2015-12-14注：加入新节点不更改运行节点参数需求已满足，将在后续文章中陆续总结。
注：目前方案不满足加入新节点（master节点或regionserver节点）而不更改已运行节点的参数的需求，具体讨论见第六部分。
一、背景知识
先看下HBase的组成：
Master：Master主要负责管理RegionServer集群，如负载均衡及资源分配等，它本身也可以以集群方式运行，但同一时刻只有一个master处于激活状态。当工作中的master宕掉后，zookeeper会切换到其它备选的master上。
RegionServer：负责具体数据块的读写操作。
ZooKeeper：负责集群元数据的维护并监控集群的状态以防止单点故障。部署HBase时可以使用自带的ZooKeeper也可以使用独立的集群，是HBase跑起来的先决条件。
HDFS：写入HBase中的数据最终都持久化到了HDFS中，也是HBase运行的先决条件。
二、skyDNS部署
skyDNS并不是必需项，但设置了skyDNS后可以为k8s的service绑定域名，进行hbase的参数设置时可以以域名代替service的IP地址。k8s环境中的skyDNS由三部分组成： 存储IP地址和域名映射关系的ETCD；进行域名解析的skyDNS；连接k8s和skyDNS的桥梁kube2sky。k8s的域名构成为 service_name.namespace.k8s_cluster_domain。k8s的文档中有对部署skyDNS的简略说明（戳这里），其中要用到Google镜像仓库中的image，国内访问不到，可以使用DockerHub上的替换方案，如：
skyDNS: docker pull shenshouer/skydns:2015-09-22
kube2sky: docker pull shenshouer/kube2sky:1.11
ETCD: 只要是2.x版本的ETCD都可以，也可以和上面的保持一致使用 docker pull shenshouer/etcd:2.0.9
pull下来之后打上tag再push到私有仓库中。
下面创建一个service和一个pod来部署skyDNS，设定 skyDNS service 的服务地址为 172.16.40.1 (53/UDP, 53/TCP)，注意该IP地址要在kube-apiserver启动时设定的service的子网范围内；k8s集群的域名后缀为 domeos.sohu，注意这个后缀的选择最好包含两部分，否则kube2sky可能会出问题（具体讨论戳这里）。
首先创建skydns.yaml文件：
apiVersion: v1
kind: Service
metadata:
name: kube-dns
labels:
app: kube-dns
version: v8
spec:
selector:
app: kube-dns
version: v8
type: ClusterIP
clusterIP: 172.16.40.1
ports:
- name: dns
port: 53
protocol: UDP
- name: dns-tcp
port: 53
protocol: TCP
---
apiVersion: v1
kind: ReplicationController
metadata:
name: kube-dns-v8
labels:
app: kube-dns
version: v8
spec:
replicas: 1
selector:
app: kube-dns
version: v8
template:
metadata:
labels:
app: kube-dns
version: v8
spec:
containers:
- name: etcd
image: 10.11.150.76:5000/openxxs/etcd:2.0.3
command:
- "etcd"
args:
- "--data-dir=/var/etcd/data"
- "--listen-client-urls=http://127.0.0.1:2379,http://127.0.0.1:4001"
- "--advertise-client-urls=http://127.0.0.1:2379,http://127.0.0.1:4001"
- "--initial-cluster-token=skydns-etcd"
volumeMounts:
- name: etcd-storage
mountPath: /var/etcd/data
- name: kube2sky
image: 10.11.150.76:5000/openxxs/kube2sky:k8s-dns
args:
- "--domain=domeos.sohu"
- "--kube_master_url=http://10.16.42.200:8080"
- name: skydns
image: 10.11.150.76:5000/openxxs/skydns:2015-09-22
args:
- "--machines=http://localhost:4001"
- "--addr=0.0.0.0:53"
- "--domain=domeos.sohu"
ports:
- containerPort: 53
name: dns
protocol: UDP
- containerPort: 53
name: dns-tcp
protocol: TCP
volumes:
- name: etcd-storage
emptyDir: {}
dnsPolicy: Default
kube2sky中的 --kube_master_url 参数用于指定 kube-apiserver 的地址；kube2sky中的 --domain 和 skydns中的 --domain 要保持一致。
然后 kubectl create -f skydns.yaml 创建服务和pod：
$kubectl create -f skydns.yaml
service "kube-dns" created
replicationcontroller "kube-dns-v8" created
$kubectl get pods
NAME READY STATUS RESTARTS AGE
kube-dns-v8-61aie 3/3 Running 0 9s
$kubectl get service
NAME CLUSTER_IP EXTERNAL_IP PORT(S) SELECTOR AGE
kube-dns 172.16.40.1 <none> 53/UDP,53/TCP app=kube-dns,version=v8 6m
最后，重启 kubelet 加上dns相关的设置参数 --cluster_dns 和 --cluster_domain，参数值要与前面yaml文件中写的一致，如：
./kubelet --logtostderr=true --v=0 --api_servers=http://bx-42-200:8080 --address=0.0.0.0 --hostname_override=bx-42-198 --allow_privileged=false --pod-infra-container-image=10.11.150.76:5000/kubernetes/pause:latest --cluster_dns=172.16.40.1 --cluster_domain=domeos.sohu &
注意：只有在kubelet加了dns设置参数重启之后创建的pods才会使用skyDNS。
此时进入到etcd的container中就可以发现k8s的service域名信息已被写入etcd当中了：
$ docker exec -it 13e243510e3e sh / # etcdctl ls --recursive / /skydns /skydns/sohu /skydns/sohu/domeos /skydns/sohu/domeos/default /skydns/sohu/domeos/default/kube-dns /skydns/sohu/domeos/default/kubernetes /skydns/sohu/domeos/default/zookeeper-1 /skydns/sohu/domeos/default/zookeeper-2 /skydns/sohu/domeos/default/zookeeper-3 /skydns/sohu/domeos/svc /skydns/sohu/domeos/svc/default /skydns/sohu/domeos/svc/default/zookeeper-2 /skydns/sohu/domeos/svc/default/zookeeper-2/b8757496 /skydns/sohu/domeos/svc/default/zookeeper-3 /skydns/sohu/domeos/svc/default/zookeeper-3/8687b21f /skydns/sohu/domeos/svc/default/kube-dns /skydns/sohu/domeos/svc/default/kube-dns/a9f11e6f /skydns/sohu/domeos/svc/default/kubernetes /skydns/sohu/domeos/svc/default/kubernetes/cf07aead /skydns/sohu/domeos/svc/default/zookeeper-1 /skydns/sohu/domeos/svc/default/zookeeper-1/75512011 / # etcdctl get /skydns/sohu/domeos/default/zookeeper-1 {"host":"172.16.11.1","priority":10,"weight":10,"ttl":30,"targetstrip":0}
