博客
关于我
强烈建议你试试无所不能的chatGPT,快点击我
xfsdump命令使用
阅读量:5782 次
发布时间:2019-06-18

本文共 6162 字,大约阅读时间需要 20 分钟。

一:使用xfsdump备份和恢复xfs文件系统

首先了解一下xfsdump的备份级别有以下两种,默认为0(即完全备份)

0                                    完全备份

1    <=   level <= 9      增量备份

1.创建xfs文件系统并挂载到/data目录下

-x4xj:~ # mkfs.xfs /dev/sdb

linux-x4xj:~ # mkdir /data

linux-x4xj:~ # mount /dev/sdb /data

linux-x4xj:~ # df -h

Filesystem      Size Used Avail Use% Mounted on

/dev/sda2       4.0G 1.6G  2.2G  43% /

devtmpfs        465M 168K  465M   1% /dev

tmpfs           465M     0 465M   0% /dev/shm

/dev/sdb         20G  33M   20G   1% /data

linux-x4xj:~ #

2.备份

首先复制一些文件到/data目录下,以便测试,如下所示

linux-x4xj:~ # ls -l /data

total 1792

-rwxr--r-- 1 nobody nobody 1832960 Sep 1314:12 iozone3_429.tar

drwxr-xr-x 2 nobody nobody     102 Sep 16 20:46 lustre-client-el5.8-2.1.6

linux-x4xj:~ #

◆首先备份/data整个目录到/tmp目录下

测试过程中发现:如果命令这样写xfsdump -f /tmp/dump_data /data/,就会提示错误(xfsdump:ERROR: /data/ does not identify a file system)

 

linux-x4xj:~ # xfsdump -f /tmp/dump_data /data/

xfsdump: using file dump (drive_simple)strategy

xfsdump: version 3.0.5 (dump format 3.0) -Running single-threaded

 

 ============================= dump labeldialog ==============================

 

please enter label for this dump session(timeout in 300 sec)

 -> dump_data

session label entered:"dump_data"

 

 ---------------------------------end dialog ---------------------------------

 

xfsdump: ERROR:/data/ does not identify a file system

xfsdump: usage: xfsdump [ -a (dump DMFdualstate files as offline) ]

                           *

                           *

                           *

 

 

必须要这样写xfsdump -f /tmp/dump_data /data(即/data后面的/不能要)这样才能成功

 

linux-x4xj:~ # xfsdump -f /tmp/data_dump /data

xfsdump: using file dump (drive_simple)strategy

xfsdump: version 3.0.5 (dump format 3.0) -Running single-threaded

 

 ============================= dump labeldialog ==============================

 

please enter label for this dump session(timeout in 300 sec)

 -> data_dump

session label entered:"data_dump"

 

 --------------------------------- end dialog---------------------------------

 

xfsdump: level 0 dump of linux-x4xj:/data

xfsdump: dump date: Tue Sep 16 20:58:582014

xfsdump: session id:ca3390ef-1a8c-429b-8d87-4e60d22d3c54

xfsdump: session label: "data_dump"

xfsdump: ino map phase 1: constructinginitial dump list

xfsdump: ino map phase 2: skipping (nopruning necessary)

xfsdump: ino map phase 3: skipping (onlyone dump stream)

xfsdump: ino map construction complete

xfsdump: estimated dump size: 70670208bytes

 

 ============================= media labeldialog =============================

 

please enter label for media in drive 0(timeout in 300 sec)

 -> media0

media label entered: "media0"

 

 --------------------------------- end dialog---------------------------------

 

xfsdump: creating dump session media file 0(media 0, file 0)

xfsdump: dumping ino map

xfsdump: dumping directories

xfsdump: dumping non-directory files

xfsdump: ending media file

xfsdump: media file size 70680352 bytes

xfsdump: dump size (non-dir files) :70656816 bytes

xfsdump: dump complete: 9 seconds elapsed

xfsdump: Dump Status: SUCCESS

linux-x4xj:~ # ls -l /tmp/data_dump

-rw-r--r-- 1 root root 70680352 Sep 1620:59 /tmp/data_dump

linux-x4xj:~ #

以上为会话模式,即会提示让你输入label for session和label for dump,可以用下面的命令运行,就不会出现会话模式了

linux-x4xj:~ # xfsdump -f /tmp/data_dump /data -L data_dump -M data_dump

 

◆备份/data目录下的lustre-client-el5.8-2.1.6子目录到/tmp目录下,命令如下

linux-x4xj:~ # xfsdump -f /tmp/data_lustre_client_dump -s lustre-client-el5.8-2.1.6 /data -L data_lustre_client_dump -M media0 (注意子目录前面不能有/,最后面也不能有/)

