实验说明:
本实验目的是通过使用Oracle的HDFS直接连接器从oracle数据库的外部表中直接访问HDFS的数据文件。支持的数据文件格式取决于ORACLE_LOADER的驱动程序。
一、 安装准备
1. 软件下载:
Oracle Big Data Connectors:ODCH
http://www.oracle.com/technetwork/bdc/big-data-connectors/downloads/index.html
Jave SE 下载地址:
http://www.oracle.com/technetwork/java/javase/downloads/jdk6u38-downloads-1877406.html
Oracle11g下载地址:
http://www.oracle.com/technetwork/database/enterprise-edition/downloads/index.html
Oracle Enterprise Linux下载地址:需要注册oracle帐号才能下载
https://edelivery.oracle.com/linux
2. 安装环境
软件版本:
OS:Oracle Enterprise Linux 5.6
Oracle DB:Oracle 11.2.0.2
Jave SE:jdk1.6.0_18
ODCH:Oracle SQL Connector for Hadoop Distributed File System Release 2.0 for Linux
Hadoop:hadoop-0.20.2
Hadoop\Oracle架构:
主机名 |
IP |
系统版本 |
Oracle node |
Hadoop node |
hadoop进程名 |
gc |
192.168.2.100 |
OEL5.6 |
Oracle11.2.0.2 |
master |
namenode,jobtracker |
rac1 |
192.168.2.101 |
OEL5.6 |
/ |
slave |
datanode,tasktracker |
rac2 |
192.168.2.102 |
OEL5.6 |
/ |
slave |
datanode,tasktracker |
主要环境变量:
变量名 |
变量值 |
ORACLE_HOME |
/u01/app/ora11g/product/11.2.0/db_1 |
ORACLE_SID |
gcdb |
HADOOP_HOME |
/home/grid/hadoop-0.20.2 |
HADOOP_CONF_DIR |
$HADOOP_HOME/conf |
OSCH_HOME |
/opt/odch/orahdfs-2.0.0 |
ORAHDFS_JAR |
$OSCH_HOME/jlib/orahdfs.jar |
HDFS_BIN_PATH |
/opt/odch/orahdfs-2.0.0 |
HADOOP_CLASSPATH |
$OSCH_HOME/jlib/* |
3. 软件安装
linux和oracle安装见:
http://f.dataguru.cn/thread-55231-1-1.html
Hadoop安装见:
http://blog.csdn.net/lichangzai/article/details/8206834
二、 通过ODCH把HDFS->ORACLE
1. 上传odch文件并解压
[grid@gc odch]$ pwd
/opt/odch
[grid@gc odch]$ unzip orahdfs-2.0.0.zip
2. 修改环境变量
--添加hadoop\oracle\ODCH主要要环境变量
[grid@gc ~]$ vi .bash_profile
export ORACLE_BASE=/u01/app/ora11g
export ORACLE_HOME=$ORACLE_BASE/product/11.2.0/db_1
export ORACLE_SID=gcdb
export PATH=/usr/sbin:$PATH
export PATH=$ORACLE_HOME/bin:$PATH
export LD_LIBRARY_PATH=$ORACLE_HOME/lib:/lib64:/usr/lib64:/usr/local/lib64:/usr/X11R6/lib64/
export CLASSPATH=$ORACLE_HOME/JRE:$ORACLE_HOME/jlib:$ORACLE_HOME/rdbms/jlib
export HADOOP_HOME=/home/grid/hadoop-0.20.2
PATH=$PATH:$HOME/bin:/usr/java/jdk1.6.0_18/bin: $HADOOP_HOME/bin
export HADOOP_CONF_DIR=$HADOOP_HOME/conf
export OSCH_HOME=/opt/odch/orahdfs-2.0.0
export ORAHDFS_JAR=$OSCH_HOME/jlib/orahdfs.jar
export HDFS_BIN_PATH=$OSCH_HOME/bin
export JAVA_HOME=/usr
export PATH
3. 修改用户属组
说明:因为之前的hadoop安装在grid系统用户下,而oracle安装在了oracle系统用户下,为了操作方便,使grid用户有操作oracle操作库的权限,现在修改grid系统用户属组,添加到和oracle同样的用户组下。
--查看之前的用户属组
[root@gc ~]# id grid
uid=501(grid) gid=54326(hadoop) groups=54326(hadoop)
[root@gc ~]# id oracle
uid=500(oracle) gid=54321(oinstall) groups=54321(oinstall),54322(dba),54323(oper),54324(asmadmin)
--修改用户属组
[root@gc ~]# vi /etc/group
oinstall:x:54321:grid
dba:x:54322:oracle,grid
oper:x:54323:oracle,grid
asmadmin:x:54324:oracle,grid
--再次查看用户属组
[root@gc ~]# id grid
uid=501(grid) gid=54326(hadoop) groups=54326(hadoop),54322(dba),54321(oinstall),54324(asmadmin),54323(oper)
[root@gc ~]# id oracle
uid=500(oracle) gid=54321(oinstall) groups=54321(oinstall),54322(dba),54323(oper),54324(asmadmin)
注意:添加用户的用户组后,可能需要启系统才能真正生效,虽然id命令已经显示在所在的用户组。
--测试grid用户能访问oracle
[grid@gc ~]$ sqlplus "/as sysdba"
SQL*Plus: Release 11.2.0.2.0 Production on 星期三 1月 23 03:36:21 2013
Copyright (c) 1982, 2010, Oracle. All rights reserved.
