EXT3文件系统收缩和扩容

Seker

查看系统现在结构
[root@seker /]# fdisk -l

Disk /dev/sda: 17.1 GB, 17179869184 bytes
255 heads, 63 sectors/track, 2088 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *           1          13      104391   83  Linux
/dev/sda2              14        1057     8385930   83  Linux
/dev/sda3            1058        1122      522112+  82  Linux swap / Solaris
/dev/sda4            1123        2088     7759395    5  Extended
[root@seker /]#

[root@seker /]# df -Th
文件系统      类型    容量  已用 可用 已用% 挂载点
/dev/sda2     ext3    7.8G  3.0G  4.5G  40% /
/dev/sda1     ext3     99M   15M   80M  16% /boot
tmpfs        tmpfs    125M     0  125M   0% /dev/shm
[root@seker /]#

根据上面输出,先计算出一个值来,我们以后要用到:
sda1 = 100M 从柱面1到柱面13. 也就是说12个柱面是100M空间

1.
在逻辑分区建立一个新的分区 大小为300M

[root@seker /]# fdisk /dev/sda

The number of cylinders for this disk is set to 2088.
There is nothing wrong with that, but this is larger than 1024,
and could in certain setups cause problems with:
1) software that runs at boot time (e.g., old versions of LILO)
2) booting and partitioning software from other OSs
   (e.g., DOS FDISK, OS/2 FDISK)

Command (m for help): n
First cylinder (1123-2088, default 1123):
Using default value 1123
Last cylinder or +size or +sizeM or +sizeK (1123-2088, default 2088): +300M

Command (m for help): p

Disk /dev/sda: 17.1 GB, 17179869184 bytes
255 heads, 63 sectors/track, 2088 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *           1          13      104391   83  Linux
/dev/sda2              14        1057     8385930   83  Linux
/dev/sda3            1058        1122      522112+  82  Linux swap / Solaris
/dev/sda4            1123        2088     7759395    5  Extended
/dev/sda5            1123        1159      297171   83  Linux

Command (m for help): w

这里,再计算一下 给sda5 300M 使用柱面恰好是占用了 1159 - 1123 = 36 个.和我们前面计算的100M=12个柱面相吻合

[root@seker /]# partprobe /dev/sda
[root@seker /]# mke2fs -j /dev/sda5
让你新建的分区马上生效,之后再格式化成ext3.

[root@seker /]# mount /dev/sda5 /mnt
[root@seker /]# cp /etc/rc.d/init.d/* /mnt/
[root@seker /]# ls /mnt/ | wc -l
74
[root@seker /]# sync;sync
[root@seker /]#
将新分区挂载 并写入写文件进去.现在时74个文件在新分区上.

2.
一切准备就绪,我们先缩小它.
[root@seker /]# umount /mnt
[root@seker /]#
要想缩小或者是放大,都必须要离线.

[root@seker /]# tune2fs -O ^has_journal /dev/sda5
tune2fs 1.39 (29-May-2006)
删除分区上的日志,也就是把系统先转化成ext2格式.(也可以不做这里的转换,直接用ext3的就支持.引用man: resize2fs - ext2/ext3 file system resizer.)

[root@seker /]# e2fsck -f /dev/sda5
e2fsck 1.39 (29-May-2006)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
/dev/sda5: 84/74296 files (2.4% non-contiguous), 11687/297168 blocks
[root@seker /]#
接着对文件系统做强制检查.

[root@seker /]# resize2fs /dev/sda5 150M
resize2fs 1.39 (29-May-2006)
Resizing the filesystem on /dev/sda5 to 153600 (1k) blocks.
The filesystem on /dev/sda5 is now 153600 blocks long.

[root@seker /]# 开始收缩,收缩至150M

[root@seker /]# tune2fs -j /dev/sda5
tune2fs 1.39 (29-May-2006)
Creating journal inode: done
This filesystem will be automatically checked every 33 mounts or
180 days, whichever comes first.  Use tune2fs -c or -i to override.
[root@seker /]# 再将ext2转成ext3.


[root@seker /]# mount /dev/sda5 /mnt
[root@seker /]# df -Th /dev/sda5
文件系统      类型    容量  已用 可用 已用% 挂载点
/dev/sda5     ext3    146M  5.8M  134M   5% /mnt
[root@seker /]# ls /mnt/ | wc  -l
74
[root@seker /]# 验证大小,以及上面的文件是否还在. 一切安好.

再做写入测试
[root@seker mnt]# dd if=/dev/zero of=./big_file bs=5M count=28
dd: 写入 “./big_file”: 设备上没有空间
28+0 records in
27+0 records out
145690624 bytes (146 MB) copied, 2.58229 seconds, 56.4 MB/s
[root@seker mnt]#

[root@seker mnt]# df -Th /dev/sda5
文件系统      类型    容量  已用 可用 已用% 挂载点
/dev/sda5     ext3    146M  146M     0 100% /mnt
[root@seker mnt]# 写到146M后提示空间已满.

缩小成功.

3.放大
放大比缩小困难一点.因为缩小不需要去更改柱面位置.

