1, Single machine environment construction

1.1 download

Download the corresponding version of Zookeeper, here I download version 3.4.14. Official download address:

# https://www.apache.org/dyn/closer.lua/zookeeper/zookeeper-3.4.14/zookeeper-3.4.14.tar.gz
#perhaps
wget https://archive.apache.org/dist/zookeeper/zookeeper-3.4.14/zookeeper-3.4.14.tar.gz

1.2 decompression

# tar -zxvf zookeeper-3.4.14.tar.gz  -C /home/hadoop/

1.3 configure environment variables

# vim /etc/profile

Add environment variable:

export ZOOKEEPER_HOME=/home/hadoop/zookeeper-3.4.14
export PATH=$ZOOKEEPER_HOME/bin:$PATH

Make the configured environment variables effective:

# source /etc/profile

1.4 modify configuration

Enter the conf / directory of the installation directory, copy the configuration sample and modify it:

# cp zoo_sample.cfg  zoo.cfg

Specify the data storage directory and log file directory (the directory does not need to be created in advance, and the program will automatically create it). After modification, the complete configuration is as follows:

# The number of milliseconds of each tick
tickTime=2000
# The number of ticks that the initial
# synchronization phase can take
initLimit=10
# The number of ticks that can pass between
# sending a request and getting an acknowledgement
syncLimit=5
# the directory where the snapshot is stored.
# do not use /tmp for storage, /tmp here is just
# example sakes.
dataDir=/data/zookeeper/data
dataLogDir=/data/zookeeper/log
# the port at which the clients will connect
clientPort=2181
# the maximum number of client connections.
# increase this if you need to handle more clients
#maxClientCnxns=60
#
# Be sure to read the maintenance section of the
# administrator guide before turning on autopurge.
#
# http://zookeeper.apache.org/doc/current/zookeeperAdmin.html#sc_maintenance
#
# The number of snapshots to retain in dataDir
#autopurge.snapRetainCount=3
# Purge task interval in hours
# Set to "0" to disable auto purge feature
#autopurge.purgeInterval=1

1.5 start up

Since the environment variable has been configured, you can start it directly with the following command:

zkServer.sh start

1.6 verification

Use JPS to verify whether the process has been started, and the presence of QuorumPeerMain indicates that the start is successful.

[hadoop@hadoop-nn-01 conf]$ jps -m
37193 QuorumPeerMain /home/hadoop/zookeeper-3.4.14/bin/../conf/zoo.cfg
37210 Jps -m

2, Cluster environment construction

In order to ensure the high availability of the cluster, the number of nodes in Zookeeper cluster should be odd, and there should be at least three nodes, so this is a demonstration of building a three node cluster. Here, I use three hosts to build. The host names are respectively hadoop-nn-01, hadoop-nn-02 and hadoop-dn-01.

2.1 modify configuration

Extract a zookeeper installation package and modify its configuration file zoo.cfg , as follows. Then use the scp command to distribute the installation package to three servers and configure the environment variables on three machines:

tickTime=2000
initLimit=10
syncLimit=5
dataDir=/data/zookeeper/data
dataLogDir=/data/zookeeper/log
clientPort=2181

# server.1 this 1 is the identification of the server, which can be any valid number. This is the identification of the server node. This identification should be written to the myid file under the dataDir directory
# Name inter cluster communication port and election port
server.1=hadoop-nn-01:2888:3888
server.2=hadoop-nn-02:2888:3888
server.3=hadoop-dn-01:2888:3888

(host name, Heartbeat port, data port)

# Distribute the installation package to hadoop-nn-02
scp -r /home/hadoop/zookeeper-3.4.14/  hadoop-nn-02:/home/hadoop/
# Distribute the installation package to hadoop-dn-01
scp -r /home/hadoop/zookeeper-3.4.14/  hadoop-dn-01:/home/hadoop/

2.2 identification node

Create a new myid file in the dataDir directory of the three hosts, and write the corresponding node ID. Zookeeper cluster identifies cluster nodes through myid file, and carries out node communication through node communication port and election port configured above, and selects Leader node.