连接到:
Oracle Database 11g Enterprise Edition Release 11.2.0.2.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
SQL> show parameter db_name;
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
db_name string gcdb
4. 配置hdfs_stream script文件
--添加以下内容
[grid@gc ~]$ vi ${OSCH_HOME}/bin/hdfs_stream
export HADOOP_HOME=/home/grid/hadoop-0.20.2
export OSCH_HOME=/opt/odch/orahdfs-2.0.0
export PATH=/usr/bin:/bin:$HADOOP_HOME/bin
5. 确保oracle用户对$ODCH_LOG_DIR/logs目录有读写权限
因为Oracle用户需要在 {$ODCH_LOG_DIR} 目录中创建log/bad文件,所以要确保oracle用户对$ODCH_LOG_DIR/log目录有读写权限
--修改目录权限,测试读写文件
[root@gc opt]# chmod -R 777 odch/
[root@gc opt]# su - oracle
[oracle@gc ~]$ cd /opt/odch/orahdfs-2.0.0/log/
[oracle@gc log]$ touch ora_access_test
[oracle@gc log]$ rm ora_access_test
6. 配置操作系统目录和数据库的Directory对象
--创建所用的系统目录
[root@gc ~]# mkdir -p /opt/odch/orahdfs-2.0.0/logs
[root@gc ~]# mkdir -p /opt/odch/orahdfs-2.0.0/extdir
[root@gc ~]# chmod 777 /opt/odch/orahdfs-2.0.0/logs
[root@gc ~]# chmod 777 /opt/odch/orahdfs-2.0.0/extdir
--创建oracle Directory对象
[root@gc ~]# su - grid
[grid@gc ~]$ sqlplus "/as sysdba"
SQL*Plus: Release 11.2.0.2.0 Production on 星期三 1月 23 05:35:48 2013
Copyright (c) 1982, 2010, Oracle. All rights reserved.
连接到:
Oracle Database 11g Enterprise Edition Release 11.2.0.2.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
SQL> --可以放在命令行批量执行下面的语句
create or replace directory ODCH_LOG_DIR as '/opt/odch/orahdfs-2.0.0/logs';
grant read, write on directory ODCH_LOG_DIR to SCOTT;
create or replace directory ODCH_DATA_DIR as '/opt/odch/orahdfs-2.0.0/extdir';
grant read, write on directory ODCH_DATA_DIR to SCOTT;
create or replace directory HDFS_BIN_PATH as '/opt/odch/orahdfs-2.0.0/bin';
grant execute on directory HDFS_BIN_PATH to SCOTT;
grant read, write on directory HDFS_BIN_PATH to SCOTT;
--目录对象说明
HDFS_BIN_PATH::hdfs_stream脚本所在目录.
HDFS_DATA_DIR:用来存放“位置文件”(location files)的目录。“位置文件”(location files) 是一个配置文件,里面包含HDFS的文件路径/文件名以及文件编码格式。
ODCH_LOG_DIR:Oracle用来存放外部表的log/bad等文件的目录.
