CDH5.12.1 安装部署

###通过http://192.168.50.200:7180/cmf/login 访问CM控制台

4.CDH安装

4.1CDH集群安装向导

1.admin/admin登陆到CM

2.同意license协议，点击继续

3.选择60试用，点击继续

4.点击“继续”

5.输入主机IP或者名称，点击搜索找到主机名后点击继续

6.点击“继续”

7.使用parcel选项，点击“更多选项”，点击“-”删除其他所有的地址，输入http://ip-192-168-50-200.hns.com/cdh5.12.1/点击“保存更改”

8.选择自定义存储库，输入cm的http地址

9.点击“继续”，进入下一步安装jdk

10.点击“继续”，进入下一步，默认多用户模式

11.点击“继续”，进入下一步配置ssh账号密码：

12.点击“继续”，进入下一步，安装Cloudera Manager相关到各个节点

13.点击“继续”，进入下一步安装cdh到各个节点

14.点击“继续”，进入下一步主机检查，确保所有检查项均通过

上述的解决方案：

在每台机器上执行如下操作：

[root@ip-192-168-50-200 ~]# echo never > /sys/kernel/mm/transparent_hugepage/enabled
[root@ip-192-168-50-200 ~]# echo never > /sys/kernel/mm/transparent_hugepage/defrag

[root@ip-192-168-50-200 ~]# echo "vm.swappiness = 10" >> /etc/sysctl.conf

[root@ip-192-168-50-200 ~]# sysctl -p

点击完成进入服务安装向导！！！

4.2 集群设置安装向导

1.选择需要安装的服务，此处使用自定义服务，如下图

2.点击“继续”，进入集群角色分配

HDFS角色分配：

Hive角色分配：

Cloudera Manager Service 角色分配：

Spark角色分配：（Spark on Yarn 所以没有spark的master和worker 角色）

Yarn角色分配：

Zookeeper角色分配：（至少3个Server）

3.角色分配完成点击“继续”，进入下一步，测试数据库连接

4.测试成功，点击“继续”，进入目录设置，此处使用默认默认目录，根据实际情况进行目录修改

5.点击“继续”，等待服务启动成功！！！

6.点击“继续”，显示集群安装成功！

7.安装成功后，进入home管理界面

5.快速组建服务验证

5.1HDFS验证（mkdir+put+cat +get）

mkdir操作：

put 操作：

cat 操作：

get 操作：

5.2 Hive 验证

使用hive命令行操作

hive> create external table test_table(
    > s1 string,
    > s2 string
    > )row format delimited fields terminated by ','
    > stored as textfile location '/hns/test';
OK
Time taken: 0.074 seconds
hive> show tables;
OK
test_table
Time taken: 0.012 seconds, Fetched:  row(s)

hive> select * from test_table;
OK
1 test
2 hns
3 zhangsan
Time taken: 0.054 seconds, Fetched: 3 row(s)
hive>

hive> insert into test_table values("","lisi");
Query ID = hdfs_20181013220202_823a17d7-fb58-40e9-bf33-11f44d0de10a
Total jobs =
Launching Job  out of
Number of reduce tasks is set to  since there's no reduce operator
Starting Job = job_1539418452562_0003, Tracking URL = http://ip-192-168-50-200.hns.com:8088/proxy/application_1539418452562_0003/
Kill Command = /opt/cloudera/parcels/CDH-5.12.-.cdh5.12.1.p0./lib/hadoop/bin/hadoop job  -kill job_1539418452562_0003
Hadoop job information for Stage-: number of mappers: ; number of reducers:
-- ::, Stage- map = %,  reduce = %
-- ::, Stage- map = %,  reduce = %, Cumulative CPU 0.93 sec
MapReduce Total cumulative CPU time:  msec
Ended Job = job_1539418452562_0003
Stage- is selected by condition resolver.
Stage- is filtered out by condition resolver.
Stage- is filtered out by condition resolver.
Moving data to: hdfs://ip-192-168-50-200.hns.com:8020/hns/test/.hive-staging_hive_2018-10-13_22-02-31_572_2687237229927791201-1/-ext-10000
Loading data to table default.test_table
Table default.test_table stats: [numFiles=, numRows=, totalSize=, rawDataSize=]
MapReduce Jobs Launched:
Stage-Stage-: Map:    Cumulative CPU: 0.93 sec   HDFS Read:  HDFS Write:  SUCCESS
Total MapReduce CPU Time Spent:  msec
OK
Time taken: 19.016 seconds
hive> select * from test_table;
OK
       lisi
       test
       hns
       zhangsan
Time taken: 0.121 seconds, Fetched:  row(s)
hive> 

