassandra-operator的文档也有一个关于Ingress的部分，我也建议阅读。
• Kubernetes没有默认的DNS名称。
• 您将必须注册一个域，将DNS指向负载均衡器的IP，这样它将解析网络LoadBalancer的IP。
• 网络负载平衡器已绑定到静态公共IP，Kubernetes节点中的任何更改都不会导致服务不可用。

如有任何疑问，请在评论中告诉我。