Linux系统dd命令详解

Linux系统中dd命令可以用指定大小的块拷贝一个文件，并在拷贝的同时进行指定的转换。下面将通过几个案例为大家详细讲解一下dd命令的具体使用方法。

参数注释：

  bs=BYTES  read and write BYTES bytes at a time (also see ibs=,obs=)
  cbs=BYTES  convert BYTES bytes at a time
  conv=CONVS  convert the file as per the comma separated symbol list
  count=N   copy only N input blocks
  ibs=BYTES  read BYTES bytes at a time (default: 512)
  if=FILE   read from FILE instead of stdin(默认为标准输入)
  iflag=FLAGS  read as per the comma separated symbol list
  obs=BYTES  write BYTES bytes at a time (default: 512)
  of=FILE   write to FILE instead of stdout(默认为标准输出)
  oflag=FLAGS  write as per the comma separated symbol list
  seek=BLOCKS  skip BLOCKS obs-sized blocks at start of output
  skip=BLOCKS  skip BLOCKS ibs-sized blocks at start of input
  status=WHICH WHICH info to suppress outputting to stderr;
      'noxfer' suppresses transfer stats, 'none' suppresses all

CONVS的可选参数

  ascii  from EBCDIC to ASCII
  ebcdic from ASCII to EBCDIC
  ibm  from ASCII to alternate EBCDIC
  block  pad newline-terminated records with spaces to cbs-size
  unblock replace trailing spaces in cbs-size records with newline
  lcase  change upper case to lower case
  nocreat do not create the output file
  excl  fail if the output file already exists
  notrunc do not truncate the output file
  ucase  change lower case to upper case
  sparse try to seek rather than write the output for NUL input blocks
  swab  swap every pair of input bytes
  noerror continue after read errors
  sync  pad every input block with NULs to ibs-size; when used
    with block or unblock, pad with spaces rather than NULs
  fdatasync physically write output file data before finishing
  fsync  likewise, but also write metadata

FLAGS的可选参数

  append append mode (makes sense only for output; conv=notrunc suggested)
  direct use direct I/O for data
  directory fail unless a directory
  dsync  use synchronized I/O for data
  sync  likewise, but also for metadata
  fullblock accumulate full blocks of input (iflag only)
  nonblock use non-blocking I/O
  noatime do not update access time
  noctty do not assign controlling terminal from file
  nofollow do not follow symlinks
  count_bytes treat 'count=N' as a byte count (iflag only)

注意：指定数字的地方若以下列字符结尾，则乘以相应的数字：

 c =1, w =2, b =512, kB =1000, K =1024, MB =1000*1000, M =1024*1024, xM =M
 
 GB =1000*1000*1000, G =1024*1024*1024, and so on for T, P, E, Z, Y

二、使用实例

1、将本地的/dev/hdb整盘备份到/dev/hdd

 dd` `if``=``/dev/hdb` `of=``/dev/hdd

2、将/dev/hdb全盘数据备份到指定路径的image文件

 dd` `if``=``/dev/hdb` `of=``/root/image

3、备份/dev/hdb全盘数据，并利用gzip工具进行压缩，保存到指定路径

 dd` `if``=``/dev/hdb` `| ``gzip` `> ``/root/image``.gz

4、把一个文件拆分为3个文件

 #文件大小为2.3k
 [Oracle@rhel6 ~]$ ll db1_db_links.sql
 -rw-r--r-- 1 oracle oinstall 2344 Nov 21 10:39 db1_db_links.sql
 #把这个文件拆成每个文件1k，bs=1k，count=1，使用skip参数指定在输入文件中跳过多少个bs支读取
 [oracle@rhel6 ~]$ dd if=db1_db_links.sql of=dd01.sql bs=1k count=1
 1+0 records in
 1+0 records out
 1024 bytes (1.0 kB) copied, 4.5536e-05 s, 22.5 MB/s
 [oracle@rhel6 ~]$ dd if=db1_db_links.sql of=dd02.sql bs=1k count=1 skip=1
 1+0 records in
 1+0 records out
 1024 bytes (1.0 kB) copied, 0.000146387 s, 7.0 MB/s
 [oracle@rhel6 ~]$ dd if=db1_db_links.sql of=dd03.sql bs=1k count=1 skip=2
 0+1 records in
 0+1 records out
 296 bytes (296 B) copied, 0.000204216 s, 1.4 MB/s
 #拆分出的文件
 [oracle@rhel6 ~]$ ll dd*sql
 -rw-r--r-- 1 oracle oinstall 1024 May 20 14:58 dd01.sql
 -rw-r--r-- 1 oracle oinstall 1024 May 20 14:58 dd02.sql
 -rw-r--r-- 1 oracle oinstall 296 May 20 14:58 dd03.sql

5、把拆分出的文件合并为1个

 #合并操作，此时用到seek参数，用于指定在输入文件中跳过的bs数
 [oracle@rhel6 ~]$ dd of=1.sql if=dd01.sql
 2+0 records in
 2+0 records out
 1024 bytes (1.0 kB) copied, 0.000176 s, 5.8 MB/s
 [oracle@rhel6 ~]$ dd of=1.sql if=dd02.sql bs=1k seek=1
 1+0 records in
 1+0 records out
 1024 bytes (1.0 kB) copied, 0.000124038 s, 8.3 MB/s
 [oracle@rhel6 ~]$ dd of=1.sql if=dd03.sql bs=1k seek=2
 0+1 records in
 0+1 records out
 296 bytes (296 B) copied, 0.00203881 s, 145 kB/s
 #与拆分前的文件进行校验
 [oracle@rhel6 ~]$ diff 1.sql db1_db_links.sql
 [oracle@rhel6 ~]$

6、在输出文件中指定的位置插入数据，而不截断输出文件

需要使用conv=notrunc参数

 [oracle@rhel6 ~]$ ``dd` `if``=2.sql of=1.sql bs=1k seek=1 count=2 conv=notrunc

以上就是良许教程网为各位朋友分享的Linux系统相关内容。想要了解更多Linux相关知识记得关注公众号“良许Linux”，或扫描下方二维码进行关注，更多干货等着你！