7. 创建oracle外部表
SQL> conn scott/tiger
已连接。
SQL> --可以放在命令行批量执行下面的语句
CREATE TABLE admin_ext_employees
(employee_id NUMBER(4),
first_name VARCHAR2(20),
last_name VARCHAR2(25),
job_id VARCHAR2(10),
manager_id NUMBER(4),
hire_date DATE,
salary NUMBER(8,2),
commission_pct NUMBER(2,2),
department_id NUMBER(4),
email VARCHAR2(25)
)
ORGANIZATION EXTERNAL
(
TYPE ORACLE_LOADER
DEFAULT DIRECTORY ODCH_DATA_DIR
ACCESS PARAMETERS
(
records delimited by newline
preprocessor HDFS_BIN_PATH:hdfs_stream --使用hadoop预处理器
badfile ODCH_LOG_DIR:'empxt%a_%p.bad'
logfile ODCH_LOG_DIR:'empxt%a_%p.log'
fields terminated by ','
missing field values are null
( employee_id, first_name, last_name, job_id, manager_id,
hire_date char date_format date mask "dd-mon-yyyy",
salary, commission_pct, department_id, email
)
)
--在LOCATION对应的2个文件现在还不存在,我们会在下面的步骤中生成
--在LOCATION中使用多个文件,可以使Oracle可以多个程序并行访问HDFS
LOCATION ('empxt1.dat', 'empxt2.dat')
)
PARALLEL
REJECT LIMIT UNLIMITED;
SQL>
Table created
注:外部表使用实例
http://docs.oracle.com/cd/E11882_01/server.112/e25494/tables013.htm#ADMIN12896
8. 在Hadoop中放入示例文件
因ODCH要从Hadoop文件系统中读取数据. 所以我们先要在Hadoop中放入几个的数据文件。
--示例文件内容
[grid@gc ~]$ cat empxt1.dat
360,Jane,Janus,ST_CLERK,121,17-MAY-2001,3000,0,50,jjanus
361,Mark,Jasper,SA_REP,145,17-MAY-2001,8000,.1,80,mjasper
362,Brenda,Starr,AD_ASST,200,17-MAY-2001,5500,0,10,bstarr
363,Alex,Alda,AC_MGR,145,17-MAY-2001,9000,.15,80,aalda
[grid@gc ~]$ cat empxt2.dat
401,Jesse,Cromwell,HR_REP,203,17-MAY-2001,7000,0,40,jcromwel
402,Abby,Applegate,IT_PROG,103,17-MAY-2001,9000,.2,60,aapplega
403,Carol,Cousins,AD_VP,100,17-MAY-2001,27000,.3,90,ccousins
404,John,Richardson,AC_ACCOUNT,205,17-MAY-2001,5000,0,110,jrichard
--先在Hadoop中建立一个目录,然后把empxt*.dat放入该目录中
[grid@gc ~]$ hadoop dfs -mkdir odch_data
[grid@gc ~]$ wc -l empxt*
4 empxt1.dat
4 empxt2.dat
8 总计
[grid@gc ~]$ hadoop dfs -put empxt*.dat odch_data
[grid@gc ~]$ hadoop dfs -ls odch_data
Found 2 items
-rw-r--r-- 2 grid supergroup 228 2013-01-24 23:13 /user/grid/odch_data/empxt1.dat
-rw-r--r-- 2 grid supergroup 252 2013-01-24 23:13 /user/grid/odch_data/empxt2.dat
9. 生成“位置文件”
--执行下面的命令
export HADOOP_CLASSPATH="$OSCH_HOME/jlib/*"
hadoop jar \
${ORAHDFS_JAR} oracle.hadoop.hdfs.exttab.ExternalTable \
-D oracle.hadoop.hdfs.exttab.tableName=admin_ext_employees \
-D oracle.hadoop.hdfs.exttab.datasetPaths=odch_data \
-D oracle.hadoop.hdfs.exttab.datasetRegex=empxt.dat \
-D oracle.hadoop.hdfs.exttab.connection.url="jdbc:oracle:thin:@//192.168.2.100/gcdb" \
-D oracle.hadoop.hdfs.exttab.connection.user=SCOTT \
-publish
--输出如图
参数说明:
ExternalTable:使用hadoop ExternalTable命令工具
-D:指定相关参数
tableName:外部表名字
datasetPaths:源数据存放路径(HDFS)
datasetRegex:数据源格式
connection.url:oracle数据库连接串
connection.user:数据库用户名scott
命令执行后还要输入用户名密码:tiger
注意:
上面的输出结果中提示下面的内容,
DEPRECATED: The class oracle.hadoop.hdfs.exttab.ExternalTable is deprecated.
It is replaced by oracle.hadoop.exttab.ExternalTable.
意思类已经过时了,但是测试还是能用的,不用管它。被替换的新类使用方法可以见官方文档。
10. 查看“位置文件”内容
[grid@gc ~]$ cd /opt/odch/orahdfs-2.0.0/extdir/
[grid@gc extdir]$ cat osch-20130124045151-3641-1
参数说明:
CompressionCodec:HDFS 文件指向
默认值 hdfs://gc:9000/user/grid/odch_data/empxt1.dat
11. 检查结果
--查看外部表数据
--查看执行计划
--把外部表导入到数据库当中
SQL> create table ext1 as select * from admin_ext_employees;
表已创建。
到此完成了从HDFS数据到oracle数据库的导入!