To create a storage directory:

# All three hosts execute the command
mkdir -vp  /data/zookeeper/data/

Create and write the node ID to the myid file:

# hadoop-nn-01 host
echo "1" > /data/zookeeper/data/myid
# hadoop-nn-02 host
echo "2" > /data/zookeeper/data/myid
# hadoop-dn-01 host
echo "3" > /data/zookeeper/data/myid

2.3 start cluster

On three hosts, execute the following command to start the service:

/home/hadoop/zookeeper-3.4.14/bin/zkServer.sh start /home/hadoop/zookeeper-3.4.14/conf/zoo.cfg

2.4 cluster verification

Use after startup zkServer.sh Status view the status of each node in the cluster, and hadoop-nn-02 is the leader node, hadoop-nn-01 and hadoop-dn-01 are the follower nodes.

[hadoop@hadoop-nn-01 bin]$ zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /home/hadoop/zookeeper-3.4.14/bin/../conf/zoo.cfg
Mode: follower

[hadoop@hadoop-nn-02 bin]$ ./zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /home/hadoop/zookeeper-3.4.14/bin/../conf/zoo.cfg
Mode: leader

[hadoop@hadoop-dn-01 conf]$ zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /home/hadoop/zookeeper-3.4.14/bin/../conf/zoo.cfg
Mode: follower

3, Some operations of zookeeper

3.1 change zoo.cfg

scp /home/hadoop/zookeeper-3.4.14/conf/zoo.cfg  hadoop-nn-02:/home/hadoop/zookeeper-3.4.14/conf/
scp /home/hadoop/zookeeper-3.4.14/conf/zoo.cfg  hadoop-dn-01:/home/hadoop/zookeeper-3.4.14/conf/

three point two zkServer.sh Common operations

####1. View zkServer.sh Help information

[hadoop@hadoop-nn-01 conf]$ zkServer.sh help
ZooKeeper JMX enabled by default
Using config: /home/hadoop/zookeeper-3.4.14/bin/../conf/zoo.cfg
Usage: /home/hadoop/zookeeper-3.4.14/bin/zkServer.sh {start|start-foreground|stop|restart|status|upgrade|printcmd}

####2. Start / shut down / restart zk server

/home/hadoop/zookeeper-3.4.14/bin/zkServer.sh stop
/home/hadoop/zookeeper-3.4.14/bin/zkServer.sh start /home/hadoop/zookeeper-3.4.14/conf/zoo.cfg
/home/hadoop/zookeeper-3.4.14/bin/zkServer.sh restart /home/hadoop/zookeeper-3.4.14/conf/zoo.cfg

####3. View server status

zkServer.sh status

####4. View port

[hadoop@hadoop-nn-01 version-2]$ netstat -nltp | grep 2181
(Not all processes could be identified, non-owned process info
 will not be shown, you would have to be root to see it all.)
tcp        0      0 :::2181                     :::*                        LISTEN      37668/java 

3.3 common Shell commands of zookeeper

1 start service and connection service

#If the connection service does not specify a service address, it will connect to by default localhost:2181
zkCli.sh -server hadoop-nn-01:2181

2 help command

Use help to view all commands and formats.

3 view node list

There are two commands to view the node list: ls path and ls2 path. The latter is an enhancement of the former. You can not only view all nodes under the specified path, but also view the information of the current node.

