cdh集羣hive升級，數據不丟失

1.下載hive-1.2.1安裝包 http://archive.apache.org/dist/hive/hive-1.2.1/apache-hive-1.2.1-bin.tar.gz

2.將安裝包傳到集羣全部節點上

3.全部節點root用戶下

cd /opt/cloudera/parcels/CDH/lib/hive
mkdir lib121

4.全部節點解壓 apache-hive-1.2.1-bin.tar.gz

5.全部節點將解壓出來hive/lib下全部文件拷貝到lib121

6.Coudera 界面中止Hive服務，若是安裝了Hue，Ozzie，Impala先停掉再停掉Hive

6.全部節點修改HIVE_LIB變量 HIVE_LIB=${HIVE_HOME}/lib ==> HIVE_LIB=${HIVE_HOME}/lib121

vim /opt/cloudera/parcels/CDH/lib/hive/bin/hive

7.全部節點更新hadoop上jline jar包，並刪除老的jlien jar包

ln -s /opt/cloudera/parcels/CDH/lib/hive/lib121/jline-2.12.jar /opt/cloudera/parcels/CDH/lib/hadoop-yarn/lib/jline-2.12.jar

rm -rf /opt/cloudera/parcels/CDH/lib/hadoop-yarn/lib/jline-0.94.jar

8.元數據節點備份MySQL中hive元數據庫

mysqldump -uroot -hhserver -padmin123 hive > hive013.sql

9.元數據節點進入mysql

create database hive013 DEFAULT CHARSET utf8 COLLATE utf8_general_ci;
grant all privileges on *.* to 'root'@'hserver' identified by 'admin123' with grant option;
flush privileges;
use hive013;
source hive013.sql;

10.元數據節點升級Hive元數據庫

升級前查看升級腳本，評估影響，特別是生產環境，根據業務需求更改過元數據的，若是是接手別人的，那就比較苦逼了，每一個元數據庫表結構都要看一遍。

若是沒有動過就直接升級就好

cd /home/dp/apache-hive-1.2.1-bin/scripts/metastore/upgrade/mysql
mysql -uroot -hhserver -padmin123
use hive;
source upgrade-0.13.0-to-0.14.0.mysql.sql;
source upgrade-0.14.0-to-1.1.0.mysql.sql;
source upgrade-1.1.0-to-1.2.0.mysql.sql;
exit

11.cloudera 管理界面啓動Hive服務

12.任意節點

hive #進入前hive shell界面
show databases;  
use testupdate;
show tables;
select * from test1;  #查看是否原來的數據還在
insert into test1 values("AAAA","BBBB","CCCC");
select * from test1;

#hive支持行級操做了！

參考了https://blog.csdn.net/xjping0794/article/details/77542096