HDFS中的fsck命令(检查数据块是否健康)

在HDFS中，提供了fsck命令，用于检查HDFS上文件和目录的健康状态、获取文件的block信息和位置信息等。

我们在master机器上执行hdfs fsck就可以看到这个命令的用法。

[hadoop-twq@master ~]$ hdfs fsck
Usage: hdfs fsck <path> [-list-corruptfileblocks | [-move | -delete | -openforwrite] [-files [-blocks [-locations | -racks]]]] [-includeSnapshots] [-storagepolicies] [-blockId <blk_Id>]
	<path>	start checking from this path
	-move	move corrupted files to /lost+found
	-delete	delete corrupted files
	-files	print out files being checked
	-openforwrite	print out files opened for write
	-includeSnapshots	include snapshot data if the given path indicates a snapshottable directory or there are snapshottable directories under it
	-list-corruptfileblocks	print out list of missing blocks and files they belong to
	-blocks	print out block report
	-locations	print out locations for every block
	-racks	print out network topology for data-node locations
	-storagepolicies	print out storage policy summary for the blocks
	-blockId	print out which file this blockId belongs to, locations (nodes, racks) of this block, and other diagnostics info (under replicated, corrupted or not, etc)

查看文件目录的健康信息

执行如下的命令：

hdfs fsck /user/hadoop-twq/cmd

可以查看/user/hadoop-twq/cmd目录的健康信息：

其中有一个比较重要的信息，就是Corrupt blocks，表示损坏的数据块的数量

查看文件中损坏的块 (-list-corruptfileblocks)

[hadoop-twq@master ~]$ hdfs fsck /user/hadoop-twq/cmd -list-corruptfileblocks
Connecting to namenode via http://master:50070/fsck?ugi=hadoop-twq&listcorruptfileblocks=1&path=%2Fuser%2Fhadoop-twq%2Fcmd
The filesystem under path '/user/hadoop-twq/cmd' has 0 CORRUPT files

上面的命令可以找到某个目录下面的损坏的数据块，但是上面表示没有看到坏的数据块

损坏文件的处理

将损坏的文件移动至/lost+found目录 (-move)

hdfs fsck /user/hadoop-twq/cmd -move

　删除有损坏数据块的文件 (-delete)

hdfs fsck /user/hadoop-twq/cmd -delete

检查并列出所有文件状态(-files)

执行如下的命令：

hdfs fsck /user/hadoop-twq/cmd -files

　显示结果如下：

　

上面的命令可以检查指定路径下的所有文件的信息，包括：数据块的数量以及数据块的备份情况

检查并打印正在被打开执行写操作的文件(-openforwrite)

执行下面的命令可以检查指定路径下面的哪些文件正在执行写操作：

hdfs fsck /user/hadoop-twq/cmd -openforwrite

打印文件的Block报告(-blocks)

　执行下面的命令，可以查看一个指定文件的所有的Block详细信息，需要和-files一起使用：　

hdfs fsck /user/hadoop-twq/cmd/big_file.txt -files -blocks

　　结果如下：

如果，我们在上面的命令再加上-locations的话，就是表示还需要打印每一个数据块的位置信息，如下命令：

hdfs fsck /user/hadoop-twq/cmd/big_file.txt -files -blocks -locations

　　结果如下：

如果，我们在上面的命令再加上-racks的话，就是表示还需要打印每一个数据块的位置所在的机架信息，如下命令：

hdfs fsck /user/hadoop-twq/cmd/big_file.txt -files -blocks -locations -racks

　　结果如下：

hdfs fsck的使用场景

场景一

当我们执行如下的命令：

hdfs fsck /user/hadoop-twq/cmd

可以查看/user/hadoop-twq/cmd目录的健康信息： 　　

我们可以看出，有两个文件的数据块的备份数量不足，这个我们可以通过如下的命令，重新设置两个文件数据块的备份数：

## 将文件big_file.txt对应的数据块备份数设置为1
hadoop fs -setrep -w 1 /user/hadoop-twq/cmd/big_file.txt
## 将文件parameter_test.txt对应的数据块备份数设置为1
hadoop fs -setrep -w 1 /user/hadoop-twq/cmd/parameter_test.txt