4 new node

create [-s] [-e] path data acl   #Where - s is an ordered node and - e is a temporary node

Create node and write data:

create /test 123456

[zk: hadoop-nn-01:2181(CONNECTED) 10] create -s /test/a  "aaa"
Created /test/a0000000000
[zk: hadoop-nn-01:2181(CONNECTED) 11] create -s /test/b  "bbb"
Created /test/b0000000001
[zk: hadoop-nn-01:2181(CONNECTED) 12] create -s /test/c  "ccc"
Created /test/c0000000002

Create a temporary node, which will be deleted after the session expires:

[zk: hadoop-nn-01:2181(CONNECTED) 17] create -e /test/tmp  "tmp"
Created /test/tmp

5 view node

# format
get path [watch] 

[zk: hadoop-nn-01:2181(CONNECTED) 18] get /test
123456  #Node data
cZxid = 0x300000007
ctime = Thu Jun 18 16:09:14 CST 2020
mZxid = 0x300000007
mtime = Thu Jun 18 16:09:14 CST 2020
pZxid = 0x30000000b
cversion = 4
dataVersion = 0
aclVersion = 0
ephemeralOwner = 0x0
dataLength = 6
numChildren = 4

Each attribute of the node is shown in the following table. One of the important concepts is Zxid (ZooKeeper transaction ID). Every change of ZooKeeper node has a unique Zxid. If Zxid1 is smaller than Zxid2, the change of Zxid1 occurs before the change of Zxid2.

State properties	explain
cZxid	Transaction ID when data node is created
ctime	Time when the data node was created
mZxid	Transaction ID when the data node was last updated
mtime	Time when the data node was last updated
pZxid	Transaction ID of the data node's child node when it was last modified
cversion	Number of changes to child nodes
dataVersion	Number of changes to node data
aclVersion	Number of ACL changes for node
ephemeralOwner	If the node is a temporary node, it indicates the SessionID of the session that created the node; if the node is a persistent node, the property value is 0
dataLength	Length of data content
numChildren	Current number of child nodes of data node

6. View node status

You can use the stat command to see the status of a node, which returns values similar to the get command, but does not return node data.

[zk: hadoop-nn-01:2181(CONNECTED) 19] stat /test
cZxid = 0x300000007
ctime = Thu Jun 18 16:09:14 CST 2020
mZxid = 0x300000007
mtime = Thu Jun 18 16:09:14 CST 2020
pZxid = 0x30000000b
cversion = 4
dataVersion = 0
aclVersion = 0
ephemeralOwner = 0x0
dataLength = 6
numChildren = 4

7 update node

The command to update a node is set, which can be modified directly, as follows:

[zk: hadoop-nn-01:2181(CONNECTED) 22] set /test 345
cZxid = 0x300000007
ctime = Thu Jun 18 16:09:14 CST 2020
mZxid = 0x30000000d
mtime = Thu Jun 18 16:23:57 CST 2020
pZxid = 0x30000000b
cversion = 4
dataVersion = 1
aclVersion = 0
ephemeralOwner = 0x0
dataLength = 3
numChildren = 4

It can also be changed based on the version number. At this time, it is similar to the optimistic locking mechanism. When the data version you passed in does not match the data version number of the current node, zookeeper will reject this modification:

[zk: hadoop-nn-01:2181(CONNECTED) 23] set /test 345 0
version No is not valid : /test  #Invalid version number

[zk: hadoop-nn-01:2181(CONNECTED) 26] set /test 33345 1
cZxid = 0x300000007
ctime = Thu Jun 18 16:09:14 CST 2020
mZxid = 0x300000010
mtime = Thu Jun 18 16:26:01 CST 2020
pZxid = 0x30000000b
cversion = 4
dataVersion = 2
aclVersion = 0
ephemeralOwner = 0x0
dataLength = 5
numChildren = 4

1.7 deleting nodes

The syntax for deleting a node is as follows:

delete path [version]

Just like updating node data, you can also pass in version number. When the data version number you pass in does not match the data version number of the current node, zookeeper will not delete it.

[zk: hadoop-nn-01:2181(CONNECTED) 28] delete /test/a0000000000 1 
version No is not valid : /test/a0000000000   #Invalid version number
[zk: hadoop-nn-01:2181(CONNECTED) 29] delete /test/a0000000000 0

To delete a node and all its descendants, you can use recursive deletion with the command rmr path.

