CDH安装

大数据

准备工作

以下步骤要在所有节点上执行

1、修改hostname

vim /etc/sysconfig/network

2、关闭SELinux

若SELinux没有关闭，按照下述方式关闭
[root@BigDataMaster ~]# vi /etc/selinux/config
# This file controls the state of SELinux on the system.
# SELINUX= can take one of these three values:
#     enforcing - SELinux security policy is enforced.
#     permissive - SELinux prints warnings instead of enforcing.
#     disabled - No SELinux policy is loaded.
SELINUX=disabled
# SELINUXTYPE= can take one of these two values:
#     targeted - Targeted processes are protected,
#     mls - Multi Level Security protection.
SELINUXTYPE=targeted

3、安装软件包依赖关系

[root@BigDataNode04 ~]# yum install bind-utils chkconfig cyrus-sasl-gssapi cyrus-sasl-plain fuse fuse-libs gcc httpd init-functions libxslt mod_ssl MySQL-python openssl openssl-devel perl portmap psmisc python-psycopg2 python-setuptools sed service sqlite swig useradd zlib

关闭防火墙

service iptables stop

配置NTP服务（所有节点）（略）

修改hosts文件

把所有要添加到集群中的主机都要加入hosts中，格式如下：
127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6
172.16.132.160 BigDataMaster
172.16.132.161 BigDataMaster02
172.16.132.162 BigDataNode01
172.16.132.163 BigDataNode02
172.16.132.164 BigDataNode03
172.16.132.165 BigDataNode04
172.16.132.166 BigDataNode05
172.16.132.167 BigDataNode06
172.16.132.168 BigDataNode07
172.16.132.169 BigDataNode08

提前下载(本教程中，CM和CDH版本均为5.7.1)

软件的下载地址下载地址
下载CDH Parcel文件，只需要下载图中所示的3个文件

这里注意下，将.sha1文件后缀更改为.sha，同时把内容只保留hash码部分。

SSH免密登陆

1、生成一个key
[root@BigDataMaster ~]# ssh-keygen -t rsa
2、将这个key传到所有节点
[root@BigDataMaster ~]# ssh-copy-id -i .ssh/id_rsa.pub root@172.16.132.161

安装JAVA环境

[root@BigDataNode01 ~]# mkdir /usr/local/java && tar zxvf jdk-8u101-linux-x64.tar.gz -C /usr/local/java
[root@BigDataMaster parcel-repo]# vim /etc/profile
export JAVA_HOME=/usr/local/java/jdk1.8.0_101
export PATH=$PATH:$JAVA_HOME/bin
[root@BigDataMaster parcel-repo]# . /etc/profile

配置Agent

[root@BigDataMaster ~]# vim /opt/cm-5.7.6/etc/cloudera-scm-agent/config.ini
[General]
# Hostname of the CM server.
server_host=172.16.132.160
# Port that the CM server is listening on.
server_port=7182

优化建议

echo "vm.swappiness = 0" >>/etc/sysctl.conf && sysctl -p && echo never >/sys/kernel/mm/redhat_transparent_hugepage/defrag && echo "echo never >/sys/kernel/mm/redhat_transparent_hugepage/defrag" >>/etc/rc.local

