Docker部署Elasticsearch集群

Docker部署Elasticsearch集群

参考文档:

https://hub.docker.com/r/library/elasticsearch/

https://hub.tenxcloud.com/repos/tenxcloud/elasticsearch

https://hub.alauda.cn/repos/library/elasticsearch

环境:

CentOS 7.2

docker-engine-1.11.2

elasticsearch-2.3.3

前言:

虚拟机节点部署请参看Elasticsearch 负载均衡集群，这里简单介绍下docker部署

本实验采用不同类型的节点集群(client x1, master x3, data x2)

ela-client.example.com:192.168.8.10(client node)

ela-master1.example.com:192.168.8.101(master node)

ela-master2.example.com:192.168.8.102(master node)

ela-master3.example.com:192.168.8.103(master node)

ela-data1.example.com:192.168.8.201(data node)

ela-data2.example.com:192.168.8.202(data node)

一.安装docker-engine(所有节点)

请参看CentOS6/7 docker安装

docker镜像加速请参看Docker Hub加速及私有镜像搭建

二.获取elasticsearch docker镜像

docker pull elasticsearch:2.3.3

测试运行elasticsearch

docker run -d -p 9200:9200 -p 9300:9300 --name=elasticsearch-test elasticsearch:2.3.3

查看容器配置

docker inspect elasticsearch-test

删除测试容器

docker rm -f $(docker ps -aq)

提示:对于持续增长的目录或文件可以通过映射到docker主机来提高可定制性，如data,logs

三.配置并启动节点

创建集群名为elasticsearch_cluster的集群(默认的集群名为elasticsearch)

A.client node

ela-client.example.com:192.168.8.10(client node)

docker run -d --restart=always -p 9200:9200 -p 9300:9300 --name=elasticsearch-client --oom-kill-disable=true --memory-swappiness=1 -v /opt/elasticsearch/data:/usr/share/elasticsearch/data -v /opt/elasticsearch/logs:/usr/share/elasticsearch/logs elasticsearch:2.3.3

cat >elasticsearch.yml <<HERE

cluster.name: elasticsearch_cluster

node.name: ${HOSTNAME}

node.master: false

node.data: false

path.data: /usr/share/elasticsearch/data

path.logs: /usr/share/elasticsearch/logs

bootstrap.mlockall: true

network.host: 0.0.0.0

network.publish_host: 192.168.8.10

transport.tcp.port: 9300

http.port: 9200

index.refresh_interval: 5s

script.inline: true

script.indexed: true

HERE

docker cp elasticsearch.yml elasticsearch-client:/usr/share/elasticsearch/config/elasticsearch.yml

docker restart elasticsearch-client

直接将修改好的配置文件cp到容器对应位置后重启容器

man docker-run

       --net="bridge"

          Set the Network mode for the container

                                      'bridge': create a network stack on the default Docker bridge

                                      'none': no networking

                                      'container:': reuse another container's network stack

                                      'host': use the Docker host network stack. Note: the host mode gives the container full access to local system services such as D-bus and is  therefore  consid‐

       ered insecure.

说明:docker默认的网络模式为bridge,会自动为容器分配一个私有地址，如果是多台宿主机之间集群通信需要借助Consul,Etcd,Doozer等服务发现，自动注册组件来协同。请参看Docker集群之Swarm+Consul+Shipyard

必须通过network.publish_host: 192.168.8.10参数指定elasticsearch节点对外的监听地址,非常重要,不指定的话,集群节点间无法正常通信,报错如下

[2016-06-21 05:50:19,123][INFO ][discovery.zen ] [consul-s2.example.com] failed to send join request to master [{consul-s1.example.com}{DeKixlVMS2yoynzX8Y-gdA}{172.17.0.1}{172.17.0.1:9300}{data=false, master=true}], reason [RemoteTransportException[[consul-s2.example.com][172.17.0.1:9300]

最简单的，还可以网络直接设为host模式--net=host，直接借用宿主机的网络接口,换言之,不做网络层的layer

同时可禁用OOM，还可根据宿主机内存来设置-m参数(默认为0，无限)限制容器内存大小

[root@ela-client ~]# docker ps

CONTAINER ID        IMAGE                 COMMAND                  CREATED             STATUS              PORTS               NAMES

762e4d21aaf8        elasticsearch:2.3.3   "/docker-entrypoint.s"   2 minutes ago       Up 2 minutes                            elasticsearch-client

[root@ela-client ~]# netstat -tunlp|grep java

tcp        0      0 0.0.0.0:9200            0.0.0.0:*               LISTEN      18952/java          

tcp        0      0 0.0.0.0:9300            0.0.0.0:*               LISTEN      18952/java          

[root@ela-client ~]# ls /opt/elasticsearch/

data  logs

[root@ela-client ~]# docker logs $(docker ps -q)

[2016-06-13 16:09:51,308][INFO ][node                     ] [Sunfire] version[2.3.3], pid[1], build[218bdf1/2016-05-17T15:40:04Z]

[2016-06-13 16:09:51,311][INFO ][node                     ] [Sunfire] initializing ...