以 /skydns/sohu/domeos/default/zookeeper-1 这条记录为例，其对应的域名即为 zookeeper-1.default.domeos.sohu ，IP 为 172.16.11.1，服务名称为zookeeper-1，k8s的namespace为default，k8s设定的域为 domeos.sohu。在任一重启kubelet之后创建的pod中都可以以 zookeeper-1.default.domeos.sohu 的方式访问zookeeper-1服务，如：
[@bx_42_199 ~]# docker exec -it 0662660e8708 /bin/bash [root@test-3-2h0fx /]# curl zookeeper-1.default.domeos.sohu:2181 curl: (52) Empty reply from server
三、HDFS集群部署
HDFS由namenode和datanode组成，首先从DockerHub上pull合适的镜像再push到自己的私有仓库中：
# pull 远程images docker pull bioshrek/hadoop-hdfs-datanode:cdh5 docker pull bioshrek/hadoop-hdfs-namenode:cdh5 # 打上tag # docker tag <image的ID> <自己的私有仓库IP:PORT/名称:TAG> docker tag c89c3ebcccae 10.11.150.76:5000/hdfs-datanode:latest docker tag ca19d4c7e359 10.11.150.76:5000/hdfs-namenode:latest # push到仓库中 docker push 10.11.150.76:5000/hdfs-datanode:latest docker push 10.11.150.76:5000/hdfs-namenode:latest
然后创建如下hdfs.yaml文件：
1 apiVersion: v1 2 kind: Service 3 metadata: 4 name: hdfs-namenode-service 5 spec: 6 selector: 7 app: hdfs-namenode 8 type: ClusterIP 9 clusterIP: "172.16.20.1" 10 ports: 11 - name: rpc 12 port: 4231 13 targetPort: 8020 14 - name: p1 15 port: 50020 16 - name: p2 17 port: 50090 18 - name: p3 19 port: 50070 20 - name: p4 21 port: 50010 22 - name: p5 23 port: 50075 24 - name: p6 25 port: 8031 26 - name: p7 27 port: 8032 28 - name: p8 29 port: 8033 30 - name: p9 31 port: 8040 32 - name: p10 33 port: 8042 34 - name: p11 35 port: 49707 36 - name: p12 37 port: 22 38 - name: p13 39 port: 8088 40 - name: p14 41 port: 8030 42 --- 43 apiVersion: v1 44 kind: ReplicationController 45 metadata: 46 name: hdfs-namenode-1 47 spec: 48 replicas: 1 49 template: 50 metadata: 51 labels: 52 app: hdfs-namenode 53 spec: 54 containers: 55 - name: hdfs-namenode 56 image: 10.11.150.76:5000/hdfs-namenode:latest 57 volumeMounts: 58 - name: data1 59 mountPath: /var/lib/hadoop-hdfs/cache/hdfs/dfs/name 60 - name: data2 61 mountPath: /home/chianyu/shared_with_docker_container/cdh5/nn 62 ports: 63 - containerPort: 50020 64 - containerPort: 50090 65 - containerPort: 50070 66 - containerPort: 50010 67 - containerPort: 50075 68 - containerPort: 8031 69 - containerPort: 8032 70 - containerPort: 8033 71 - containerPort: 8040 72 - containerPort: 8042 73 - containerPort: 49707 74 - containerPort: 22 75 - containerPort: 8088 76 - containerPort: 8030 77 - containerPort: 8020 78 nodeSelector: 79 kubernetes.io/hostname: bx-42-199 80 volumes: 81 - hostPath: 82 path: /data1/kubernetes/hdfs-namenode/data1 83 name: data1 84 - hostPath: 85 path: /data1/kubernetes/hdfs-namenode/data2 86 name: data2 87 --- 88 apiVersion: v1 89 kind: ReplicationController 90 metadata: 91 name: hdfs-datanode-1 92 spec: 93 replicas: 1 94 template: 95 metadata: 96 labels: 97 app: hdfs-datanode 98 server-id: "1" 99 spec: 100 containers: 101 - name: hdfs-datanode-1 102 image: 10.11.150.76:5000/hdfs-datanode:latest 103 volumeMounts: 104 - name: data1 105 mountPath: /var/lib/hadoop-hdfs/cache/hdfs/dfs/name 106 - name: data2 107 mountPath: /home/chianyu/shared_with_docker_container/cdh5/dn 108 env: 109 - name: HDFSNAMENODERPC_SERVICE_HOST 110 value: "172.16.20.1" 111 - name: HDFSNAMENODERPC_SERVICE_PORT 112 value: "4231" 113 ports: 114 - containerPort: 50020 115 - containerPort: 50090 116 - containerPort: 50070 117 - containerPort: 50010 118 - containerPort: 50075 119 - containerPort: 8031 120 - containerPort: 8032 121 - containerPort: 8033 122 - containerPort: 8040 123 - containerPort: 8042 124 - containerPort: 49707 125 - containerPort: 22 126 - containerPort: 8088 127 - containerPort: 8030 128 - containerPort: 8020 129 nodeSelector: 130 kubernetes.io/hostname: bx-42-199 131 volumes: 132 - hostPath: 133 path: /data1/kubernetes/hdfs-datanode1/data1 134 name: data1 135 - hostPath: 136 path: /data1/kubernetes/hdfs-datanode1/data2 137 name: data2 138 --- 139 apiVersion: v1 140 kind: ReplicationController 141 metadata: 142 name: hdfs-datanode-2 143 spec: 144 replicas: 1 145 template: 146 metadata: 147 labels: 148 app: hdfs-datanode 149 server-id: "2" 150 spec: 151 containers: 152 - name: hdfs-datanode-2 153 image: 10.11.150.76:5000/hdfs-datanode:latest 154 volumeMounts: 155 - name: data1 156 mountPath: /var/lib/hadoop-hdfs/cache/hdfs/dfs/name 