Solr 4.10.1 + tomcat 7 + zookeeper + HDFS 集成 （SolrCloud+HDFS）配置

关于搭建solrCloud + zookeeper +HDFS的实践经验

 

 

搭建zookeeper：

1、        到zookeeper官网下载zookeeper（用于管理solrcloud云的配置文件）http://mirror.bit.edu.cn/apache/zookeeper/zookeeper-3.4.6/

2、        准备三台服务器,或者搭建3台虚拟机:                                                 例如：host3.com   192.168.2.87    

Host5.com   192.168.2.89

Host4.com   192.168.2.94

3、        上传zookeeper-3.4.6.tar.gz到任意一台服务器/usr/local/目录，并解压到当前目录：zookeeper： tar –zxvf zookeeper-3.4.6.tar.gz                            改名：zookeeper-3.4.6为 zookeeper：mv  zookeeper-3.4.6   zookeeper

4、        在zookeeper目录下建立data和 logs目录，同时将zookeeper目录下conf/zoo_simple.cfg文件复制一份成 zoo.cfg

5、        修改zoo.cfg  

# 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 passbetween

# sending a request and getting anacknowledgement

syncLimit=5

# the directory where the snapshot isstored.

# do not use /tmp for storage, /tmp hereis just

# example sakes.

dataDir=/usr/local/zookeeper/data  

# the port at which the clients willconnect

clientPort=2181

# the maximum number of clientconnections.

# increase this if you need to handle moreclients

#maxClientCnxns=60

#

# Be sure to read the maintenance sectionof the

# administrator guide before turning onautopurge.

#

#http://zookeeper.apache.org/doc/current/zookeeperAdmin.html#sc_maintenance

#

# The number of snapshots to retain indataDir

#autopurge.snapRetainCount=3

# Purge task interval in hours

# Set to "0" to disable autopurge feature

#autopurge.purgeInterval=1

dataLogDir=/usr/local/zookeeper/logs

server.1=192.168.2.89:2888:3888

server.2=192.168.3.87:2888:3888

server.3=192.168.3.94:2888:3888

 

6、    拷贝zookeeper目录到另外两台服务器：                                                                    

  

scp-r /usr/local/zookeeper [email protected]:/usr/local/</a>

scp–r /usr/local/zookeeper [email protected]:/usr/local/</a>

 

分别在几台服务器的data目录下建立myid 其 ip对应相应的server.*   server.1 的myid内容为1  server.2的myid内容为2 server.3的myid为 3  

7、        启动ZooKeeper集群，在每个节点上分别启动ZooKeeper服务：

cd/usr/local/zookeeper/

bin/zkServer.sh  start

8、  可以查看ZooKeeper集群的状态，保证集群启动没有问题：分别查看每台服务器的zookeeper状态 zookeeper#bin/zkServer.shstatus查看那些是following那个是leader

Eg:  

[[email protected]]# bin/zkServer.sh status

JMXenabled by default

Usingconfig: /home/hadoop/zookeeper-3.3.6/bin/../conf/zoo.cfg

Mode:follower

[root@host5/]# cd /home/hadoop/zookeeper-3.3.6/

[[email protected]]# bin/zkServer.sh status

JMXenabled by default

Usingconfig: /home/hadoop/zookeeper-3.3.6/bin/../conf/zoo.cfg

Mode:leader

 

[root@host3multicore]# cd /home/hadoop/zookeeper-3.3.6/

[[email protected]]# bin/zkServer.sh status

JMXenabled by default

Usingconfig: /home/hadoop/zookeeper-3.3.6/bin/../conf/zoo.cfg

Mode:follower

 

建立SolrCloud：

1、        到Apache官网下载solr安装文件 solr-4.10.1.tgz    http://mirror.bit.edu.cn/apache/lucene/solr/4.10.1/                            并解压tar –xvzf solr-4.10.1.tgz                                                         更改solr-4.10.1目录：mv solr-4.10.1 solr

2、         创建在根目录下创建目录mkdir –p /data0/solrcloud    将solr目录移动到data0下，那么data0下包含两个目录{solr、solrcloud}

3、        将/data0/solr/example/webapps/solr.war放到tomcat的webapps目录下，启动tomcat，这是tomcat下多出solr目录