... ...

[2016-06-13 16:09:56,408][INFO ][node                     ] [Sunfire] started

[2016-06-13 16:09:56,417][INFO ][gateway                  ] [Sunfire] recovered [0] indices into cluster_state

或者

卷映射,个人认为卷映射更便于管理与备份

mkdir -p /opt/elasticsearch/config

cat >/opt/elasticsearch/config/elasticsearch.yml <<HERE

cluster.name: elasticsearch_cluster

node.name: ${HOSTNAME}

node.master: false

node.data: false

path.data: /usr/share/elasticsearch/data

path.logs: /usr/share/elasticsearch/logs

bootstrap.mlockall: true

network.host: 0.0.0.0

network.publish_host: 192.168.8.10

transport.tcp.port: 9300

http.port: 9200

index.refresh_interval: 5s

script.inline: true

script.indexed: true

HERE

docker run -tid --restart=always 

    -p 9200:9200

    -p 9300:9300

    --oom-kill-disable=true

    --memory-swappiness=1

    -v /opt/elasticsearch/data:/usr/share/elasticsearch/data

    -v /opt/elasticsearch/logs:/usr/share/elasticsearch/logs

    -v /opt/elasticsearch/config/elasticsearch.yml:/usr/share/elasticsearch/config/elasticsearch.yml

    --name=elasticsearch-client

    elasticsearch:2.3.3

B.master node

ela-master1.example.com:192.168.8.101(master node)

mkdir -p /opt/elasticsearch/config

cat >/opt/elasticsearch/config/elasticsearch.yml <<HERE

cluster.name: elasticsearch_cluster

node.name: ${HOSTNAME}

node.master: true

node.data: false

path.data: /usr/share/elasticsearch/data

path.logs: /usr/share/elasticsearch/logs

bootstrap.mlockall: true

network.host: 0.0.0.0

network.publish_host: 192.168.8.101

transport.tcp.port: 9300

http.port: 9200

index.refresh_interval: 5s

script.inline: true

script.indexed: true

HERE

docker run -tid --restart=always 

    -p 9200:9200

    -p 9300:9300

    --oom-kill-disable=true

    --memory-swappiness=1

    -v /opt/elasticsearch/data:/usr/share/elasticsearch/data

    -v /opt/elasticsearch/logs:/usr/share/elasticsearch/logs

    -v /opt/elasticsearch/config/elasticsearch.yml:/usr/share/elasticsearch/config/elasticsearch.yml

    --name=elasticsearch-master1 

    elasticsearch:2.3.3

ela-master2.example.com:192.168.8.102(master node)

mkdir -p /opt/elasticsearch/config

cat >/opt/elasticsearch/config/elasticsearch.yml <<HERE

cluster.name: elasticsearch_cluster

node.name: ${HOSTNAME}

node.master: true

node.data: false

path.data: /usr/share/elasticsearch/data

path.logs: /usr/share/elasticsearch/logs

bootstrap.mlockall: true

network.host: 0.0.0.0

network.publish_host: 192.168.8.102

transport.tcp.port: 9300

http.port: 9200

index.refresh_interval: 5s

script.inline: true

script.indexed: true

HERE

docker run -tid --restart=always 

    -p 9200:9200

    -p 9300:9300

    --oom-kill-disable=true

    --memory-swappiness=1

    -v /opt/elasticsearch/data:/usr/share/elasticsearch/data

    -v /opt/elasticsearch/logs:/usr/share/elasticsearch/logs

    -v /opt/elasticsearch/config/elasticsearch.yml:/usr/share/elasticsearch/config/elasticsearch.yml