[root@seker ~]# umount /mnt
[root@seker ~]# 还是必须离线

[root@seker ~]# tune2fs -O ^has_journal /dev/sda5
tune2fs 1.39 (29-May-2006)
[root@seker ~]# 还是转化成ext2

[root@seker ~]# e2fsck -f /dev/sda5
e2fsck 1.39 (29-May-2006)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
/dev/sda5: 85/38152 files (2.4% non-contiguous), 149445/153600 blocks
[root@seker ~]# 还是强制检查

[root@seker ~]# resize2fs /dev/sda5 500M
resize2fs 1.39 (29-May-2006)
The containing partition (or device) is only 297171 (1k) blocks.
You requested a new size of 512000 blocks.

[root@seker ~]# 尝试加大已经不好用了.提示只能是297171个blocks.因为分区时柱面结束为止被锁定.

我们突破一下:
[root@seker ~]# fdisk /dev/sda

The number of cylinders for this disk is set to 2088.
There is nothing wrong with that, but this is larger than 1024,
and could in certain setups cause problems with:
1) software that runs at boot time (e.g., old versions of LILO)
2) booting and partitioning software from other OSs
   (e.g., DOS FDISK, OS/2 FDISK)

Command (m for help): p

Disk /dev/sda: 17.1 GB, 17179869184 bytes
255 heads, 63 sectors/track, 2088 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *           1          13      104391   83  Linux
/dev/sda2              14        1057     8385930   83  Linux
/dev/sda3            1058        1122      522112+  82  Linux swap / Solaris
/dev/sda4            1123        2088     7759395    5  Extended
/dev/sda5            1123        1159      297171   83  Linux

Command (m for help): d
Partition number (1-5): 5
        # 将sda5删掉 不用担心,数据不会丢.之后马上新建.
Command (m for help): n
First cylinder (1123-2088, default 1123):
Using default value 1123
Last cylinder or +size or +sizeM or +sizeK (1123-2088, default 2088):
        # 结束的柱面位置是关键.我们在上面算过了100M用到12个柱面,现在我想扩大到500M
        # 那么就是从开始位置(1123)处再加上12*5的数值 结果就是1183
        # 那就在lasr cylinder:这里输出:1183

Command (m for help): p

/dev/sda5            1123        1183      489951   83  Linux
        # 查看确认一下,存盘退出.
Command (m for help): w
The partition table has been altered!

再次尝试放大
[root@seker ~]# resize2fs /dev/sda5 500M
resize2fs 1.39 (29-May-2006)
The containing partition (or device) is only 489951 (1k) blocks.
You requested a new size of 512000 blocks.

因为我们计算的是个粗略的值,毕竟存在误差.故500M无法满足,但这步能提示出能放大的最大数:
is only 489951 (1k) blocks.

[root@seker ~]# resize2fs /dev/sda5 489951K
resize2fs 1.39 (29-May-2006)
Resizing the filesystem on /dev/sda5 to 489951 (1k) blocks.
The filesystem on /dev/sda5 is now 489951 blocks long.

[root@seker ~]# 这次成功了.


[root@seker ~]# mount /dev/sda5 /mnt
[root@seker ~]# df -Th /dev/sda5
文件系统      类型    容量  已用 可用 已用% 挂载点
/dev/sda5     ext2    464M  142M  303M  32% /mnt
[root@seker ~]#  加载上来看看,已经是464M了,和我们的计算基本吻合.看看已用用量142M是我们之前的数据,还存在的. 只是类型上时ext2.

拿下来,再转换成ext3吧

[root@seker ~]# umount /mnt/
[root@seker ~]# tune2fs -j /dev/sda5
tune2fs 1.39 (29-May-2006)
Creating journal inode: done
This filesystem will be automatically checked every 33 mounts or
180 days, whichever comes first.  Use tune2fs -c or -i to override.

[root@seker ~]# mount /dev/sda5 /mnt
[root@seker ~]# df -Th /dev/sda5
文件系统      类型    容量  已用 可用 已用% 挂载点
/dev/sda5     ext3    464M  151M  295M  34% /mnt
[root@seker ~]# ls /mnt/ |wc -l
75
[root@seker ~]# 原来的文件还在.

完美了.

做写入测试:

[root@seker ~]# dd if=/dev/zero of=/mnt/seker_file_1 bs=1M count=20
20+0 records in
20+0 records out
20971520 bytes (21 MB) copied, 0.124672 seconds, 168 MB/s
[root@seker ~]# dd if=/dev/zero of=/mnt/seker_file_2 bs=1M count=20
20+0 records in
20+0 records out
20971520 bytes (21 MB) copied, 0.278653 seconds, 75.3 MB/s
[root@seker ~]# dd if=/dev/zero of=/mnt/seker_file_3 bs=1M count=20
20+0 records in
20+0 records out
20971520 bytes (21 MB) copied, 0.373639 seconds, 56.1 MB/s
[root@seker ~]# df -Th /dev/sda5
文件系统      类型    容量  已用 可用 已用% 挂载点
/dev/sda5     ext3    464M  211M  235M  48% /mnt
[root@seker ~]#

写了三个20M的文件进去.查看一下一切安好.


