[hadoop3.x]HDFS存储策略和冷热温三阶段数据存储(六)概述

@[toc]

前言

目前博客Hadoop文章大都停留在Hadoop2.x阶段，本系列将依据黑马程序员大数据Hadoop3.x全套教程，对2.x没有的新特性进行补充更新，一键三连加关注，下次不迷路！

历史文章

[hadoop3.x系列]HDFS REST HTTP API的使用(一)WebHDFS

[hadoop3.x系列]HDFS REST HTTP API的使用(二)HttpFS

[hadoop3.x系列]Hadoop常用文件存储格式及BigData File Viewer工具的使用(三)

✨[hadoop3.x]新一代的存储格式Apache Arrow(四)

[hadoop3.x]HDFS存储类型和存储策略(五)概述

[hadoop3.x]HDFS存储策略和冷热温三阶段数据存储(六)概述

[hadoop3.x]HDFS中的内存存储支持(七)概述

🍑 1.1 💃存储策略命令💃

1 💃列出存储策略💃

列出所有存储策略。

命令：

[root@node1 Examples]# 💃hdfs storagepolicies -listPolicies💃
Block Storage Policies:    BlockStoragePolicy{PROVIDED:1, storageTypes=[PROVIDED, DISK], creationFallbacks=[PROVIDED, DISK], replicationFallbacks=[PROVIDED, DISK]}    BlockStoragePolicy{COLD:2, storageTypes=[ARCHIVE], creationFallbacks=[], replicationFallbacks=[]}    BlockStoragePolicy{WARM:5, storageTypes=[DISK, ARCHIVE], creationFallbacks=[DISK, ARCHIVE], replicationFallbacks=[DISK, ARCHIVE]}    BlockStoragePolicy{HOT:7, storageTypes=[DISK], creationFallbacks=[], replicationFallbacks=[ARCHIVE]}    BlockStoragePolicy{ONE_SSD:10, storageTypes=[SSD, DISK], creationFallbacks=[SSD, DISK], replicationFallbacks=[SSD, DISK]}    BlockStoragePolicy{ALL_SSD:12, storageTypes=[SSD], creationFallbacks=[DISK], replicationFallbacks=[DISK]}    BlockStoragePolicy{LAZY_PERSIST:15, storageTypes=[RAM_DISK, DISK], creationFallbacks=[DISK], replicationFallbacks=[DISK]}

hdfs storagepolicies -listPolicies

2 💃设置存储策略💃

给一个文件或目录设置存储策略

hdfs storagepolicies -setStoragePolicy -path <path> -policy <policy>

参数：

3 💃取消存储策略💃

取消文件或目录的存储策略。在执行unset命令之后，将应用当前目录最近的祖先存储策略，如果没有任何祖先的策略，则将应用默认的存储策略。

hdfs storagepolicies -unsetStoragePolicy -path <path>

参数：

4 💃获取存储策略💃

获取文件或目录的存储策略。

hdfs storagepolicies -getStoragePolicy -path <path>

🍑2.1 💃冷热温三阶段数据存储💃

为了更加充分的利用存储资源，我们可以将数据分为冷、热、温三个阶段来存储。

1 💃配置DataNode存储目录💃

为了能够支撑不同类型的数据，我们需要在hdfs-site.xml中配置不同存储类型数据的位置。

进入到Hadoop配置目录，编辑hdfs-site.xml

cd /export/server/hadoop-3.1.4/etc/hadoop
vim hdfs-site.xml
<property>
  <name>dfs.datanode.data.dir</name>
  <value>[DISK]file:///export/server/hadoop-3.1.4/data/datanode,[ARCHIVE]file:///export/server/hadoop-3.1.4/data/archive</value>
  <description>DataNode存储名称空间和事务日志的本地文件系统上的路径</description>
</property>