157 - name: data2 158 mountPath: /home/chianyu/shared_with_docker_container/cdh5/dn 159 env: 160 - name: HDFSNAMENODERPC_SERVICE_HOST 161 value: "172.16.20.1" 162 - name: HDFSNAMENODERPC_SERVICE_PORT 163 value: "4231" 164 ports: 165 - containerPort: 50020 166 - containerPort: 50090 167 - containerPort: 50070 168 - containerPort: 50010 169 - containerPort: 50075 170 - containerPort: 8031 171 - containerPort: 8032 172 - containerPort: 8033 173 - containerPort: 8040 174 - containerPort: 8042 175 - containerPort: 49707 176 - containerPort: 22 177 - containerPort: 8088 178 - containerPort: 8030 179 nodeSelector: 180 kubernetes.io/hostname: bx-42-199 181 volumes: 182 - name: data1 183 hostPath: 184 path: /data2/kubernetes/hdfs-datanode2/data1 185 - name: data2 186 hostPath: 187 path: /data2/kubernetes/hdfs-datanode2/data2 188 --- 189 apiVersion: v1 190 kind: ReplicationController 191 metadata: 192 name: hdfs-datanode-3 193 spec: 194 replicas: 1 195 template: 196 metadata: 197 labels: 198 app: hdfs-datanode 199 server-id: "3" 200 spec: 201 containers: 202 - name: hdfs-datanode-3 203 image: 10.11.150.76:5000/hdfs-datanode:latest 204 volumeMounts: 205 - name: data1 206 mountPath: /var/lib/hadoop-hdfs/cache/hdfs/dfs/name 207 - name: data2 208 mountPath: /home/chianyu/shared_with_docker_container/cdh5/dn 209 env: 210 - name: HDFSNAMENODERPC_SERVICE_HOST 211 value: "172.16.20.1" 212 - name: HDFSNAMENODERPC_SERVICE_PORT 213 value: "4231" 214 ports: 215 - containerPort: 50020 216 - containerPort: 50090 217 - containerPort: 50070 218 - containerPort: 50010 219 - containerPort: 50075 220 - containerPort: 8031 221 - containerPort: 8032 222 - containerPort: 8033 223 - containerPort: 8040 224 - containerPort: 8042 225 - containerPort: 49707 226 - containerPort: 22 227 - containerPort: 8088 228 - containerPort: 8030 229 nodeSelector: 230 kubernetes.io/hostname: bx-42-199 231 volumes: 232 - name: data1 233 hostPath: 234 path: /data3/kubernetes/hdfs-datanode3/data1 235 - name: data2 236 hostPath: 237 path: /data3/kubernetes/hdfs-datanode3/data2
通过 kubectl create -f hdfs.yaml 即创建一个名为hdfs-namenode-service的service，四个分别名为hdfs-namenode-1、hdfs-datanode-1、hdfs-datanode-2、hdfs-datanode-3的RC。通过 kubectl get services/rc/pods 可以看到对应的service和pod都已经正常启动了。
下面对HDFS进行测试是否可以正常使用：
# 查看HDFS pods kubectl get pods # 通过describe查看pods跑在哪个k8s node上 kubectl describe pod hdfs-datanode-3-h4jvt # 进入容器内部 docker ps | grep hdfs-datanode-3 docker exec -it 2e2c4df0c0a9 /bin/bash # 切换至 hdfs 用户 su hdfs # 创建目录 hadoop fs -mkdir /test # 创建本地文件 echo "Hello" > hello # 将本地文件复制到HDFS文件系统中 hadoop fs -put hello /test # 查看HDFS中的文件信息 hadoop fs -ls /test # 类似的，可以 docker exec 到其它datanode中查看文件信息，如： root@hdfs-datanode-1-nek2l:/# hadoop fs -ls /test Found 1 items -rw-r--r-- 2 hdfs hadoop 6 2015-11-27 08:36 /test/hello
四、ZooKeeper集群部署
在 fabric8/zookeeper 的image基础上进行修改，修改后Dockerfile文件内容如下：
1 FROM jboss/base-jdk:7 2 3 MAINTAINER iocanel@gmail.com 4 5 USER root 6 7 ENV ZOOKEEPER_VERSION 3.4.6 8 EXPOSE 2181 2888 3888 9 10 RUN yum -y install wget bind-utils && yum clean all 11 && wget -q -O - http://apache.mirrors.pair.com/zookeeper/zookeeper-${ZOOKEEPER_VERSION}/zookeeper-${ZOOKEEPER_VERSION}.tar.gz | tar -xzf - -C /opt 12 && mv /opt/zookeeper-${ZOOKEEPER_VERSION} /opt/zookeeper 13 && cp /opt/zookeeper/conf/zoo_sample.cfg /opt/zookeeper/conf/zoo.cfg 14 && mkdir -p /opt/zookeeper/{data,log} 15 16 WORKDIR /opt/zookeeper 17 VOLUME ["/opt/zookeeper/conf", "/opt/zookeeper/data", "/opt/zookeeper/log"] 18 19 COPY config-and-run.sh ./bin/ 20 COPY zoo.cfg ./conf/ 21 22 CMD ["/opt/zookeeper/bin/config-and-run.sh"]
zoo.cfg 文件内容如下：
1 # The number of milliseconds of each tick 2 tickTime=2000 3 # The number of ticks that the initial 4 # synchronization phase can take 5 initLimit=10 6 # The number of ticks that can pass between 7 # sending a request and getting an acknowledgement 8 syncLimit=5 9 # the directory where the snapshot is stored. 10 dataDir=/opt/zookeeper/data 11 #This option will direct the machine to write the transaction log to the dataLogDir rather than the dataDir. This allows a dedicated log device to be used, and helps avoid competition between logging and snaphots. 12 dataLogDir=/opt/zookeeper/log 13 14 # the port at which the clients will connect 15 clientPort=2181 16 # 17 # Be sure to read the maintenance section of the 18 # administrator guide before turning on autopurge. 19 # 20 # http://zookeeper.apache.org/doc/current/zookeeperAdmin.html#sc_maintenance 21 # 22 # The number of snapshots to retain in dataDir 23 #autopurge.snapRetainCount=3 24 # Purge task interval in hours 25 # Set to "0" to disable auto purge feature 26 #autopurge.purgeInterval=1