4、        将/data0/solr/example/lib/ext/下的所有的jar文件复制到tomcat/webapps/solr/WEB-INF/lib/下，建立mkdir –p /data0/solrcloud/{multicore，solr-lib}目录，在将tomcat/webapps/solr/WEB-INF/lib/* 复制一份到solr-lib/下 

cp  /usr/local/tomcat/webapps/solr/WEB-INF/lib/*  /data0/solrcloud/solr-lib/

 

 

 

5、        在建立一个装在配置文件的目录

 

mkdir–p  /data0/solrcloud/multicore/collection/{conf,data}  

 

同时将 /data0/solr/ example/solr/collection1/conf/* 

复制到/data0/solrcloud/multicore/collection/conf目录下

 

将example/solr/multicore下的solr.xml和zoo.cfg复制到/data0/solrcloud/multicore目录下 eg：

 

 

Collection目录：

 

 Data-config.xml文件是数据导入的配置文件查看：

 

http://blog.csdn.net/john_hongming/article/details/40181451

 

需要自己建立的 solrcore.properties文件

文件内容：

solr.shard.data.dir=/data0/solrcloud/multicore/collection/data

说明：属性solr.shard.data.dir在solrconfig.xml文件中被引用过，指定索引数据的存放位置。

 

Solr.xml文件的说明：

 

 

 

6、        通过zookeeper管理配置文件：

#zookeeper上传配置文件 #

java-classpath .:/data0/solrcloud/solr-lib/* org.apache.solr.cloud.ZkCLI -cmdupconfig -zkhost host3.com:2181,host4.com:2181,host5.com:2181 -confdir/data0/solrcloud/multicore/collection/conf -confname myconf

#zookeeper连接多个节点#

java-classpath .:/data0/solrcloud/solr-lib/* org.apache.solr.cloud.ZkCLI -cmdlinkconfig -collection collection1 -confname myconf -zkhosthost3.com:2181,host4.com:2181,host5.com:2181

 

7、        在启动文件tomcat/bin/Catalina.sh中添加如下配置：

#配置tomcat/bin/catlina.sh下的启动参数#

JAVA_OPTS="-server -Xmx2048m-Xms1024m -verbose:gc  -Xloggc:solr_gc.log -Dsolr.solr.home=/data0/solrcloud/multicore 

-DzkHost=host3.com:2181,host4.com:2181,host5.com:2181"

 

8、        修改tomcat/webapps/solr/WEB-INF/web.xml

context.xml

      

<!-- The contents of this file will be loaded for each web application -->
<Context docBase="solr.war" debug="0" crossContext="false">

    <!-- Default set of monitored resources -->
    <WatchedResource>WEB-INF/web.xml</WatchedResource>

    <!-- Uncomment this to disable session persistence across Tomcat restarts -->
    <!--
    <Manager pathname="" />
    -->

    <!-- Uncomment this to enable Comet connection tacking (provides events
         on session expiration as well as webapp lifecycle) -->
    <!--
    <Valve className="org.apache.catalina.valves.CometConnectionManagerValve" />
    -->
<Environment name="solr/home" type="java.lang.String" value="/data0/solrcloud/multicore" override="true"/>
</Context>

web.xml

<env-entry>
       <env-entry-name>solr/home</env-entry-name>
       <env-entry-value>/data0/solrcloud/multicore</env-entry-value>
       <env-entry-type>java.lang.String</env-entry-type>
</env-entry>

 

<env-entry>
       <env-entry-name>solr/home</env-entry-name>
       <env-entry-value>/data0/solrcloud/multicore</env-entry-value>
       <env-entry-type>java.lang.String</env-entry-type>
    </env-entry>

 

9、        Solrcloud的IK分词器的配置：

首先在solrcloud中的multicore/collection/下建立lib目录，将IK分词器的配置文件最主要的就是IKAnalyzer.cfg.xml和 stopword.dic移动到lib下

之后修改multicore/collection/conf下的schema.xml文件

添加：

<fieldTypename="ikanalyzer"class="solr.TextField">

<analyzertype="index"isMaxWordLength="false"class="org.wltea.analyzer.lucene.IKAnalyzer"/>

<analyzertype="query"isMaxWordLength="true"class="org.wltea.analyzer.lucene.IKAnalyzer"/>

<analyzertype="multiterm">

<tokenizerclass="solr.KeywordTokenizerFactory"/>

</analyzer>

</fieldType>

Field会根据type的属性进行分词

至此，IKAnalyzer中文分词基本添加完成，更新下zookeeper的solr配置：

java -classpath .:/usr/local/solrcloud/solr-lib/*org.apache.solr.cloud.ZkCLI -cmd upconfig -zkhost 192.168.3.119:2181,192.168.3.111:2181,192.168.3.127:2181 -confdir/usr/local/solrcloud/multicore/collection/conf -confname myconf