分发到不同另外两个节点中

scp hdfs-site.xml node2.itcast.cn:$PWD

scp hdfs-site.xml node3.itcast.cn:$PWD

重启HDFS集群

stop-dfs.sh
start-dfs.sh

配置好后，我们在WebUI的Datanodes页面中点击任意一个DataNode节点

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-1HGUDOAb-1633607380830)(https://gitee.com/the_efforts_paid_offf/picture-blog/raw/master/img/20211006210405.jpg)]

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-VHsmIDyH-1633607380831)(https://gitee.com/the_efforts_paid_offf/picture-blog/raw/master/img/20211006210409.jpg)]

可以看到，现在配置的是两个目录，一个StorageType为ARCHIVE、一个Storage为DISK。

2 💃配置策略💃

创建测试目录结构

hdfs dfs -mkdir -p /data/hdfs-test/data_phase/hot
hdfs dfs -mkdir -p /data/hdfs-test/data_phase/warm
hdfs dfs -mkdir -p /data/hdfs-test/data_phase/cold

2. 查看当前HDFS支持的存储策略

[root@node1 Examples]# hdfs storagepolicies -listPolicies
Block Storage Policies:
BlockStoragePolicy{PROVIDED:1, storageTypes=[PROVIDED, DISK], creationFallbacks=[PROVIDED, DISK], replicationFallbacks=[PROVIDED, DISK]}
BlockStoragePolicy{COLD:2, storageTypes=[ARCHIVE], creationFallbacks=[], replicationFallbacks=[]}
BlockStoragePolicy{WARM:5, storageTypes=[DISK, ARCHIVE], creationFallbacks=[DISK, ARCHIVE], replicationFallbacks=[DISK, ARCHIVE]}
BlockStoragePolicy{HOT:7, storageTypes=[DISK], creationFallbacks=[], replicationFallbacks=[ARCHIVE]}
BlockStoragePolicy{ONE_SSD:10, storageTypes=[SSD, DISK], creationFallbacks=[SSD, DISK], replicationFallbacks=[SSD, DISK]}
BlockStoragePolicy{ALL_SSD:12, storageTypes=[SSD], creationFallbacks=[DISK], replicationFallbacks=[DISK]}
BlockStoragePolicy{LAZY_PERSIST:15, storageTypes=[RAM_DISK, DISK], creationFallbacks=[DISK], replicationFallbacks=[DISK]}

3. 分别设置三个目录的存储策略

hdfs storagepolicies -setStoragePolicy -path /data/hdfs-test/data_phase/hot -policy HOT
hdfs storagepolicies -setStoragePolicy -path /data/hdfs-test/data_phase/warm -policy WARM
hdfs storagepolicies -setStoragePolicy -path /data/hdfs-test/data_phase/cold -policy COLD

4. 查看三个目录的存储策略

hdfs storagepolicies -getStoragePolicy -path /data/hdfs-test/data_phase/hot
hdfs storagepolicies -getStoragePolicy -path /data/hdfs-test/data_phase/warm 
hdfs storagepolicies -getStoragePolicy -path /data/hdfs-test/data_phase/cold 

[root@node1 Examples]# hdfs storagepolicies -getStoragePolicy -path /data/hdfs-test/data_phase/hot
The storage policy of /data/hdfs-test/data_phase/hot:
BlockStoragePolicy{HOT:7, storageTypes=[DISK], creationFallbacks=[], replicationFallbacks=[ARCHIVE]}

[root@node1 Examples]# hdfs storagepolicies -getStoragePolicy -path /data/hdfs-test/data_phase/warm 
The storage policy of /data/hdfs-test/data_phase/warm:
BlockStoragePolicy{WARM:5, storageTypes=[DISK, ARCHIVE], creationFallbacks=[DISK, ARCHIVE], replicationFallbacks=[DISK, ARCHIVE]}