三、 经验总结
1. 各种目录权限
如:odch和hadoop最好不要安装在/home/usr下,因linux系统的用户之间的/home/usr是不相互访问的,即使/home/usr设置了比较大的访问权限
2. 各种软件版本
可以参考:
http://www.dataguru.cn/forum.php?mod=viewthread&tid=39412
四、 参考资料
http://docs.oracle.com/cd/E37231_01/doc.20/e36961/sqlch.htm#CHDHHHAJ
http://docs.oracle.com/cd/E11882_01/server.112/e25494/tables013.htm#ADMIN12896
http://docs.oracle.com/cd/E37231_01/doc.20/e36961/odi.htm
http://www.dataguru.cn/forum.php?mod=viewthread&tid=39412
tigerfish ppt
五、 遇到的问题
1. 执行生成“位置文件”命令后,不显示结果跳出
现象:
当执行生成“位置文件”命令后,没有报错,没提示输入密码,也没有结果跳出执行。
[grid@gc logs]$ hadoop jar \
> ${ORAHDFS_JAR} oracle.hadoop.hdfs.exttab.ExternalTable \
> -D oracle.hadoop.hdfs.exttab.tableName=odch_ext_table \
> -D oracle.hadoop.hdfs.exttab.datasetPaths=odch_data \
> -D oracle.hadoop.hdfs.exttab.datasetRegex=odch1.dat \
> -D oracle.hadoop.hdfs.exttab.connection.url="jdbc:oracle:thin:@//192.168.2.100:1521/gcdb" \
> -D oracle.hadoop.hdfs.exttab.connection.user=SCOTT \
> -publish
DEPRECATED: The class oracle.hadoop.hdfs.exttab.ExternalTable is deprecated.
It is replaced by oracle.hadoop.exttab.ExternalTable.
Oracle SQL Connector for HDFS Release 2.0.0 - Production
Copyright (c) 2012, Oracle and/or its affiliates. All rights reserved.
解决:
检查执行命令正确无误后,感觉就是环境变量的问题
设置环境变量HADOOP_CLASSPATH后,再次运行命令正常。
export HADOOP_CLASSPATH="$OSCH_HOME/jlib/*"
2. IO 错误: The Network Adapter could not establish the connection
现象:
当使用新版本的ODCH的oracle数据库连接串时,出现下面的错误
[grid@gc ~]$ export HADOOP_CLASSPATH="$OSCH_HOME/jlib/*"
[grid@gc ~]$ hadoop jar \
> ${ORAHDFS_JAR} oracle.hadoop.exttab.ExternalTable \
> -D oracle.hadoop.exttab.tableName=odch_ext_table \
> -D oracle.hadoop.exttab.datasetPaths=odch_data \
> -D oracle.hadoop.exttab.datasetRegex=odch1.dat \
> -D oracle.hadoop.exttab.connection.url=jdbc:oracle:thin:@localhost:1521/gcdb \
> -D oracle.hadoop.exttab.connection.user=SCOTT \
> -publish
Oracle SQL Connector for HDFS Release 2.0.0 - Production
Copyright (c) 2012, Oracle and/or its affiliates. All rights reserved.
[Enter Database Password:]
java.sql.SQLRecoverableException: IO 错误: The Network Adapter could not establish the connection
oracle.net.ns.NetException: The Network Adapter could not establish the connection
To get a complete stacktrace, rerun the command with property: oracle.hadoop.exttab.printStackTrace=true
--使用printStackTrace参数出现帮助信息
[grid@gc ~]$ hadoop jar \
> ${ORAHDFS_JAR} oracle.hadoop.exttab.ExternalTable \
> -D oracle.hadoop.exttab.tableName=odch_ext_table \
> -D oracle.hadoop.exttab.datasetPaths=odch_data \
> -D oracle.hadoop.exttab.datasetRegex=odch1.dat \
> -D oracle.hadoop.exttab.connection.url=jdbc:oracle:thin:@localhost:1521/gcdb \
> -D oracle.hadoop.exttab.connection.user=SCOTT \
> -D oracle.hadoop.exttab.printStackTrace=true\
> -publish
Oracle SQL Connector for HDFS Release 2.0.0 - Production
Copyright (c) 2012, Oracle and/or its affiliates. All rights reserved.