Cloudera Manager安装

数据库安装

[root@BigDataMaster] tar zxvf mysql-5.6.27-linux-glibc2.5-x86_64.tar.gz -C /usr/local/
[root@BigDataMaster] mv mysql-5.6.27-linux-glibc2.5-x86_64/ mysql
[root@BigDataMaster] mkdir -p /data/mysql/mysql3306/{data,logs,tmp}
[root@BigDataMaster] vim my3306.cnf
#my.cnf
[client]
port            = 3306
socket          = /tmp/mysql.sock
[mysql]
prompt='(test)\u@\h [\d]>'
#tee=/data/mysql/mysql3306/data/query.log
no-auto-rehash
[mysqld]
#misc
user = mysql
basedir = /usr/local/mysql/
datadir = /data/mysql/mysql3306/data
port = 3306
socket = /tmp/mysql.sock
event_scheduler = 0
#tmp
tmpdir=/data/mysql/mysql3306/tmp
lower_case_table_names = 1
#timeout
interactive_timeout = 300
wait_timeout = 300
#character set
character-set-server = utf8
open_files_limit = 65535
max_connections = 100
max_connect_errors = 100000
#skip-name-resolve = 1
#logs
log-output=file
slow_query_log = 1
slow_query_log_file = slow.log
log-error = error.log
pid-file = mysql.pid
long_query_time = 1
#log-slow-admin-statements = 1
log-queries-not-using-indexes = 1
log-slow-slave-statements = 1
explicit_defaults_for_timestamp=1
#binlog
binlog_format = row
server-id = 2203306
log-bin = /data/mysql/mysql3306/logs/mysql-bin
binlog_cache_size = 4M
max_binlog_size = 1G
max_binlog_cache_size = 2G
sync_binlog = 1
innodb_flush_log_at_trx_commit = 1
expire_logs_days = 10
binlog_checksum = NONE   # for GR 
#relay log
relay-log = /data/mysql/mysql3306/logs/relay-bin
skip_slave_start = 1
max_relay_log_size = 1G
relay_log_purge = 0
relay_log_recovery = 1
log_slave_updates = 1
#slave-skip-errors=1032,1053,1062
# replication
relay_log_recovery = ON
master_info_repository = TABLE 
relay_log_info_repository = TABLE 
#buffers & cache
table_open_cache = 2048
table_definition_cache = 2048
table_open_cache = 2048
max_heap_table_size = 96M
sort_buffer_size = 2M
join_buffer_size = 2M
thread_cache_size = 256
query_cache_size = 0
query_cache_type = 0
query_cache_limit = 256K
query_cache_min_res_unit = 512
thread_stack = 192K
tmp_table_size = 96M
key_buffer_size = 8M
read_buffer_size = 2M
read_rnd_buffer_size = 16M
bulk_insert_buffer_size = 32M
#myisam
myisam_sort_buffer_size = 128M
myisam_max_sort_file_size = 10G
myisam_repair_threads = 1
#innodb
innodb_buffer_pool_size = 100M
innodb_buffer_pool_instances = 4
innodb_data_file_path = ibdata1:1G:autoextend
innodb_flush_log_at_trx_commit = 1
innodb_log_buffer_size = 64M
innodb_log_file_size = 500M
innodb_log_files_in_group = 3
innodb_max_dirty_pages_pct = 50
innodb_file_per_table = 1
innodb_rollback_on_timeout
innodb_status_file = 1
innodb_io_capacity = 2000
transaction_isolation = READ-COMMITTED
innodb_flush_method = O_DIRECT
[mysqld_safe]
#malloc-lib=/usr/local/mysql/lib/jmalloc.so
nice=-19
open-files-limit=65535
[root@BigDataMaster] /usr/local/mysql_install_db --defaults-file=/data/mysql/mysql3306/my3306.cnf

进入mysql并创建以下数据库：

1、首先需要去MySql的官网下载JDBC驱动，http://dev.mysql.com/downloads/connector/j/，解压后，找到mysql-connector-java-5.1.35-bin.jar，放到/opt/cm-5.7.6/share/cmf/lib/中。

[root@BigDataMaster mysql-connector-java-5.1.41]# cp mysql-connector-java-5.1.41-bin.jar /opt/cm-5.7.6/share/cmf/lib/

2、在主节点初始化CM5的数据库：

mysql> CREATE USER 'scm'@'%' IDENTIFIED BY 'scm';
Query OK, 0 rows affected (0.01 sec)
mysql>  grant all privileges on *.* to 'scm'@'%'  with grant option;
Query OK, 0 rows affected (0.01 sec)
mysql>  grant all privileges on *.* to 'root'@'%'  with grant option;
Query OK, 0 rows affected (0.00 sec)
mysql> flush privileges;
Query OK, 0 rows affected (0.00 sec)
mysql> create database hive DEFAULT CHARSET utf8 COLLATE utf8_general_ci;