 

xfsdump: using file dump (drive_simple)strategy

xfsdump: version 3.0.5 (dump format 3.0) -Running single-threaded

xfsdump: level 0 dump of linux-x4xj:/data

xfsdump: dump date: Tue Sep 16 21:07:14 2014

xfsdump: session id:12d13ad4-47d2-4ea5-94df-afc2d71cb6a4

xfsdump: session label:"data_lustre_client_dump"

xfsdump: ino map phase 1: constructinginitial dump list

xfsdump: ino map phase 2: skipping (nopruning necessary)

xfsdump: ino map phase 3: skipping (onlyone dump stream)

xfsdump: ino map construction complete

xfsdump: estimated dump size: 68834880bytes

xfsdump: creating dump session media file 0(media 0, file 0)

xfsdump: dumping ino map

xfsdump: dumping directories

xfsdump: dumping non-directory files

xfsdump: ending media file

xfsdump: media file size 68845344 bytes

xfsdump: dump size (non-dir files) :68822096 bytes

xfsdump: dump complete: 0 seconds elapsed

xfsdump: Dump Status: SUCCESS

linux-x4xj:~ # ls -l /tmp/data_lustre_client_dump

-rw-r--r-- 1 root root 68845344 Sep 1621:07 /tmp/data_lustre_client_dump

linux-x4xj:~ #

 

◆克隆/data(备份所有内容到另一个目录或到另一个硬盘下),本次新加了一个盘用作备份,挂载到了/data2下,开始克隆,命令如下

linux-x4xj:~ #xfsdump - /data | xfsrestore - /data2

                             或者

linux-x4xj:~ #xfsdump -l 0 - /data | xfsrestore - /data2 (因为默认级别为0)

linux-x4xj:~ # ll /data

total 1792

-rwxr--r-- 1 nobodynobody 1832960 Sep 13 14:12 iozone3_429.tar

drwxr-xr-x 2 nobodynobody     102 Sep 16 20:46lustre-client-el5.8-2.1.6

linux-x4xj:~ # ll /data2/

total 1792

-rwxr--r-- 1 nobodynobody 1832960 Sep 13 14:12 iozone3_429.tar

drwxr-xr-x 2 nobodynobody     102 Sep 16 20:46lustre-client-el5.8-2.1.6

linux-x4xj:~ #

 

◆只克隆/data下的子目录lustre-client-el5.8-2.1.6到/data2目录下,命令如下

linux-x4xj:~ #xfsdump - -s lustre-client-el5.8-2.1.6 /data | xfsrestore - /data2

                            或者

linux-x4xj:~ #xfsdump -l 0 - -s lustre-client-el5.8-2.1.6 /data | xfsrestore - /data2

linux-x4xj:~ # ll /data2/

total 0

drwxr-xr-x 2 nobodynobody 102 Sep 16 20:46 lustre-client-el5.8-2.1.6

linux-x4xj:~ #

 

3.查看会话存档

可以使用xfsdump -I来查看session存档

linux-x4xj:~ #xfsdump -I    (从存档中选择了一个会话内容如下)

        session 2:

                mount point:    linux-x4xj:/data

                device:         linux-x4xj:/dev/sdb

                time:           Tue Sep 16 20:58:58 2014

                session label:  "data_dump"                   (会话标识)

                session id:     ca3390ef-1a8c-429b-8d87-4e60d22d3c54

                level:          0

                resumed:        NO

                subtree:        NO

                streams:        1

                stream 0:

                        pathname:       /tmp/data_dump

                        start:          ino 131 offset 0

                        end:            ino 136 offset 0

                        interrupted:    NO

                        media files:    1

                        media file 0:

                                mfileindex:    0

                                mfiletype:     data

                                mfilesize:     70680352

                                mfilestart:    ino 131 offset 0

                                mfile end:      ino 136 offset 0

                                medialabel:    "media0"             (介质标识)

                                media id:       43881f83-231f-4835-a858-4d4ade150efe

 

4.使用xfsrestore恢复数据

★模拟删除/data目录内的所有内容,然后恢复数据

linux-x4xj:~ # rm -rf /data/*

linux-x4xj:~ #xfsrestore -f /tmp/data_dump /data

★根据会话标识,只恢复某个文件或者是子目录

linux-x4xj:~#xfsrestore-f /tmp/data_dump -L data_dump -s lustre-client-el5.8-2.1.6 /data

 

5.在没有进行恢复之前可以通过xfsrestore  -f /tmp/data_dump  -t查看data_dump内容

转载于:https://www.cnblogs.com/wajika/p/6389204.html

你可能感兴趣的文章
湘潭邀请赛——Alice and Bob
查看>>
js设置定时器
查看>>
数据库除运算
查看>>
LeetCode--112--路径总和
查看>>
DeviceIOControl与驱动层 - 缓冲区模式
查看>>
感悟贴2016-05-13
查看>>
vim使用教程
查看>>
JDK在LINUX系统平台下的部署案例与总结
查看>>
跨vlan通信-----单臂路由技术
查看>>
百度编辑器ueditor 光标位置的坐标
查看>>
DEV-C++ 调试方法简明图文教程(转)
查看>>
VS2017+EF+Mysql生成实体数据模型(解决闪退的坑)
查看>>
C++多态、继承的简单分析
查看>>
库克称未来苹果用户可自己决定是否降频 网友:你是在搞笑吗?
查看>>
6倍性能差100TB容量,阿里云POLARDB咋实现?
查看>>
linux 安装 MySQLdb for python
查看>>
Sublime Text 2 技巧
查看>>
使用fscanf()函数从磁盘文件读取格式化数据
查看>>
网站一些error_log报错
查看>>
参加婚礼
查看>>