config-and-run.sh 文件内容如下：
#!/bin/bash echo "$SERVER_ID / $MAX_SERVERS" if [ ! -z "$SERVER_ID" ] && [ ! -z "$MAX_SERVERS" ]; then echo "Starting up in clustered mode" echo "" >> /opt/zookeeper/conf/zoo.cfg echo "#Server List" >> /opt/zookeeper/conf/zoo.cfg for i in $( eval echo {1..$MAX_SERVERS});do HostEnv="ZOOKEEPER_${i}_SERVICE_HOST" HOST=${!HostEnv} FollowerPortEnv="ZOOKEEPER_${i}_SERVICE_PORT_FOLLOWERS" FOLLOWERPORT=${!FollowerPortEnv} ElectionPortEnv="ZOOKEEPER_${i}_SERVICE_PORT_ELECTION" ELECTIONPORT=${!ElectionPortEnv} if [ "$SERVER_ID" = "$i" ];then echo "server.$i=0.0.0.0:$FOLLOWERPORT:$ELECTIONPORT" >> /opt/zookeeper/conf/zoo.cfg else echo "server.$i=$HOST:$FOLLOWERPORT:$ELECTIONPORT" >> /opt/zookeeper/conf/zoo.cfg fi done cat /opt/zookeeper/conf/zoo.cfg # Persists the ID of the current instance of Zookeeper echo ${SERVER_ID} > /opt/zookeeper/data/myid else echo "Starting up in standalone mode" fi exec /opt/zookeeper/bin/zkServer.sh start-foreground
修改完后创建镜像并push到私有仓库中(镜像名为10.11.150.76:5000/zookeeper-kb:3.4.6-1)。
创建zookeeper.yaml文件：
1 apiVersion: v1 2 kind: Service 3 metadata: 4 name: zookeeper-1 5 labels: 6 name: zookeeper-1 7 spec: 8 ports: 9 - name: client 10 port: 2181 11 targetPort: 2181 12 - name: followers 13 port: 2888 14 targetPort: 2888 15 - name: election 16 port: 3888 17 targetPort: 3888 18 selector: 19 name: zookeeper 20 server-id: "1" 21 type: ClusterIP 22 clusterIP: 172.16.11.1 23 --- 24 apiVersion: v1 25 kind: Service 26 metadata: 27 name: zookeeper-2 28 labels: 29 name: zookeeper-2 30 spec: 31 ports: 32 - name: client 33 port: 2181 34 targetPort: 2181 35 - name: followers 36 port: 2888 37 targetPort: 2888 38 - name: election 39 port: 3888 40 targetPort: 3888 41 selector: 42 name: zookeeper 43 server-id: "2" 44 type: ClusterIP 45 clusterIP: 172.16.11.2 46 --- 47 apiVersion: v1 48 kind: Service 49 metadata: 50 name: zookeeper-3 51 labels: 52 name: zookeeper-3 53 spec: 54 ports: 55 - name: client 56 port: 2181 57 targetPort: 2181 58 - name: followers 59 port: 2888 60 targetPort: 2888 61 - name: election 62 port: 3888 63 targetPort: 3888 64 selector: 65 name: zookeeper 66 server-id: "3" 67 type: ClusterIP 68 clusterIP: 172.16.11.3 69 --- 70 apiVersion: v1 71 kind: ReplicationController 72 metadata: 73 name: zookeeper-1 74 spec: 75 replicas: 1 76 template: 77 metadata: 78 labels: 79 name: zookeeper 80 server-id: "1" 81 spec: 82 volumes: 83 - hostPath: 84 path: /data1/kubernetes/zookeeper/data1 85 name: data 86 - hostPath: 87 path: /data1/kubernetes/zookeeper/log1 88 name: log 89 containers: 90 - name: server 91 image: 10.11.150.76:5000/zookeeper-kb:3.4.6-1 92 env: 93 - name: SERVER_ID 94 value: "1" 95 - name: MAX_SERVERS 96 value: "3" 97 ports: 98 - containerPort: 2181 99 - containerPort: 2888 100 - containerPort: 3888 101 volumeMounts: 102 - mountPath: /opt/zookeeper/data 103 name: data 104 - mountPath: /opt/zookeeper/log 105 name: log 106 nodeSelector: 107 kubernetes.io/hostname: bx-42-199 108 --- 109 apiVersion: v1 110 kind: ReplicationController 111 metadata: 112 name: zookeeper-2 113 spec: 114 replicas: 1 115 template: 116 metadata: 117 labels: 118 name: zookeeper 119 server-id: "2" 120 spec: 121 volumes: 122 - hostPath: 123 path: /data1/kubernetes/zookeeper/data2 124 name: data 125 - hostPath: 126 path: /data1/kubernetes/zookeeper/log2 127 name: log 128 containers: 129 - name: server 130 image: 10.11.150.76:5000/zookeeper-kb:3.4.6-1 131 env: 132 - name: SERVER_ID 133 value: "2" 134 - name: MAX_SERVERS 135 value: "3" 136 ports: 137 - containerPort: 2181 138 - containerPort: 2888 139 - containerPort: 3888 140 volumeMounts: 141 - mountPath: /opt/zookeeper/data 142 name: data 143 - mountPath: /opt/zookeeper/log 144 name: log 145 nodeSelector: 146 kubernetes.io/hostname: bx-42-199 147 --- 148 apiVersion: v1 149 kind: ReplicationController 150 metadata: 151 name: zookeeper-3 152 spec: 153 replicas: 