8 monitor

1. get path [watch]

A listener registered with get path [watch] can notify the client when the content of the node changes. It should be noted that the trigger of zookeeper is one-time trigger, that is, once triggered, it will fail immediately.

[zk: hadoop-nn-01:2181(CONNECTED) 31] get /test  watch
[zk: hadoop-nn-01:2181(CONNECTED) 32] set /test 45678 
WATCHER::
WatchedEvent state:SyncConnected type:NodeDataChanged path:/test #Change of node value

1. stat path [watch]

A listener registered with stat path [watch] can notify clients when the state of a node changes.

[zk: hadoop-nn-01:2181(CONNECTED) 33] stat /test watch 
[zk: hadoop-nn-01:2181(CONNECTED) 34] set /test 112233 
WATCHER::
WatchedEvent state:SyncConnected type:NodeDataChanged path:/test  #Change of node value

1. ls\ls2 path [watch]

A listener registered with ls path [watch] or ls2 path [watch] can listen to the addition and deletion of all the child nodes under the node.

[zk: hadoop-nn-01:2181(CONNECTED) 35] ls /test watch 
[b0000000001, tmp, c0000000002]
[zk: hadoop-nn-01:2181(CONNECTED) 36] create  /test/yarn "aaa" 
WATCHER::
Created /test/yarn
WatchedEvent state:SyncConnected type:NodeChildrenChanged path:/test

9 zookeeper four word command

command	Function description
conf	Print details of service configuration.
cons	Lists the full connection / session details for all clients connected to this server. Including the number of packets received / sent, session ID, operation delay, last operation, etc.
dump	Lists outstanding sessions and temporary nodes. This only applies to Leader nodes.
envi	Print details of the service environment.
ruok	Test whether the service is in the correct state. If it is correct, return "imok", otherwise, do nothing.
stat	Lists brief details of the server and connecting clients.
wchs	Lists simple information for all watch es.
wchc	Lists the details of the server watch by session.
wchp	Lists the details of the server watch by path.

[zk: hadoop-nn-01:2181(CONNECTED) 0] conf
ZooKeeper -server host:port cmd args
	stat path [watch]
	set path data [version]
	ls path [watch]
	delquota [-n|-b] path
	ls2 path [watch]
	setAcl path acl
	setquota -n|-b val path
	history 
	redo cmdno
	printwatches on|off
	delete path [version]
	sync path
	listquota path
	rmr path
	get path [watch]
	create [-s] [-e] path data acl
	addauth scheme auth
	quit 
	getAcl path
	close 
	connect host:port

4, Zookeeper automation script

################################################################################
# Function: monitor the process of zookeeper service
# Author : gzg
# Date : 2020-6-18
################################################################################
#!/bin/bash

current_dir=$(cd $(dirname $0);pwd)
cd ${current_dir}

source ~/.bashrc

now_time=$(date +"%Y-%m-%d %H:%M:%S")
ERR_FILE=/home/hadoop/log/zookeeper.log

function checkProcess()
{
    process=$(ps -ef | grep $1 | grep -Ev "grep|check_zookeeper_process.sh")
    
    if [[ "$process" != "" ]]; then
        echo "${now_time} $1 process exist."
        echo "$process"
        return 0
    fi

    echo "${now_time} java zk service restart" >> ${ERR_FILE}
    /home/hadoop/zookeeper-3.4.14/bin/zkServer.sh start /home/hadoop/zookeeper-3.4.14/conf/zoo.cfg
}

function main()
{
    checkProcess "org.apache.zookeeper.server.quorum.QuorumPeerMain"
}

main

Hang in crontab and start up automatically every time you use it

* * * * * /bin/bash /home/hadoop/scripts/check_zookeeper_process.sh > /dev/null 2>&1

reference resources:

Construction of Zookeeper single machine and cluster environment

Zookeeper common Shell commands

Blindly building Zookeeper distributed cluster

zookeeper zkServer.sh Orders zkCli.sh Command, four word command