初始化

[root@BigDataMaster schema]#  /opt/cm-5.7.6/share/cmf/schema/scm_prepare_database.sh mysql  -hlocalhost -uroot -p123456 --scm-host localhost scm scm scm
JAVA_HOME=/usr/local/java/jdk1.8.0_101
Verifying that we can write to /opt/cm-5.7.6/etc/cloudera-scm-server
Creating SCM configuration file in /opt/cm-5.7.6/etc/cloudera-scm-server
groups: cloudera-scm：无此用户
Executing:  /usr/local/java/jdk1.8.0_101/bin/java -cp /usr/share/java/mysql-connector-java.jar:/usr/share/java/oracle-connector-java.jar:/opt/cm-5.7.6/share/cmf/schema/../lib/* com.cloudera.enterprise.dbutil.DbCommandExecutor /opt/cm-5.7.6/etc/cloudera-scm-server/db.properties com.cloudera.cmf.db.
[                          main] DbCommandExecutor              INFO  Successfully connected to database.
All done, your SCM database is configured correctly!

安装

把cm传到各个节点

[root@BigDataNode04 ~]# scp -r /opt/cm-5.7.6/ root@172.16.132.169:/opt/

启动sever端

[root@BigDataNode04 ~]# /opt/cm-5.7.6/etc/init.d/cloudera-scm-server start

各个节点启动agent端

[root@BigDataNode04 ~]# /opt/cm-5.7.6/etc/init.d/cloudera-scm-agent start

登录

选择版本（本次选择免费版）

各节点安装CDH集群

选择CDH版本

集群安装

主机检查

集群设置，选择服务

自定义服务

自定义角色服务

出现问题

据说此问题是BUG，修复方法
错误日志

-yarn/stacks', u'bytes_free_warning_threshhold_bytes': 0, u'group': u'hadoop', u'user': u'yarn', u'mode': 493}, u'cpu': None, u'contents': None
Traceback (most recent call last):
  File "/opt/cm-5.7.6/lib64/cmf/agent/build/env/lib/python2.6/site-packages/cmf-5.7.6-py2.6.egg/cmf/agent.py", line 1605, in handle_heartbeat_p
    new_process.activate()
  File "/opt/cm-5.7.6/lib64/cmf/agent/build/env/lib/python2.6/site-packages/cmf-5.7.6-py2.6.egg/cmf/agent.py", line 3144, in activate
    self.write_process_conf()
  File "/opt/cm-5.7.6/lib64/cmf/agent/build/env/lib/python2.6/site-packages/cmf-5.7.6-py2.6.egg/cmf/agent.py", line 3251, in write_process_conf
    "source_parcel_environment", env))
  File "/opt/cm-5.7.6/lib64/cmf/agent/build/env/lib/python2.6/site-packages/cmf-5.7.6-py2.6.egg/cmf/util.py", line 373, in source
    raise e
ValueError: dictionary update sequence element #93 has length 1; 2 is required

修复方法：

[root@BigDataNode06 cloudera-scm-agent]# vim /opt/cm-5.7.6/lib64/cmf/agent/build/env/lib/python2.6/site-packages/cmf-5.7.6-py2.6.egg/cmf/util.py
#原样子
      LOG.warn("None environment value: %s=%s" % (k, v,))
  pipe = subprocess.Popen(['/bin/bash', '-c', ". %s; %s; env" % (path, command)],
                          stdout=subprocess.PIPE, env=caller_env)
  data = pipe.communicate()[0]
  try:
#修复
  pipe = subprocess.Popen(['/bin/bash', '-c', ". %s; %s; env| grep -v { | grep -v }" % (path, command)],
                          stdout=subprocess.PIPE, env=caller_env)
  data = pipe.communicate()[0]
  try:
    return dict((line.split("=", 1) for line in data.splitlines()))
  except Exception, e:

通过CM启动hive报错.org.apache.hadoop.hive.metastore.HiveMetaException: Failed to load driver

[root@BigDataNode04 tools]# cp mysql-connector-java-5.1.41-bin.jar /opt/cloudera/parcels/CDH/lib/hive/lib/

最好从已经运行的slave机器上scp对应的文件夹，不用修改server_host了，但是需要把对应的[/opt/cm-5.7.0/lib/cloudera-scm-agent]下的文件全部删除；同时删除［/opt/cm-5.7.0/log/cloudera-scm-agent］目录下的所有内容。

安装完成