Zookeeper installation and pseudo-cluster deployment

Overview (Overview)
Http://zookeeper.apache.org/doc/r3.4.6/zookeeperOver.html

Getting Started (Getting started)
Http://zookeeper.apache.org/doc/r3.4.6/zookeeperStarted.html

Tutorial (Tutorial)
Http://zookeeper.apache.org/doc/r3.4.6/zookeeperTutorial.html

Java Example (Java sample)
Http://zookeeper.apache.org/doc/r3.4.6/javaExample.html

Programmer ' s Guide (developer Guides)
Http://zookeeper.apache.org/doc/r3.4.6/zookeeperProgrammers.html

Recipes and Solutions (Tips and solutions)
Http://zookeeper.apache.org/doc/r3.4.6/recipes.html

3.4.6 API Online (online API quick Check)

Http://zookeeper.apache.org/doc/r3.4.6/api/index.html

Another recommended zookeeper series of friends Sunddenly
Http://www.cnblogs.com/sunddenly/category/620563.html

First, download

http://apache.opencas.org/zookeeper/

or http://mirrors.cnnic.cn/apache/zookeeper/download the latest version

Download unzip
Extract to 3 directories (simulate 3 ZK servers):
/soft/zookeeper-1
/soft/zookeeper-2
/soft/zookeeper-3
Create a conf/zoo.cfg configuration file under each directory

/soft/zookeeper-1/conf/zoo.cfg content is as follows:

ticktime=2000
initlimit=10
Synclimit=5
Datadir=/soft/tmp/zk1/data
Datalogdir=/soft/tmp/zk1/log
clientport=2181
server.1=localhost:2287:3387
server.2=localhost:2288:3388
server.3=localhost:2289:3389

/soft/zookeeper-2/conf/zoo.cfg content is as follows:

ticktime=2000
initlimit=10
Synclimit=5
Datadir=/soft/tmp/zk2/data
Datalogdir=/soft/tmp/zk2/log
clientport=2182
server.1=localhost:2287:3387
server.2=localhost:2288:3388
server.3=localhost:2289:3389

/soft/zookeeper-3/conf/zoo.cfg content is as follows:

ticktime=2000
initlimit=10
Synclimit=5
Datadir=/soft/tmp/zk3/data
Datalogdir=/soft/tmp/zk3/log
clientport=2183
server.1=localhost:2287:3387
server.2=localhost:2288:3388
server.3=localhost:2289:3389

Note: Because the cluster is simulated on a single machine, the ports cannot be duplicated, and the 2181~2183,2287~2289 and 3387~3389 are staggered. In addition, each ZK instance, need to set up a separate data storage directory, log storage directory, so DataDir, datalogdir these two nodes corresponding directories, need to be created manually first.
Mkdir/soft/tmp/zk1/data

Mkdir/soft/tmp/zk1/log

Mkdir/soft/tmp/zk2/data

Mkdir/soft/tmp/zk2/log

Mkdir/soft/tmp/zk3/data

Mkdir/soft/tmp/zk3/log
There is also a very critical setting in which a file named myID must be created in the same directory as the DataDir for each ZK server configuration file, where the content must be the same as the X in server.x in zoo.cfg, i.e.:
cd/soft/tmp/zk1/data/

Vimyid Input 1 Save

cd/soft/tmp/zk2/data/

Vimyid Input 2 Save

cd/soft/tmp/zk3/data/

Vimyid Input 3 Save

The content in/soft/tmp/zk1/data/myid is 1, corresponding to 1 in Server.1
The content in/soft/tmp/zk2/data/myid is 2, corresponding to 2 in Server.2
The content in/soft/tmp/zk3/data/myid is 3, corresponding to 3 in SERVER.3

In a production environment, the steps for distributed cluster deployment are basically the same as above, except that because each ZK server is distributed on different machines, localhost in the above configuration file is replaced with the real IP of each server. Distributed in different machines, there is no port conflict problem, you can make each server ZK use the same port, so it is easier to manage.

Start validation

/soft/zookeeper-1/bin/zkserver.sh start

/soft/zookeeper-2/bin/zkserver.sh start

/soft/zookeeper-3/bin/zkserver.sh start

Input JPS Check

Ps–ef|grep Zookeeper can also display all boot zookeeper ports

Use a client connection attempt:

Sh/soft/zookeeper-2/bin/zkcli.sh-server localhost:2181

Startup success:

Watcher::

Watchedevent state:syncconnected Type:none Path:null

[zk:localhost:2181 (CONNECTED) 0]

Client command:

LS/view root node; Ls/root View all nodes under the/root node

Create Path data Creation node

Set path data [watcher] sets the node value

Get path data gets the specified node value

Set on the same machine as a pseudo-cluster, but the same implementation of the random server on which to make the node changes, the other nodes will synchronize the effect of time.