    --name=elasticsearch-master2 

    elasticsearch:2.3.3

ela-master3.example.com:192.168.8.103(master node)

mkdir -p /opt/elasticsearch/config

cat >/opt/elasticsearch/config/elasticsearch.yml <<HERE

cluster.name: elasticsearch_cluster

node.name: ${HOSTNAME}

node.master: true

node.data: false

path.data: /usr/share/elasticsearch/data

path.logs: /usr/share/elasticsearch/logs

bootstrap.mlockall: true

network.publish_host: 192.168.8.103

transport.tcp.port: 9300

network.host: 0.0.0.0

http.port: 9200

index.refresh_interval: 5s

script.inline: true

script.indexed: true

HERE

docker run -tid --restart=always 

    -p 9200:9200

    -p 9300:9300

    --oom-kill-disable=true

    --memory-swappiness=1

    -v /opt/elasticsearch/data:/usr/share/elasticsearch/data

    -v /opt/elasticsearch/logs:/usr/share/elasticsearch/logs

    -v /opt/elasticsearch/config/elasticsearch.yml:/usr/share/elasticsearch/config/elasticsearch.yml

    --name=elasticsearch-master3 

    elasticsearch:2.3.3

C.data node

ela-data1.example.com:192.168.8.201(data node)

mkdir -p /opt/elasticsearch/config

cat >/opt/elasticsearch/config/elasticsearch.yml <<HERE

cluster.name: elasticsearch_cluster

node.name: ${HOSTNAME}

node.master: false

node.data: true

path.data: /usr/share/elasticsearch/data

path.logs: /usr/share/elasticsearch/logs

bootstrap.mlockall: true

network.publish_host: 192.168.8.201

transport.tcp.port: 9300

network.host: 0.0.0.0

http.port: 9200

index.refresh_interval: 5s

script.inline: true

script.indexed: true

HERE

docker run -tid --restart=always 

    -p 9200:9200

    -p 9300:9300

    --oom-kill-disable=true

    --memory-swappiness=1

    -v /opt/elasticsearch/data:/usr/share/elasticsearch/data

    -v /opt/elasticsearch/logs:/usr/share/elasticsearch/logs

    -v /opt/elasticsearch/config/elasticsearch.yml:/usr/share/elasticsearch/config/elasticsearch.yml

    --name=elasticsearch-data1 

    elasticsearch:2.3.3

ela-data2.example.com:192.168.8.202(data node)

mkdir -p /opt/elasticsearch/config

cat >/opt/elasticsearch/config/elasticsearch.yml <<HERE

cluster.name: elasticsearch_cluster

node.name: ${HOSTNAME}

node.master: false

node.data: true

path.data: /usr/share/elasticsearch/data

path.logs: /usr/share/elasticsearch/logs

bootstrap.mlockall: true

network.host: 0.0.0.0

network.publish_host: 192.168.8.202

transport.tcp.port: 9300

http.port: 9200

index.refresh_interval: 5s

script.inline: true

script.indexed: true

HERE

docker run -tid --restart=always 

    -p 9200:9200

    -p 9300:9300

    --oom-kill-disable=true

    --memory-swappiness=1

    -v /opt/elasticsearch/data:/usr/share/elasticsearch/data

    -v /opt/elasticsearch/logs:/usr/share/elasticsearch/logs

    -v /opt/elasticsearch/config/elasticsearch.yml:/usr/share/elasticsearch/config/elasticsearch.yml

    --name=elasticsearch-data2 

    elasticsearch:2.3.3

四.配置集群(所有节点)

通过discovery模块来将节点加入集群

在以上节点的配置文件/opt/elasticsearch/config/elasticsearch.yml中加入如下行后重启

cat >>/opt/elasticsearch/config/elasticsearch.yml <<HERE

discovery.zen.ping.timeout: 100s

discovery.zen.fd.ping_timeout: 100s

discovery.zen.ping.multicast.enabled: false

discovery.zen.ping.unicast.hosts: ["192.168.8.101:9300", "192.168.8.102:9300", "192.168.8.103:9300", "192.168.8.201:9300", "192.168.8.202:9300","192.168.8.10:9300"]

discovery.zen.minimum_master_nodes: 2

gateway.recover_after_nodes: 2

HERE

docker restart $(docker ps -a|grep elasticsearch|awk '{print $1}')

五.确认集群

等待30s左右，集群节点会自动join,在集群的含意节点上都会看到如下类似输出，说明集群运行正常

REST API调用请参看Elasticsearch REST API小记

https://www.elastic.co/guide/en/elasticsearch/reference/current/_cluster_health.html

[root@ela-client ~]#curl 'http://localhost:9200/_cat/health?v'

epoch      timestamp cluster               status node.total node.data shards pri relo init unassign pending_tasks max_task_wait_time active_shards_percent 

1465843145 18:39:05  elasticsearch_cluster green           6         2      0   0    0    0        0             0                  -                100.0% 