上面命令中 -w 参数表示等待备份数到达指定的备份数，加上这个参数后再执行的话，则需要比较长的时间　　

执行完上面的命令后，我们再来执行下面的命令：

hdfs fsck /user/hadoop-twq/cmd

结果如下：

　　

场景二

当我们访问HDFS的WEB UI的时候，出现了如下的警告信息：

表明有一个数据块丢失了，这个时候我们执行下面的命令来确定是哪一个文件的数据块丢失了：

[hadoop-twq@master ~]$ hdfs fsck / -list-corruptfileblocks
Connecting to namenode via http://master:50070/fsck?ugi=hadoop-twq&listcorruptfileblocks=1&path=%2F
The list of corrupt files under path '/' are:
blk_1073744153	/tmp/hadoop-yarn/staging/history/done_intermediate/hadoop-twq/job_1528682852398_0015_conf.xml
The filesystem under path '/' has 1 CORRUPT files

　　

发现是数据块blk_1073744153丢失了，这个数据块是淑文文件/tmp/hadoop-yarn/staging/history/done_intermediate/hadoop-twq/job_1528682852398_0015_conf.xml的。

如果出现这种场景是因为在DataNode中没有这个数据块，但是在NameNode的元数据中有这个数据块的信息，我们可以执行下面的命令，把这些没用的数据块信息删除掉，如下：

[hadoop-twq@master ~]$ hdfs fsck /tmp/hadoop-yarn/staging/history/done_intermediate/hadoop-twq/ -delete
Connecting to namenode via http://master:50070/fsck?ugi=hadoop-twq&delete=1&path=%2Ftmp%2Fhadoop-yarn%2Fstaging%2Fhistory%2Fdone_intermediate%2Fhadoop-twq
FSCK started by hadoop-twq (auth:SIMPLE) from /192.168.126.130 for path /tmp/hadoop-yarn/staging/history/done_intermediate/hadoop-twq at Tue Mar 05 19:18:00 EST 2019
....................................................................................................
..
/tmp/hadoop-yarn/staging/history/done_intermediate/hadoop-twq/job_1528682852398_0015_conf.xml: CORRUPT blockpool BP-1639452328-192.168.126.130-1525478508894 block blk_1073744153

/tmp/hadoop-yarn/staging/history/done_intermediate/hadoop-twq/job_1528682852398_0015_conf.xml: MISSING 1 blocks of total size 220262 B...................................................................................................
....................................................................................................
........................Status: CORRUPT
 Total size:	28418833 B
 Total dirs:	1
 Total files:	324
 Total symlinks:		0
 Total blocks (validated):	324 (avg. block size 87712 B)
  ********************************
  UNDER MIN REPL'D BLOCKS:	1 (0.30864197 %)
  dfs.namenode.replication.min:	1
  CORRUPT FILES:	1
  MISSING BLOCKS:	1
  MISSING SIZE:		220262 B
  CORRUPT BLOCKS: 	1
  ********************************
 Minimally replicated blocks:	323 (99.69136 %)
 Over-replicated blocks:	0 (0.0 %)
 Under-replicated blocks:	0 (0.0 %)
 Mis-replicated blocks:		0 (0.0 %)
 Default replication factor:	1
 Average block replication:	0.99691355
 Corrupt blocks:		1
 Missing replicas:		0 (0.0 %)
 Number of data-nodes:		2
 Number of racks:		1
FSCK ended at Tue Mar 05 19:18:01 EST 2019 in 215 milliseconds

　　然后执行：

[hadoop-twq@master ~]$ hdfs fsck / -list-corruptfileblocks
Connecting to namenode via http://master:50070/fsck?ugi=hadoop-twq&listcorruptfileblocks=1&path=%2F
The filesystem under path '/' has 0 CORRUPT files

　　

丢失的数据块没有的，被删除了。我们也可以刷新WEB UI，也没有了警告信息：