sisimi的点点滴滴

你若盛开,清风自来

导航

组装恢复rbd

标签: ceph,ceph实验,rbd


cluster相关环境:

# cat /etc/redhat-release 
CentOS Linux release 7.3.1611 (Core) 
# ceph -v
ceph version 12.2.1 (3e7492b9ada8bdc9a5cd0feafd42fbca27f9c38e) luminous (stable)

比如需要恢复所有前缀为 rbd_data.1041643c9869 的块设备,该块设备名为foo,文件系统类型为xfs,大小1G,挂载到了/root/foocon/目录下,并有三个文件:

[root@node3 ~]# ls foocon/
file1.txt  file2.txt  file3.txt
[root@node3 ~]# cat foocon/file2.txt
Ceph
2222222222222222222222222222222222222

恢复该foo块设备需要进行下面的操作

1. 从集群下载前缀相同的所有对象到本地:

1.1 查看rbd存储池中的对象:

[root@node3 ~]# rados -p rbd ls|sort
rbd_data.1041643c9869.0000000000000000
rbd_data.1041643c9869.0000000000000001
rbd_data.1041643c9869.000000000000001f
rbd_data.1041643c9869.000000000000003e
rbd_data.1041643c9869.000000000000005d
rbd_data.1041643c9869.000000000000007c
rbd_data.1041643c9869.000000000000007d
rbd_data.1041643c9869.000000000000007e
rbd_data.1041643c9869.000000000000009b
rbd_data.1041643c9869.00000000000000ba
rbd_data.1041643c9869.00000000000000d9
rbd_data.1041643c9869.00000000000000f8
rbd_data.1041643c9869.00000000000000ff
rbd_directory
rbd_header.1041643c9869
rbd_id.foo
rbd_info

1.2 从集群获取前缀为 rbd_data.1041643c9869 的所有对象到本地:

如:

[root@node3 ~]# rados -p rbd get rbd_data.1041643c9869.0000000000000000 rbd_data.1041643c9869.0000000000000000
[root@node3 ~]# rados -p rbd get rbd_data.1041643c9869.0000000000000001 rbd_data.1041643c9869.0000000000000001
……

2. 创建脚本组装该rbd:

2.1 创建 mkrbd.sh 脚本:

[root@node3 ~]# touch mkrbd.sh
[root@node3 ~]# chmod +x mkrbd.sh

2.2 编辑脚本:

[root@node3 ~]# vi mkrbd.sh

输入以下内容:

#!/bin/sh
# Rados object size 这是刚刚的4M的大小
obj_size=4194304
# DD bs value
rebuild_block_size=512
#rbd="${1}" 
rbd="foo"  #生成的块名
#base="${2}" #prefix
base="rbd_data.1041643c9869"
#rbd_size="${3}" #1G
rbd_size="1073741824" 
base_files=$(ls -1 ${base}.* 2>/dev/null | wc -l | awk '{print $1}')
if [ ${base_files} -lt 1 ]; then
  echo "COULD NOT FIND FILES FOR ${base} IN $(pwd)"
  exit
fi
# Create full size sparse image.  Could use truncate, but wanted
# as few required files and dd what a must.
dd if=/dev/zero of=${rbd} bs=1 count=0 seek=${rbd_size} 2>/dev/null
for file_name in $(ls -1 ${base}.* 2>/dev/null); do
  seek_loc=$(echo ${file_name} | awk -v os=${obj_size} -v rs=${rebuild_block_size} -F. '{print os*strtonum("0x" $NF)/rs}')
  dd conv=notrunc if=${file_name} of=${rbd} seek=${seek_loc} bs=${rebuild_block_size} 2>/dev/null
done

执行该脚本后在本地生成了一个foo的文件

2.3 查看挂载组装后的foo:

[root@node3 ~]# file foo
foo: SGI XFS filesystem data (blksz 4096, inosz 512, v2 dirs)
[root@node3 ~]# du -sh foo
11M     foo
[root@node3 ~]# ll -h foo
-rw-r--r-- 1 root root 1.0G 10月 17 16:04 foo

可以看出foo是xfs文件,使用了11M,大小为1G

[root@node3 ~]# mount foo /mnt
mount: 文件系统类型错误、选项错误、/dev/loop0 上有坏超级块、
       缺少代码页或助手程序,或其他错误

       有些情况下在 syslog 中可以找到一些有用信息- 请尝试
       dmesg | tail  这样的命令看看。
[root@node3 ~]# dmesg|tail
[88818.307314] XFS (rbd0): Mounting V5 Filesystem
[88818.865978] XFS (rbd0): Ending clean mount
[91099.845834] bash (6208): drop_caches: 1
[91492.345582] bash (6208): drop_caches: 1
[93485.275727] libceph: osd2 down
[93485.275739] libceph: osd5 down
[93495.518099] libceph: osd2 up
[93495.518165] libceph: osd5 up
[95288.897917] loop: module loaded
[98449.535689] XFS (loop0): Filesystem has duplicate UUID 313b2d89-f4bc-4ee6-a1d8-a996190222fd - can't mount

挂载foo时出现了上面的错误提示,这是因为原来的foo块是/dev/rbd0的克隆,所以foo的UUID是和/dev/rbd0的是一样的,这时候我们umount /foo即可:

[root@node3 ~]# umount foo/
[root@node3 ~]# mount foo /mnt
[root@node3 ~]# ls /mnt/
file1.txt  file2.txt  file3.txt
[root@node3 ~]# cat /mnt/file2.txt
Ceph
2222222222222222222222222222222222222

可以看到已将完全恢复出了原来的foo块设备的内容

参考文献:

大话Ceph--RBD那点事儿

posted on 2017-11-02 09:56  sisimi_2017  阅读(968)  评论(0编辑  收藏  举报