Hive MapReduce操作：

hive> select count(*) from test_table;
Query ID = hdfs_20181013220606_1011d0ce-9ddd-43ec-a103-18b3a32ea292
Total jobs =
Launching Job  out of
Number of reduce tasks determined at compile time:
In order to change the average load for a reducer (in bytes):
  set hive.exec.reducers.bytes.per.reducer=<number>
In order to limit the maximum number of reducers:
  set hive.exec.reducers.max=<number>
In order to set a constant number of reducers:
  set mapreduce.job.reduces=<number>
Starting Job = job_1539418452562_0004, Tracking URL = http://ip-192-168-50-200.hns.com:8088/proxy/application_1539418452562_0004/
Kill Command = /opt/cloudera/parcels/CDH-5.12.-.cdh5.12.1.p0./lib/hadoop/bin/hadoop job  -kill job_1539418452562_0004
Hadoop job information for Stage-: number of mappers: ; number of reducers:
-- ::, Stage- map = %,  reduce = %
-- ::, Stage- map = %,  reduce = %, Cumulative CPU 1.12 sec
-- ::, Stage- map = %,  reduce = %, Cumulative CPU 2.28 sec
MapReduce Total cumulative CPU time:  seconds  msec
Ended Job = job_1539418452562_0004
MapReduce Jobs Launched:
Stage-Stage-: Map:   Reduce:    Cumulative CPU: 2.28 sec   HDFS Read:  HDFS Write:  SUCCESS
Total MapReduce CPU Time Spent:  seconds  msec
OK

Time taken: 24.471 seconds, Fetched:  row(s)

5.3 MapReduce 验证：

[hdfs@ip---- hadoop-mapreduce]$ pwd
/opt/cloudera/parcels/CDH/lib/hadoop-mapreduce

[hdfs@ip---- hadoop-mapreduce]$ hadoop jar hadoop-mapreduce-examples.jar pi
Number of Maps  =
Samples per Map =
Wrote input for Map #
Wrote input for Map #
Wrote input for Map #
Wrote input for Map #
Wrote input for Map #
Starting Job
.
.
.
// :: INFO mapreduce.Job: Running job: job_1539418452562_0005
// :: INFO mapreduce.Job: Job job_1539418452562_0005 running in uber mode : false
// :: INFO mapreduce.Job:  map % reduce %
// :: INFO mapreduce.Job:  map % reduce %
// :: INFO mapreduce.Job:  map % reduce %
// :: INFO mapreduce.Job:  map % reduce %
// :: INFO mapreduce.Job:  map % reduce %
// :: INFO mapreduce.Job: Job job_1539418452562_0005 completed successfully
// :: INFO mapreduce.Job: Counters:
        File System Counters
.
.
.

5.4 Spark 验证

scala> val testFile=sc.textFile("hdfs://ip-192-168-50-200.hns.com:8020/hns/test/a.txt")
testFile: org.apache.spark.rdd.RDD[String] = hdfs://ip-192-168-50-200.hns.com:8020/hns/test/a.txt MapPartitionsRDD[1] at textFile at <console>:27

scala> testFile.count()
res2: Long =