【原創】大數據基礎之Ambari（5）經過Ambari部署Hue

ambari2.7.3（hdp3.1） 安裝 hue4.2

ambari的hdp中原生不支持hue安裝，下面介紹如何經過添加service的方式使ambari支持hue安裝：

官方：http://gethue.com/

Hue is an open source Workbench for developing and accessing Data Apps.

一 安裝SERVICE

第三方service：https://github.com/EsharEditor/ambari-hue-service

1 下載

VERSION=`hdp-select status hadoop-client | sed 's/hadoop-client - \([0-9]\.[0-9]\).*/\1/'`
rm -rf /var/lib/ambari-server/resources/stacks/HDP/$VERSION/services/HUE
sudo git clone https://github.com/EsharEditor/ambari-hue-service.git /var/lib/ambari-server/resources/stacks/HDP/$VERSION/services/HUE

2 修改

第三方service中安裝的是hue3.11，改成4.2，須要以下改動：

# cd /var/lib/ambari-server/resources/stacks/HDP/3.1/services/HUE
# sed -i 's/3.11.0/4.2.0/g' metainfo.xml
# sed -i 's/3.11.0/4.2.0/g' README.md
# sed -i 's/3.11.0/4.2.0/g' package/scripts/params.py
# sed -i 's/3.11.0/4.2.0/g' package/scripts/setup_hue.py

 

# vi package/scripts/params.py

--modify line

32 #download_url = 'cat /etc/yum.repos.d/HDP.repo | grep "baseurl" | awk -F \'=\' \'{print $2"hue/hue-4.2.0.tgz"}\''
33 download_url = 'echo "http://gethue.com/downloads/releases/4.2.0/hue-4.2.0.tgz"'

187 webhdfs_url = format('http://' + hdfs_httpfs_host + ':50070/webhdfs/v1')

--without yarn

:234,256s/^/#/g

 

# vi package/scripts/common.py

--add line 59,63,64

58 Execute('tar -zxvf hue.tgz -C {0} && rm -f hue.tgz'.format(params.hue_install_ dir))
59 Execute('mkdir -p {0}'.format(params.hue_dir))
60 # Ensure all Hue files owned by hue 61 Execute('chown -R {0}:{1} {2}'.format(params.hue_user,params.hue_group,params. hue_dir))
62 Execute('ln -s {0} /usr/hdp/current/hue-server'.format(params.hue_dir))
63 Execute('rm -rf {0}/*'.format(params.hue_dir))
64 Execute('cd /usr/local/hue-4.2.0 && make install')

--comment line

90 #services_configurations['httpfs-site'] = {}
91 #services_configurations['httpfs-site']['httpfs.proxyuser.hue.groups'] = '*'
92 #services_configurations['httpfs-site']['httpfs.proxyuser.hue.hosts'] = '*'

 

# vi package/scripts/setup_hue.py

--comment line

26 #Logger.info("Creating symlinks /usr/hdp/current/hadoop-client/lib/hue-plugins -4.2.0-SNAPSHOT.jar")
27 #Link("{0}/desktop/libs/hadoop/java-lib/*".format(params.hue_dir),to = "/usr/h dp/current/hadoop-client/lib")

 

# vi package/files/configs.sh

39 PASSWD="admin"

注意這一步

download_url = 'echo "http://gethue.com/downloads/releases/4.2.0/hue-4.2.0.tgz"'

hue-4.2.0.tgz須要手工下載並放到一個服務器可訪問的url上（好比本地的nginx），同時修改這裏的download_url

二 在ambari中部署

1 添加用戶

useradd hue

2 頁面部署安裝hue

安裝後的hue目錄位於/usr/local/hue/

3 修改配置

1）修改數據庫配置

# vi desktop/conf/hue.ini

  [[database]]

    # Database engine is typically one of:

    # postgresql_psycopg2, mysql, sqlite3 or oracle.

    #

    # Note that for sqlite3, 'name', below is a path to the filename. For other backends, it is the database name

    # Note for Oracle, options={"threaded":true} must be set in order to avoid crashes.

    # Note for Oracle, you can use the Oracle Service Name by setting "host=" and "port=" and then "name=<host>:<port>/<service_name>".

    # Note for MariaDB use the 'mysql' engine.

    ## engine=sqlite3

    ## host=

    ## port=

    ## user=

    ## password=

    # conn_max_age option to make database connection persistent value in seconds

    # https://docs.djangoproject.com/en/1.9/ref/databases/#persistent-connections

    ## conn_max_age=0

    # Execute this script to produce the database password. This will be used when 'password' is not set.

    ## password_script=/path/script

    ## name=desktop/desktop.db

修改engine、host、port、user、password、name

2）初始化數據庫

# build/env/bin/hue syncdb

其中會建立初試用戶名和密碼

# build/env/bin/hue migrate

3）若是要支持數據庫查詢好比mysql，找到

  [[databases]]

 

    # mysql, oracle, or postgresql configuration.

    ## [[[mysql]]]

      # Name to show in the UI.

      ## nice_name="My SQL DB"

 

      # For MySQL and PostgreSQL, name is the name of the database.

      # For Oracle, Name is instance of the Oracle server. For express edition

      # this is 'xe' by default.

      ## name=mysqldb

 

      # Database backend to use. This can be:

      # 1. mysql

      # 2. postgresql

      # 3. oracle

      ## engine=mysql

 

      # IP or hostname of the database to connect to.

      ## host=localhost

 

      # Port the database server is listening to. Defaults are:

      # 1. MySQL: 3306

      # 2. PostgreSQL: 5432

      # 3. Oracle Express Edition: 1521

      ## port=3306

 

      # Username to authenticate with when connecting to the database.

      ## user=example

 

      # Password matching the username to authenticate with when

      # connecting to the database.

      ## password=example

按照註釋添加便可，可添加多個

4）若是須要增長impala查詢，找到

[impala]

  # Host of the Impala Server (one of the Impalad)

  # server_host=localhost

 

  # Port of the Impala Server

  ## server_port=21050

注意這裏是21050端口，不是impala-shell鏈接的21000端口

5）能夠經過配置blacklist來隱藏一些組件的菜單

[desktop]

 app_blacklist=beeswax,impala,security,filebrowser,jobbrowser,rdbms,jobsub,pig,hbase,sqoop,zookeeper,metastore,spark,oozie,indexer

4 啓動

直接啓動

# build/env/bin/hue runcherrypyserver

後臺啓動supervisor（supervisor會監控並啓動hue server）

# nohup build/env/bin/supervisor start >> logs/hue.out 2>&1 &

日誌位於 /usr/local/hue/logs/ 下

三 使用

 

若是hue使用nginx代理，必定要設置Host，不然會報400錯誤：

location / {

  proxy_pass http://hue_backend;

  proxy_set_header Host hue.anything;

}

另外還能夠使用nginx來管理hue的靜態文件，加速訪問； 

 

參考：https://blog.csdn.net/zhouyuanlinli/article/details/83374416