1 154 template: 155 metadata: 156 labels: 157 name: zookeeper 158 server-id: "3" 159 spec: 160 volumes: 161 - hostPath: 162 path: /data1/kubernetes/zookeeper/data3 163 name: data 164 - hostPath: 165 path: /data1/kubernetes/zookeeper/log3 166 name: log 167 containers: 168 - name: server 169 image: 10.11.150.76:5000/zookeeper-kb:3.4.6-1 170 env: 171 - name: SERVER_ID 172 value: "3" 173 - name: MAX_SERVERS 174 value: "3" 175 ports: 176 - containerPort: 2181 177 - containerPort: 2888 178 - containerPort: 3888 179 volumeMounts: 180 - mountPath: /opt/zookeeper/data 181 name: data 182 - mountPath: /opt/zookeeper/log 183 name: log 184 nodeSelector: 185 kubernetes.io/hostname: bx-42-199
通过 kubectl create -f zookeeper.yaml 创建三个service和对应的RC。注意container中已经把ZooKeeper的data和log目录映射到了主机的对应目录上用于持久化存储。
创建完之后即可进行测试：
# 进入zookeeper对应的容器后找到zkCli.sh，用该客户端进行测试 /opt/zookeeper/bin/zkCli.sh [zk: localhost:2181(CONNECTED) 0] # 连接到k8s创建的zookeeper service (三个service任意一个都行) [zk: localhost:2181(CONNECTED) 0] connect 172.16.11.2:2181 [zk: 172.16.11.2:2181(CONNECTED) 1] # 查看目录信息 [zk: 172.16.11.2:2181(CONNECTED) 1] ls / [zookeeper] [zk: 172.16.11.2:2181(CONNECTED) 2] get /zookeeper cZxid = 0x0 ctime = Thu Jan 01 00:00:00 UTC 1970 mZxid = 0x0 mtime = Thu Jan 01 00:00:00 UTC 1970 pZxid = 0x0 cversion = -1 dataVersion = 0 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 0 numChildren = 1 [zk: 172.16.11.2:2181(CONNECTED) 3]
五、HBase部署
以上准备工作做好后，下面部署具有两个master和两个regionserver的HBase集群，其中两个master分别位于两个节点上，两个regionserver也分别位于两个节点上；使用独立的HDFS和ZooKeeper服务。
首先需要创建HBase的镜像，选择的HBase版本为hbase-0.98.10.1-hadoop2。Dockerfile内容如下：
1 FROM centos:6.6 2 MAINTAINER openxxs <xiaoshengxu@sohu-inc.com> 3 4 RUN yum install -y java-1.7.0-openjdk-devel.x86_64 5 ENV JAVA_HOME=/usr/lib/jvm/jre 6 7 RUN yum install -y nc 8 && yum install -y tar 9 && mkdir /hbase-setup 10 11 WORKDIR /hbase-setup 12 13 COPY hbase-0.98.10.1-hadoop2-bin.tar.gz /hbase-setup/hbase-0.98.10.1-hadoop2-bin.tar.gz 14 RUN tar zxf hbase-0.98.10.1-hadoop2-bin.tar.gz -C /opt/ 15 && ln -s /opt/hbase-0.98.10.1-hadoop2 /opt/hbase 16 17 ADD hbase-site.xml /opt/hbase/conf/hbase-site.xml 18 ADD start-k8s-hbase.sh /opt/hbase/bin/start-k8s-hbase.sh 19 RUN chmod +x /opt/hbase/bin/start-k8s-hbase.sh 20 21 WORKDIR /opt/hbase/bin 22 23 ENV PATH=$PATH:/opt/hbase/bin 24 25 CMD /opt/hbase/bin/start-k8s-hbase.sh
配置文件hbase-site.xml内容如下：
<?xml version="1.0"?> <?xml-stylesheet type="text/xsl" href="configuration.xsl"?> <configuration> <property> <name>hbase.cluster.distributed</name> <value>true</value> </property> <property> <name>hbase.master.port</name> <value>@HBASE_MASTER_PORT@</value> </property> <property> <name>hbase.master.info.port</name> <value>@HBASE_MASTER_INFO_PORT@</value> </property> <property> <name>hbase.regionserver.port</name> <value>@HBASE_REGION_PORT@</value> </property> <property> <name>hbase.regionserver.info.port</name> <value>@HBASE_REGION_INFO_PORT@</value> </property> <property> <name>hbase.rootdir</name> <value>hdfs://@HDFS_PATH@/@ZNODE_PARENT@</value> </property> <property> <name>hbase.zookeeper.quorum</name> <value>@ZOOKEEPER_IP_LIST@</value> </property> <property> <name>hbase.zookeeper.property.clientPort</name> <value>@ZOOKEEPER_PORT@</value> </property> <property> <name>zookeeper.znode.parent</name> <value>/@ZNODE_PARENT@</value> </property> </configuration>
启动脚本 start-k8s-hbase.sh 主要完成参数替换、写入/etc/hosts、启动 hbase 的功能，内容如下：
1 #!/bin/bash 2 3 export HBASE_CONF_FILE=/opt/hbase/conf/hbase-site.xml 4 export HADOOP_USER_NAME=hdfs 5 export HBASE_MANAGES_ZK=false 6 7 sed -i "s/@HBASE_MASTER_PORT@/$HBASE_MASTER_PORT/g" $HBASE_CONF_FILE 8 sed -i "s/@HBASE_MASTER_INFO_PORT@/$HBASE_MASTER_INFO_PORT/g" $HBASE_CONF_FILE 9 sed -i "s/@HBASE_REGION_PORT@/$HBASE_REGION_PORT/g" $HBASE_CONF_FILE 10 sed -i "s/@HBASE_REGION_INFO_PORT@/$HBASE_REGION_INFO_PORT/g" $HBASE_CONF_FILE 11 sed -i "s/@HDFS_PATH@/$HDFS_SERVICE:$HDFS_PORT/$ZNODE_PARENT/g" $HBASE_CONF_FILE 12 sed -i "s/@ZOOKEEPER_IP_LIST@/$ZOOKEEPER_SERVICE_LIST/g" $HBASE_CONF_FILE 13 sed -i "s/@ZOOKEEPER_PORT@/$ZOOKEEPER_PORT/g" $HBASE_CONF_FILE 14 sed -i "s/@ZNODE_PARENT@/$ZNODE_PARENT/g" $HBASE_CONF_FILE 15 16 for i in ${HBASE_MASTER_LIST[@]} 17 do 18 arr=(${i//:/ }) 19 echo "${arr[0]} ${arr[1]}" >> /etc/hosts 20 done 21 22 for i in ${HBASE_REGION_LIST[@]} 23 do 24 arr=(${i//:/ }) 25 echo "${arr[0]} ${arr[1]}" >> /etc/hosts 26 done 27 28 if [ "$HBASE_SERVER_TYPE" = "master" ]; then 29 /opt/hbase/bin/hbase master start > logmaster.log 2>&1 30 elif [ "$HBASE_SERVER_TYPE" = "regionserver" ]; then 31 /opt/hbase/bin/hbase regionserver start > logregion.log 2>&1 32 fi