Usage: bin/hadoop jar orahdfs.jar oracle.hadoop.exttab.ExternalTable [genericOptions] [-publish | -createTable [--noexecute]] | [-listLocations [--details]] | [-getDDL]
-createTable: creates an external table and also publishes HDFS data URIs to location files of the table
-publish: publishes the HDFS data URIs to the location files of an external table
-listLocations: prints the location files of an external table
-getDDL: prints the DDL of an external table
[genericOptions] : the generic command-line options supported by the Hadoop Tool interface
[options] : the following options can be used with thecommands
--noexecute : prints the execution plan of the -createTable or -publish command
--details : prints a detailed listing of the location files for the -listLocations command.
Example:
$HADOOP_HOME/bin/hadoop jar orahdfs.jar oracle.hadoop.exttab.ExternalTable -D oracle.hadoop.exttab.tableName=SALES_EXT_TAB -D oracle.hadoop.exttab.dataPaths=hdfs:/user/scott/data/ -D oracle.hadoop.connection.url=jdbc:oracle:thin:@myhost:1521/example.example.com -D oracle.hadoop.connection.user=scott -publish
[grid@gc ~]$
解决:
还是使用旧版本的ODCH连接oracle数据库的方法
hadoop jar \
/opt/odch/orahdfs-2.0.0/jlib/orahdfs.jar oracle.hadoop.hdfs.exttab.ExternalTable \
-D oracle.hadoop.hdfs.exttab.tableName=admin_ext_employees \
-D oracle.hadoop.hdfs.exttab.datasetPaths=odch_data \
-D oracle.hadoop.hdfs.exttab.datasetRegex=empxt.dat \
-D oracle.hadoop.hdfs.exttab.connection.url="jdbc:oracle:thin:@//192.168.2.100:1521/gcdb" \
-D oracle.hadoop.hdfs.exttab.connection.user=SCOTT \
-publish
3. OSCH_HOME is not set
现象:
在执行查询外部时,数据加载错误,在bad文件中出现以下内容
[grid@gc ~]$ cd /opt/odch/orahdfs-2.0.0/logs
[grid@gc logs]$ cat odch_ext_table000_14910.bad
OSCH_HOME is not set
解决:
原因是没有在hdfs_stream文件中设置环境变量
--在hdfs_stream文件中添加如下内容,
export OSCH_HOME=/opt/odch/orahdfs-2.0.0
4. Error opening job jar
现象:
当刚配置完成后命令是正常可用,过了一段时间后,再次执行时,却突然出现下面的错误。
[grid@gc ~]$ hadoop jar \
> ${ORAHDFS_JAR} oracle.hadoop.hdfs.exttab.ExternalTable \
> -D oracle.hadoop.hdfs.exttab.tableName=admin_ext_employees \
> -D oracle.hadoop.hdfs.exttab.datasetPaths=odch_data \
> -D oracle.hadoop.hdfs.exttab.datasetRegex=empxt.dat \
> -D oracle.hadoop.hdfs.exttab.connection.url="jdbc:oracle:thin:@//192.168.2.100/gcdb" \
> -D oracle.hadoop.hdfs.exttab.connection.user=SCOTT \
> -publish
Exception in thread "main" java.io.IOException: Error opening job jar: /opt/odch/orahdfs-2.0.0/jlib/orahdfs.jar
at org.apache.hadoop.util.RunJar.main(RunJar.java:90)
Caused by: java.util.zip.ZipException: error in opening zip file
at java.util.zip.ZipFile.open(Native Method)
at java.util.zip.ZipFile.<init>(ZipFile.java:114)
at java.util.jar.JarFile.<init>(JarFile.java:133)
at java.util.jar.JarFile.<init>(JarFile.java:70)
at org.apache.hadoop.util.RunJar.main(RunJar.java:88)
解决:
把orahdfs.jar包从安装文件里再重新copy一份,把原来的jar包覆盖掉,重新运行就好了。
[root@gc jlib]# mv orahdfs.jar /opt/odch/orahdfs-2.0.0/jlib/
mv:是否覆盖“/opt/odch/orahdfs-2.0.0/jlib/orahdfs.jar”? y
[root@gc jlib]# chown grid:hadoop /opt/odch/orahdfs-2.0.0/jlib/orahdfs.jar
[root@gc jlib]# chmod 777 /opt/odch/orahdfs-2.0.0/jlib/orahdfs.jar
[root@gc jlib]# ll /opt/odch/orahdfs-2.0.0/jlib/orahdfs.jar
-rwxrwxrwx 1 grid hadoop 117405 11-29 09:46 /opt/odch/orahdfs-2.0.0/jlib/orahdfs.jar