 

如果要添加扩展词典：

在tomcat/webapps/solr/WEB-INF/下建立classes目录将要添加的词典和配置文件都放在该目录下

 

 

 

编辑IKAnalyzer.cfg.xml 添加词典

如下图：

10、     将配置好的data0目录 scp到另外两个服务器：

scp –r /data0 [email protected]:/

scp –r /[email protected]:/

 

scp –r /usr/local/tomcat [email protected]:/usr/local/

scp –r /usr/local/tomcat [email protected]:/usr/local/ 

启动三台服务器bin/startup.sh start

 

11、     创建collection和shard

 

#创建collection 分3片一个副本集#

#创建三个分片，每个分片一个副本集#

curl 'http://192.168.2.89:8080/solr/admin/collections?action=CREATE&name=mycollection&numShards=3&replicationFactor=1'    

 #创建shard 的副本  在89创建shard1的副本集mycollection_shard1_replica_2#

curl'http://192.168.2.89:8080/solr/admin/cores?action=CREATE&collection=mycollection&name=mycollection_shard1_replica_2&shard=shard1'

curl'http://192.168.2.87:8080/solr/admin/cores?action=CREATE&collection=mycollection&name=mycollection_shard1_replica_3&shard=shard1'

curl'http://192.168.2.89:8080/solr/admin/cores?action=CREATE&collection=mycollection&name=mycollection_shard2_replica_2&shard=shard2'

curl'http://192.168.2.87:8080/solr/admin/cores?action=CREATE&collection=mycollection&name=mycollection_shard2_replica_3&shard=shard2'

#对shard1再次在94上shard #

curl'http://192.168.2.94:8080/solr/admin/collections?action=SPLITSHARD&collection=mycollection&shard=shard1'

 

 

 

 

 

 

 

1、         编辑/usr/local/tomcat/bin/catlina.sh 添加红色部分

 

JAVA_OPTS="-server-Xmx2048m -Xms1024m -verbose:gc -Xloggc:solr_gc.log <span style="color:#ff0000;">-XX:MaxDirectMemorySize=1g -Dsolr.directoryFactory=HdfsDirectoryFactory-Dsolr.lock.type=hdfs -Dsolr.hdfs.home=hdfs://host1xyz.com:9000/solr </span>-Dsolr.solr.home=/data0/solrcloud/multicore-DzkHost=host3.com:2181,host4.com:2181,host5.com:2181"

2、        修改/data0/solrcloud/multicore/collection/conf/solrconfig.xml文件

添加这部分：

<directoryFactoryname="DirectoryFactory"class="solr.HdfsDirectoryFactory">    

   <strname="solr.hdfs.home">hdfs://host1xyz.com:9000/solr</str>  

   <boolname="solr.hdfs.blockcache.enabled">true</bool>

   <int name="solr.hdfs.blockcache.slab.count">1</int>

   <boolname="solr.hdfs.blockcache.direct.memory.allocation">true</bool>

   <intname="solr.hdfs.blockcache.blocksperbank">16384</int>

   <boolname="solr.hdfs.blockcache.read.enabled">true</bool>

   <bool name="solr.hdfs.blockcache.write.enabled">true</bool>

   <boolname="solr.hdfs.nrtcachingdirectory.enable">true</bool>

   <intname="solr.hdfs.nrtcachingdirectory.maxmergesizemb">16</int>

   <intname="solr.hdfs.nrtcachingdirectory.maxcachedmb">192</int>

    <str name="solr.hdfs.confdir">/home/hadoop/hadoop-2.2.0/etc/hadoop</str>

  </directoryFactory>

 

 

再找到

<lockType>${solr.lock.type:native}</lockType>将其修改为<lockType>${solr.lock.type:hdfs}</lockType>

 

注意：这时<dataDir>${solr.data.dir:}</dataDir>是这种状态，如果添加路径，就会覆盖掉hdfs的路径

 

参考：

https://cwiki.apache.org/confluence/display/solr/Running+Solr+on+HDFS

http://shiyanjun.cn/archives/100.html

http://blog.csdn.net/shirdrn/article/details/9770829

http://blog.csdn.net/john_hongming/article/details/40113641

http://blog.csdn.net/john_hongming/article/details/40080947