其中导出HADOOP_USER_NAME为hdfs用户，否则会报Permission Denied的错误；HBASE_MANAGES_ZK=false表示不使用HBase自带的ZooKeeper；HBASE_MASTER_LIST为HBase集群中除当前master外的其余master的服务地址和pod名的对应关系；HBASE_REGION_LIST为HBase集群中除当前regionserver外的其余regionserver的服务地址和pod名的对应关系；最后根据 HBASE_SERVER_TYPE 的取值来确定是启master还是regionserver。
准备好这些文件后即可创建HBase的image：
docker build -t 10.11.150.76:5000/openxxs/hbase:1.0 . docker push 10.11.150.76:5000/openxxs/hbase:1.0
随后创建hbase.yaml文件，内容如下：
1 apiVersion: v1 2 kind: Service 3 metadata: 4 name: hbase-master-1 5 spec: 6 selector: 7 app: hbase-master 8 server-id: "1" 9 type: ClusterIP 10 clusterIP: "172.16.30.1" 11 ports: 12 - name: rpc 13 port: 60000 14 targetPort: 60000 15 - name: info 16 port: 60001 17 targetPort: 60001 18 --- 19 apiVersion: v1 20 kind: Service 21 metadata: 22 name: hbase-master-2 23 spec: 24 selector: 25 app: hbase-master 26 server-id: "2" 27 type: ClusterIP 28 clusterIP: "172.16.30.2" 29 ports: 30 - name: rpc 31 port: 60000 32 targetPort: 60000 33 - name: info 34 port: 60001 35 targetPort: 60001 36 --- 37 apiVersion: v1 38 kind: Service 39 metadata: 40 name: hbase-region-1 41 spec: 42 selector: 43 app: hbase-region 44 server-id: "1" 45 type: ClusterIP 46 clusterIP: "172.16.30.3" 47 ports: 48 - name: rpc 49 port: 60010 50 targetPort: 60010 51 - name: info 52 port: 60011 53 targetPort: 60011 54 --- 55 apiVersion: v1 56 kind: Service 57 metadata: 58 name: hbase-region-2 59 spec: 60 selector: 61 app: hbase-region 62 server-id: "2" 63 type: ClusterIP 64 clusterIP: "172.16.30.4" 65 ports: 66 - name: rpc 67 port: 60010 68 targetPort: 60010 69 - name: info 70 port: 60011 71 targetPort: 60011 72 --- 73 apiVersion: v1 74 kind: Pod 75 metadata: 76 name: hbase-master-1 77 labels: 78 app: hbase-master 79 server-id: "1" 80 spec: 81 containers: 82 - name: hbase-master-1 83 image: 10.11.150.76:5000/openxxs/hbase:1.0 84 ports: 85 - containerPort: 60000 86 - containerPort: 60001 87 env: 88 - name: HBASE_SERVER_TYPE 89 value: master 90 - name: HBASE_MASTER_PORT 91 value: "60000" 92 - name: HBASE_MASTER_INFO_PORT 93 value: "60001" 94 - name: HBASE_REGION_PORT 95 value: "60010" 96 - name: HBASE_REGION_INFO_PORT 97 value: "60011" 98 - name: HDFS_SERVICE 99 value: "hdfs-namenode-service.default.domeos.sohu" 100 - name: HDFS_PORT 101 value: "4231" 102 - name: ZOOKEEPER_SERVICE_LIST 103 value: "zookeeper-1.default.domeos.sohu,zookeeper-2.default.domeos.sohu,zookeeper-3.default.domeos.sohu" 104 - name: ZOOKEEPER_PORT 105 value: "2181" 106 - name: ZNODE_PARENT 107 value: hbase 108 - name: HBASE_MASTER_LIST 109 value: "172.16.30.2:hbase-master-2" 110 - name: HBASE_REGION_LIST 111 value: "172.16.30.3:hbase-region-1 172.16.30.4:hbase-region-2" 112 restartPolicy: Always 113 nodeSelector: 114 kubernetes.io/hostname: bx-42-199 115 --- 116 apiVersion: v1 117 kind: Pod 118 metadata: 119 name: hbase-master-2 120 labels: 121 app: hbase-master 122 server-id: "2" 123 spec: 124 containers: 125 - name: hbase-master-1 126 image: 10.11.150.76:5000/openxxs/hbase:1.0 127 ports: 128 - containerPort: 60000 129 - containerPort: 60001 130 env: 131 - name: HBASE_SERVER_TYPE 132 value: master 133 - name: HBASE_MASTER_PORT 134 value: "60000" 135 - name: HBASE_MASTER_INFO_PORT 136 value: "60001" 137 - name: HBASE_REGION_PORT 138 value: "60010" 139 - name: HBASE_REGION_INFO_PORT 140 value: "60011" 141 - name: HDFS_SERVICE 142 value: "hdfs-namenode-service.default.domeos.sohu" 143 - name: HDFS_PORT 144 value: "4231" 145 - name: ZOOKEEPER_SERVICE_LIST 146 value: "zookeeper-1.default.domeos.sohu,zookeeper-2.default.domeos.sohu,zookeeper-3.default.domeos.sohu" 147 - name: ZOOKEEPER_PORT 148 value: "2181" 149 - name: ZNODE_PARENT 150 value: hbase 151 - name: HBASE_MASTER_LIST 152 value: "172.16.30.1:hbase-master-1" 153 - name: HBASE_REGION_LIST 154 value: "172.16.30.3:hbase-region-1 172.16.30.4:hbase-region-2" 155 restartPolicy: Always 156 nodeSelector: 157 kubernetes.io/hostname: bx-42-198 158 --- 159 apiVersion: v1 160 kind: Pod 161 metadata: 162 name: hbase-region-1 163 labels: 164 app: hbase-region-1 165 server-id: "1" 166 spec: 167 containers: 168 - name: hbase-region-1 169 image: 10.11.150.76:5000/openxxs/hbase:1.0 170 ports: 171 - containerPort: 60010 172 - containerPort: 60011 173 env: 174 - name: HBASE_SERVER_TYPE 175 value: regionserver 176 - name: HBASE_MASTER_PORT 177 value: "60000" 178 - name: HBASE_MASTER_INFO_PORT 179 value: "60001" 180 - name: HBASE_REGION_PORT 181 value: "60010" 182 - name: HBASE_REGION_INFO_PORT 183 value: "60011" 184 - name: HDFS_SERVICE 185 value: "hdfs-namenode-service.default.domeos.sohu" 186 - name: HDFS_PORT 187 value: "4231" 188 - name: ZOOKEEPER_SERVICE_LIST 189 value: "zookeeper-1.default.domeos.sohu,zookeeper-2.default.domeos.sohu,zookeeper-3.default.domeos.sohu" 190 - name: ZOOKEEPER_PORT 191 value: "2181" 192 - name: ZNODE_PARENT 193 value: hbase 194 - name: HBASE_MASTER_LIST 195 value: "172.16.30.1:hbase-master-1 172.16.30.2:hbase-master-2" 196 - name: HBASE_REGION_LIST 197 value: "172.16.30.4:hbase-region-2" 198 restartPolicy: Always 199 nodeSelector: 200 kubernetes.io/hostname: bx-42-199 201 --- 202 apiVersion: v1 203 kind: Pod 204 metadata: 205 name: hbase-region-2 206 labels: 207 app: hbase-region-2 208 server-id: "2" 209 spec: 210 containers: 211 - name: hbase-region-2 212 image: 10.11.150.76:5000/openxxs/hbase:1.0 213 ports: 214 - containerPort: 60010 215 - containerPort: 60011 216 env: 217 - name: HBASE_SERVER_TYPE 218 value: regionserver 219 - name: HBASE_MASTER_PORT 220 value: "60000" 221 - name: HBASE_MASTER_INFO_PORT 222 value: "60001" 223 - name: HBASE_REGION_PORT 224 value: "60010" 225 - name: HBASE_REGION_INFO_PORT 226 value: "60011" 227 - name: HDFS_SERVICE 228 value: "hdfs-namenode-service.default.domeos.sohu" 229 - name: HDFS_PORT 230 value: "4231" 231 - name: ZOOKEEPER_SERVICE_LIST 232 value: "zookeeper-1.default.domeos.sohu,zookeeper-2.default.domeos.sohu,zookeeper-3.default.domeos.sohu" 233 - name: ZOOKEEPER_PORT 234 value: "2181" 235 - name: ZNODE_PARENT 236 value: hbase 237 - name: HBASE_MASTER_LIST 238 value: "172.16.30.1:hbase-master-1 172.16.30.2:hbase-master-2" 239 - name: HBASE_REGION_LIST 240 value: "172.16.30.3:hbase-region-1" 241 restartPolicy: Always 242 nodeSelector: 243 kubernetes.io/hostname: bx-42-198
说明：该yaml文件共创建了两个master服务、两个regionserver服务，以及对应的两个master Pods和两个regionserver Pods；Pod的restartPolicy设为Always表示如果该Pod挂掉的话将一直尝试重新启动它；以环境变量的形式将参数传递进Pod中，其中HDFS_SERVICE为HDFS服务经过skyDNS之后的对应域名，若未设置skyDNS则此处值设为HDFS服务对应的IP地址，ZOOKEEPER_SERVICE_LIST同理；HBASE_MASTER_LIST的值格式为 <master服务IP地址>:<master对应Pod名>，多个项之间以空格分隔，HBASE_REGION_LIST同理。
接着就可以创建和查看HBase服务了：
# 创建 $kubectl create -f hbase.yaml service "hbase-master-1" created service "hbase-master-2" created service "hbase-region-1" created service "hbase-region-2" created pod "hbase-master-1" created pod "hbase-master-2" created pod "hbase-region-1" created pod "hbase-region-2" created # 查看pods $kubectl get pods NAME READY STATUS RESTARTS AGE hbase-master-1 1/1 Running 0 5s hbase-master-2 0/1 Pending 0 5s hbase-region-1 1/1 Running 0 5s hbase-region-2 0/1 Pending 0 5s hdfs-datanode-1-nek2l 1/1 Running 3 7d hdfs-datanode-2-vkbbt 1/1 Running 3 7d hdfs-datanode-3-h4jvt 1/1 Running 3 7d hdfs-namenode-1-cl0pj 1/1 Running 3 7d kube-dns-v8-x8igc 3/3 Running 0 4h zookeeper-1-ojhmy 1/1 Running 0 12h zookeeper-2-cr73i 1/1 Running 0 12h zookeeper-3-79ls0 1/1 Running 0 12h # 查看service $kubectl get service NAME CLUSTER_IP EXTERNAL_IP PORT(S) SELECTOR AGE hbase-master-1 172.16.30.1 <none> 60000/TCP,60001/TCP app=hbase-master,server-id=1 17m hbase-master-2 172.16.30.2 <none> 60000/TCP,60001/TCP app=hbase-master,server-id=2 17m hbase-region-1 172.16.30.3 <none> 60010/TCP,60011/TCP app=hbase-region,server-id=1 17m hbase-region-2 172.16.30.4 <none> 60010/TCP,60011/TCP app=hbase-region,server-id=2 17m hdfs-namenode-service 172.16.20.1 <none> 4231/TCP,50020/TCP,50090/TCP,50070/TCP,50010/TCP,50075/TCP,8031/TCP,8032/TCP,8033/TCP,8040/TCP,8042/TCP,49707/TCP,22/TCP,8088/TCP,8030/TCP app=hdfs-namenode 7d kube-dns 172.16.40.1 <none> 53/UDP,53/TCP app=kube-dns,version=v8 10h kubernetes 172.16.0.1 <none> 443/TCP <none> 12d zookeeper-1 172.16.11.1 <none> 2181/TCP,2888/TCP,3888/TCP name=zookeeper,server-id=1 13h zookeeper-2 172.16.11.2 <none> 2181/TCP,2888/TCP,3888/TCP name=zookeeper,server-id=2 13h zookeeper-3 172.16.11.3 <none> 2181/TCP,2888/TCP,3888/TCP name=zookeeper,server-id=3 13h
通过ZooKeeper的zkCli.sh可以看到/hbase下对应的master和rs的记录（显示乱码是由于系统显示时编码的原因，无影响）：