[root@ela-client ~]#curl 'localhost:9200/_cat/nodes?v'

host            ip              heap.percent ram.percent load node.role master name                   

192.168.8.102 192.168.8.102           14          99 0.00 -         *      ela-master2.example.com  

192.168.8.103 192.168.8.103            4          99 0.14 -         m      ela-master3.example.com  

192.168.8.202 192.168.8.202           11          99 0.00 d         -      ela-data2.example.com  

192.168.8.10  192.168.8.10            10          98 0.17 -         -      ela-client.example.com 

192.168.8.201 192.168.8.201           11          99 0.00 d         -      ela-data1.example.com  

192.168.8.101 192.168.8.101           12          99 0.01 -         m      ela-master1.example.com 

[root@ela-master2 ~]#curl 'http://localhost:9200/_nodes/process?pretty'

{

  "cluster_name" : "elasticsearch_cluster",

  "nodes" : {

    "naMz_y4uRRO-FzyxRfTNjw" : {

      "name" : "ela-data2.example.com",

      "transport_address" : "192.168.8.202:9300",

      "host" : "192.168.8.202",

      "ip" : "192.168.8.202",

      "version" : "2.3.3",

      "build" : "218bdf1",

      "http_address" : "192.168.8.202:9200",

      "attributes" : {

        "master" : "false"

      },

      "process" : {

        "refresh_interval_in_millis" : 1000,

        "id" : 1,

        "mlockall" : false

      }

    },

    "7FwFY20ESZaRtIWhYMfDAg" : {

      "name" : "ela-data1.example.com",

      "transport_address" : "192.168.8.201:9300",

      "host" : "192.168.8.201",

      "ip" : "192.168.8.201",

      "version" : "2.3.3",

      "build" : "218bdf1",

      "http_address" : "192.168.8.201:9200",

      "attributes" : {

        "master" : "false"

      },

      "process" : {

        "refresh_interval_in_millis" : 1000,

        "id" : 1,

        "mlockall" : false

      }

    },

    "X0psLpQyR42A4ThiP8ilhA" : {

      "name" : "ela-master3.example.com",

      "transport_address" : "192.168.8.103:9300",

      "host" : "192.168.8.103",

      "ip" : "192.168.8.103",

      "version" : "2.3.3",

      "build" : "218bdf1",

      "http_address" : "192.168.8.103:9200",

      "attributes" : {

        "data" : "false",

        "master" : "true"

      },

      "process" : {

        "refresh_interval_in_millis" : 1000,

        "id" : 1,

        "mlockall" : false

      }

    },

    "MG_GlSAZRkqLq8gMqaZITw" : {

      "name" : "ela-master1.example.com",

      "transport_address" : "192.168.8.101:9300",

      "host" : "192.168.8.101",

      "ip" : "192.168.8.101",

      "version" : "2.3.3",

      "build" : "218bdf1",

      "http_address" : "192.168.8.101:9200",

      "attributes" : {

        "data" : "false",

        "master" : "true"

      },

      "process" : {

        "refresh_interval_in_millis" : 1000,

        "id" : 1,

        "mlockall" : false

      }

    },

    "YxNHUPqVRNK3Liilw_hU9A" : {

      "name" : "ela-master2.example.com",

      "transport_address" : "192.168.8.102:9300",

      "host" : "192.168.8.102",

      "ip" : "192.168.8.102",

      "version" : "2.3.3",

      "build" : "218bdf1",

      "http_address" : "192.168.8.102:9200",

      "attributes" : {

        "data" : "false",

        "master" : "true"

      },

      "process" : {

        "refresh_interval_in_millis" : 1000,

        "id" : 1,

        "mlockall" : false

      }

    },

    "zTKJJ4ipQg6xAcwy1aE-9g" : {

      "name" : "ela-client.example.com",

      "transport_address" : "192.168.8.10:9300",

      "host" : "192.168.8.10",

      "ip" : "192.168.8.10",

      "version" : "2.3.3",

      "build" : "218bdf1",

      "http_address" : "192.168.8.10:9200",

      "attributes" : {

        "data" : "false",

        "master" : "true"

      },

      "process" : {

        "refresh_interval_in_millis" : 1000,

        "id" : 1,

        "mlockall" : false

      }

    }

  } 

}

问题:

修改jvm内存,初始值-Xms256m -Xmx1g

https://hub.docker.com/r/itzg/elasticsearch/

https://hub.docker.com/_/elasticsearch/

实测:修改JAVA_OPTS,ES_JAVA_OPTS都只能追加，而不能覆盖，只能通过

-e ES_HEAP_SIZE="32g" 来覆盖