## 1.2 使用快照备份
hdfs dfsadmin -allowSnapshot /user/fgedu/data
hdfs dfs -createSnapshot /user/fgedu/data snapshot1
hdfs dfs -ls /user/fgedu/data/.snapshot/

## 1.3 使用外部存储备份
hadoop distcp hdfs://source-cluster:8020/user/fgedu/data file:///backup/hdfs/

# 2. HBase数据备份
## 2.1 使用Export备份
hbase org.apache.hadoop.hbase.mapreduce.Export fgedu_table /user/fgedu/backup/hbase/fgedu_table

## 2.2 使用Snapshot备份
hbase shell
> snapshot ‘fgedu_table’, ‘snapshot1’
> list_snapshots
> clone_snapshot ‘snapshot1’, ‘fgedu_table_clone’

## 2.3 使用复制表备份
hbase shell
> create ‘fgedu_table_backup’, ‘cf1’
> disable ‘fgedu_table_backup’
> alter ‘fgedu_table_backup’, {NAME => ‘cf1’, VERSIONS => 1}
> enable ‘fgedu_table_backup’
> put ‘fgedu_table_backup’, ‘row1’, ‘cf1:col1’, ‘value1’

# 3. Hive数据备份
## 3.1 导出表结构和数据
hive -e “EXPORT TABLE fgedu_db.fgedu_table TO ‘/user/fgedu/backup/hive/fgedu_table’;”

## 3.2 备份元数据
mysqldump -u root -p hive > hive_metadata_backup.sql

# 4. 定期备份脚本
## 4.1 编写备份脚本
vi backup.sh
#!/bin/bash
# backup.sh
# from:www.itpux.com.qq113257174.wx:itpux-com
# web: `http://www.fgedu.net.cn`

echo “开始备份HDFS数据…”
hdfs dfs -cp /user/fgedu/data hdfs://backup-cluster/user/fgedu/backup/$(date +%Y%m%d)

echo “开始备份HBase数据…”
hbase org.apache.hadoop.hbase.mapreduce.Export fgedu_table /user/fgedu/backup/hbase/fgedu_table_$(date +%Y%m%d)

echo “开始备份Hive数据…”
hive -e “EXPORT TABLE fgedu_db.fgedu_table TO ‘/user/fgedu/backup/hive/fgedu_table_$(date +%Y%m%d)’;”

echo “备份完成”

## 4.2 执行备份脚本
chmod +x backup.sh
./backup.sh

## 4.3 设置定时任务
crontab -e
0 0 * * * /bigdata/app/scripts/backup.sh

## 1.2 从快照恢复
hdfs dfs -cp /user/fgedu/data/.snapshot/snapshot1/* /user/fgedu/data/

## 1.3 从外部存储恢复
hadoop distcp file:///backup/hdfs/data hdfs://target-cluster:8020/user/fgedu/data

# 2. HBase灾难恢复
## 2.1 从Export备份恢复
hbase org.apache.hadoop.hbase.mapreduce.Import fgedu_table /user/fgedu/backup/hbase/fgedu_table

## 2.2 从Snapshot恢复
hbase shell
> restore_snapshot ‘snapshot1’

## 2.3 从复制表恢复
hbase shell
> scan ‘fgedu_table_backup’

# 3. Hive灾难恢复
## 3.1 导入表结构和数据
hive -e “IMPORT TABLE fgedu_db.fgedu_table FROM ‘/user/fgedu/backup/hive/fgedu_table’;”

## 3.2 恢复元数据
mysql -u root -p hive < hive_metadata_backup.sql # 4. 灾难恢复演练 ## 4.1 编写恢复脚本 vi recovery.sh #!/bin/bash # recovery.sh # from:www.itpux.com.qq113257174.wx:itpux-com # web: `http://www.fgedu.net.cn` echo "开始恢复HDFS数据..." hdfs dfs -cp hdfs://backup-cluster/user/fgedu/backup/$(date +%Y%m%d) /user/fgedu/data echo "开始恢复HBase数据..." hbase org.apache.hadoop.hbase.mapreduce.Import fgedu_table /user/fgedu/backup/hbase/fgedu_table_$(date +%Y%m%d) echo "开始恢复Hive数据..." hive -e "IMPORT TABLE fgedu_db.fgedu_table FROM '/user/fgedu/backup/hive/fgedu_table_$(date +%Y%m%d)';" echo "恢复完成" ## 4.2 执行恢复脚本 chmod +x recovery.sh ./recovery.sh

## 1.2 配置JournalNode
vi /bigdata/app/hadoop/etc/hadoop/hdfs-site.xml dfs.journalnode.edits.dir
/bigdata/fgdata/hdfs/journalnode

# 2. YARN高可用配置
## 2.1 配置ResourceManager高可用
vi /bigdata/app/hadoop/etc/hadoop/yarn-site.xml yarn.resourcemanager.ha.enabled
true yarn.resourcemanager.cluster-id
cluster1 yarn.resourcemanager.ha.rm-ids
rm1,rm2 yarn.resourcemanager.hostname.rm1
fgedu01 yarn.resourcemanager.hostname.rm2
fgedu02 yarn.resourcemanager.zk-address
fgedu01:2181,fgedu02:2181,fgedu03:2181 yarn.resourcemanager.recovery.enabled
true yarn.resourcemanager.store.class
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore

# 3. ZooKeeper高可用配置
## 3.1 配置ZooKeeper集群
vi /bigdata/app/zookeeper/conf/zoo.cfg
tickTime=2000
initLimit=10
syncLimit=5
dataDir=/bigdata/fgdata/zookeeper
clientPort=2181
server.1=fgedu01:2888:3888
server.2=fgedu02:2888:3888
server.3=fgedu03:2888:3888

## 3.2 创建myid文件
echo “1” > /bigdata/fgdata/zookeeper/myid
echo “2” > /bigdata/fgdata/zookeeper/myid
echo “3” > /bigdata/fgdata/zookeeper/myid