[root@node1 Examples]# hdfs storagepolicies -getStoragePolicy -path /data/hdfs-test/data_phase/cold 
The storage policy of /data/hdfs-test/data_phase/cold:
BlockStoragePolicy{COLD:2, storageTypes=[ARCHIVE], creationFallbacks=[], replicationFallbacks=[]}

3 💃上传测试💃

分别上传文件到三个目录中测试

hdfs dfs -put /etc/profile /data/hdfs-test/data_phase/hot
hdfs dfs -put /etc/profile /data/hdfs-test/data_phase/warm
hdfs dfs -put /etc/profile /data/hdfs-test/data_phase/cold

查看不同存储策略文件的block位置

[root@node1 hadoop]# hdfs fsck /data/hdfs-test/data_phase/hot/profile -files -blocks -locations
Connecting to namenode via http://node1.itcast.cn:9870/fsck?ugi=root&files=1&blocks=1&locations=1&path=%2Fdata%2Fhdfs-test%2Fdata_phase%2Fhot%2Fprofile
FSCK started by root (auth:SIMPLE) from /192.168.88.100 for path /data/hdfs-test/data_phase/hot/profile at Sun Oct 11 22:03:05 CST 2020

/data/hdfs-test/data_phase/hot/profile 3158 bytes, replicated: replication=3, 1 block(s):  OK
0. BP-538037512-192.168.88.100-1600884040401:blk_1073742535_1750 len=3158 Live_repl=3  [DatanodeInfoWithStorage[192.168.88.101:9866,DS-96feb29a-5dfd-4692-81ea-9e7f100166fe,DISK], DatanodeInfoWithStorage[192.168.88.100:9866,DS-79739be9-5f9b-4f96-a005-aa5b507899f5,DISK], DatanodeInfoWithStorage[192.168.88.102:9866,DS-e28af2f2-21ae-4aa6-932e-e376dd04ddde,DISK]]

hdfs fsck /data/hdfs-test/data_phase/warm/profile -files -blocks -locations

/data/hdfs-test/data_phase/warm/profile 3158 bytes, replicated: replication=3, 1 block(s):  OK
0. BP-538037512-192.168.88.100-1600884040401:blk_1073742536_1751 len=3158 Live_repl=3  [DatanodeInfoWithStorage[192.168.88.102:9866,DS-636f34a0-682c-4d1b-b4ee-b4c34e857957,ARCHIVE], DatanodeInfoWithStorage[192.168.88.101:9866,DS-ff6970f8-43e0-431f-9041-fc440a44fdb0,ARCHIVE], DatanodeInfoWithStorage[192.168.88.100:9866,DS-79739be9-5f9b-4f96-a005-aa5b507899f5,DISK]]


hdfs fsck /data/hdfs-test/data_phase/cold/profile -files -blocks -locations
/data/hdfs-test/data_phase/cold/profile 3158 bytes, replicated: replication=3, 1 block(s):  OK
0. BP-538037512-192.168.88.100-1600884040401:blk_1073742537_1752 len=3158 Live_repl=3  [DatanodeInfoWithStorage[192.168.88.102:9866,DS-636f34a0-682c-4d1b-b4ee-b4c34e857957,ARCHIVE], DatanodeInfoWithStorage[192.168.88.101:9866,DS-ff6970f8-43e0-431f-9041-fc440a44fdb0,ARCHIVE], DatanodeInfoWithStorage[192.168.88.100:9866,DS-ca9759a0-f6f0-4b8b-af38-d96f603bca93,ARCHIVE]]

我们可以看到：

hot目录中的block，3个block都在DISK磁盘

warm目录中的block，1个block在DISK磁盘，另外两个在archive磁盘

cold目录中的block，3个block都在archive磁盘

后记

📢博客主页：https://manor.blog.csdn.net
📢欢迎点赞 👍 收藏 ⭐留言 📝 如有错误敬请指正！
📢本文由 manor 原创，首发于 CSDN博客🙉
📢Hadoop系列文章会每天更新！✨