[zk: localhost:2181(CONNECTED) 0] ls /hbase [meta-region-server, backup-masters, table, draining, region-in-transition, table-lock, running, master, namespace, hbaseid, online-snapshot, replication, splitWAL, recovering-regions, rs] [zk: localhost:2181(CONNECTED) 1] ls /hbase/rs [172.27.0.0,60010,1448896399329, 172.28.0.115,60010,1448896360650] [zk: localhost:2181(CONNECTED) 2] get /hbase/master ?master:60000??E*?O=PBUF base-master-1???????* cZxid = 0x100000186 ctime = Mon Nov 30 15:12:42 UTC 2015 mZxid = 0x100000186 mtime = Mon Nov 30 15:12:42 UTC 2015 pZxid = 0x100000186 cversion = 0 dataVersion = 0 aclVersion = 0 ephemeralOwner = 0x151563a5e37001a dataLength = 60 numChildren = 0
可以通过docker exec进入到HBase对应的容器中进行表操作以测试HBase的工作状态：
# 进入198的hbase-master-2容器中 [@bx_42_198 /opt/scs/openxxs]# docker exec -it f131fcf15a72 /bin/bash # 使用hbase shell对hbase进行操作 [root@hbase-master-2 bin]# hbase shell 2015-11-30 15:15:58,632 INFO [main] Configuration.deprecation: hadoop.native.lib is deprecated. Instead, use io.native.lib.available HBase Shell; enter 'help<RETURN>' for list of supported commands. Type "exit<RETURN>" to leave the HBase Shell Version 0.98.10.1-hadoop2, rd5014b47660a58485a6bdd0776dea52114c7041e, Tue Feb 10 11:34:09 PST 2015 # 通过status查看状态，这里显示的 2 dead 是之前测试时遗留的记录，无影响 hbase(main):001:0> status 2015-11-30 15:16:03,551 WARN [main] util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable 2 servers, 2 dead, 1.5000 average load # 创建表 hbase(main):002:0> create 'test','id','name' 0 row(s) in 0.8330 seconds => Hbase::Table - test # 查看表 hbase(main):003:0> list TABLE member test 2 row(s) in 0.0240 seconds => ["member", "test"] # 插入数据 hbase(main):004:0> put 'test','test1','id:5','addon' 0 row(s) in 0.1540 seconds # 查看数据 hbase(main):005:0> get 'test','test1' COLUMN CELL id:5 timestamp=1448906130803, value=addon 1 row(s) in 0.0490 seconds # 进入199的hbase-master-1容器中查看从198上插入的数据 hbase(main):001:0> get 'test','test1' 2015-11-30 18:01:23,944 WARN [main] util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable COLUMN CELL id:5 timestamp=1448906130803, value=addon 1 row(s) in 0.2430 seconds
从以上结果可以看出HBase算是在k8s中以不怎么完善的方式跑起来了。
六、讨论
不得不维护/etc/hosts记录的原因是HBase中master之间、regionserver之间、master与regionserver之间都是通过hostname来彼此识别对方的，而k8s的DNS只针对service并没有对Pod进行解析。而如果把master和regionserver放在同一个Pod下的话存在磁盘资源共享冲突的问题（还没仔细研究）。Github讨论中下面的这段话很直白地说明了HBase对hostname的依赖：
1. Master uses ZooKeeper to run master election, the winner puts its hostname:port in ZooKeeper. 2. RegionServers will register themselves to the winning Master through its hostname registered in ZooKeeper. 3. We have a RegionServer with meta table/region (which stores Region -> RegionServer map), it stores its hostname:port in ZooKeeper. 4. When a client starts a request to HBase, it gets the hostname of the RegionServer with meta table from ZooKeeper, scan the meta table, and find out the hostname of the RegionServer it interests, and finally talks to the RegionServer through hostname.
为解决这个问题目前尝试过的方法如下：
1. 配置skyDNS：skyDNS只针对service进行解析，无法解析Pod的名称。如果向skyDNS中插入hostname的相关记录并动态维护的话或许可以解决该问题，目前正在尝试中。
2. 更改创建 ReplicationController、Server、Pod、Container 时的各种设置参数，如 name、generateName等，然并卵。
3. 创建container后启动 master 前通过脚本更改 hostname：Docker只允许在Create Container时进行hostname的修改（docker run自身有一个hostname的参数可以指定Container的hostname），但在容器运行之后并不允许修改，修改则报如下错误：docker Error: hostname: you must be root to change the hostname. 这个错误有些误导，事实上是docker的机制不允许你去修改hostname而不是权限问题，用root也没法改。
4. 修改HBase参数使其上报到ZK中的值不是hostname而是IP地址：这个一度是前景光明的解决方案，但将hostname写入ZK在HBase中是硬编码在代码中的，并没有参数可以去设置此项。有人给出了个patch（戳这里），但测试结果并不好。
关于第五部分HBase部署方案的说明：选择使用单Pod而不是ReplicationController，是因为k8s会在RC中Container的hostname后面加上随机字符以区分彼此，而单Pod的Pod name和hostname是一致的；restartPolicy设为Always算是为单Pod方式鲁棒性提供点小小的补偿吧；如果将Pod name设置为对应service的IP或域名怎样？然而hostname并不允许带点号；写入 /etc/hosts 中的IP选择了service的而非Pod的，因为Pod中的IP在运行前并不能获取到，而且在重启Pod后也会发生改变，而service的IP是不变的，因此选择了 serviceIP:PodName 这种对应关系。
最根本的解决方案是让k8s支持hostname（或者说Pod）的DNS解析，前面配置ZooKeeper同样存在hostname这个问题（戳这里），后面将要部署的Kafka也会有这个问题。k8s的开发团队已经进行了许多讨论并准备解决这个问题了（戳这里），希望下个版本会有相关设置。