好久不来发帖了..欢迎一起研究.有什么问题 请多多指教..

[[i] 本帖最后由 Seker 于 2009-7-1 13:39 编辑 [/i]]

Seker

另附上man resize2fs..
DESCRIPTION
       The resize2fs program will resize ext2 or ext3 file systems.   It  can  be
       used  to enlarge or shrink an unmounted file system located on device.  If
       the filesystem is mounted, it can be  used  to  expand  the  size  of  the
       mounted filesystem, assuming the kernel supports on-line resizing.  (As of
       this writing, the Linux 2.6 kernel supports on-line resize for filesystems
       mounted using ext3 only.).

       The size parameter specifies the requested new size of the filesystem.  If
       no units are specified, the units of  the  size  parameter  shall  be  the
       filesystem  blocksize  of  the filesystem.  Optionally, the size parameter
       may be suffixed by one of the following the units designators:  ’s’,  ’K’,
       ’M’,  or  ’G’,  for  512 byte sectors, kilobytes, megabytes, or gigabytes,
       respectively.  The size of the filesystem may never  be  larger  than  the
       size  of  the  partition.   If  size  parameter  is not specified, it will
       default to the size of the partition.

       The resize2fs program does not manipulate the size of partitions.  If  you
       wish  to enlarge a filesystem, you must first make sure you can expand the
       size of the underlying partition first.  This can be done  using  fdisk(8)
       by  deleting  the  partition and recreating it with a larger size or using
       lvextend(8), if you’re using the  logical  volume  manager  lvm(8).   When
       recreating  the  partition, make sure you create it with the same starting
       disk cylinder as before!  Otherwise, the resize operation  will  certainly
       not  work,  and  you  may  lose  your  entire  filesystem.   After running
       fdisk(8), run resize2fs to resize the ext2 filesystem to use  all  of  the
       space in the newly enlarged partition.

       If you wish to shrink an ext2 partition, first use resize2fs to shrink the
       size of filesystem.  Then you may use fdisk(8) to shrink the size  of  the
       partition.  When shrinking the size of the partition, make sure you do not
       make it smaller than the new size of the ext2 filesystem!


下面这段好像说是要内核支持才能在线搞ext3的.不知道怎么弄.
       The resize2fs program will resize ext2 or ext3 file systems.   It  can  be
       used  to enlarge or shrink an unmounted file system located on device.  If
       the filesystem is mounted, it can be  used  to  expand  the  size  of  the
       mounted filesystem, assuming the kernel supports on-line resizing.  (As of
       this writing, the Linux 2.6 kernel supports on-line resize for filesystems
       mounted using ext3 only.).

整体来说 这部是关键,虽然我在上面收缩时并没有去改结束的柱面位置,但也可以达到目的了.到达指定大小后就报警了.但不知道有什么副作用没.
This can be done  using  fdisk(8) by  deleting  the  partition and recreating it with a larger size




测试了一下在线情况...没办法搞定,,,有高手知道怎么online做给指点一下..谢了
[root@stu33 ~]# resize2fs /dev/sda5 100M
resize2fs 1.39 (29-May-2006)
Filesystem at /dev/sda5 is mounted on /mnt; on-line resizing required
On-line shrinking from 200780 to 102400 not supported.
[root@stu33 ~]# e2fsck -f /dev/sda5
e2fsck 1.39 (29-May-2006)
/dev/sda5 is mounted.  

WARNING!!!  Running e2fsck on a mounted filesystem may cause
SEVERE filesystem damage.

Do you really want to continue (y/n)? no

check aborted.
[root@stu33 ~]# tune2fs -O ^has_journal /dev/sda5
tune2fs 1.39 (29-May-2006)
The has_journal flag may only be cleared when the filesystem is
unmounted or mounted read-only.
[root@stu33 ~]#



OS版本

[root@stu33 ~]# cat /etc/issue
Red Hat Enterprise Linux Server release 5.1 (Tikanga)

webyuhang

矛盾;
因为收缩或放大的命令式resize2fs只针对ext2文件系统可用
The resize2fs program will resize ext2 or ext3 file systems.


The resize2fs program will resize ext2 or ext3 file systems.   It  can  be
       used  to enlarge or shrink an unmounted file system located on device.  If
       the filesystem is mounted, it can be  used  to  expand  the  size  of  the
       mounted filesystem, assuming the kernel supports on-line resizing.  (As of
       this writing, the Linux 2.6 kernel supports on-line resize for filesystems
       mounted using ext3 only.).
这部分说明了linux 2.6支持mount情况下的扩容

marsaber

/dev/sda1   *           1          13      104391   83  Linux
/dev/sda2              14        1057     8385930   83  Linux
/dev/sda3            1058        1122      522112+  82  Linux swap / Solaris
/dev/sda4            1123        2088     7759395    5  Extended
/dev/sda5            1123        1159      297171   83  Linux

1、sda1 = 100M 从柱面1到柱面13. 也就是说12个柱面是100M空间
2、给sda5 300M 使用柱面恰好是占用了 1159 - 1123 = 36 个.和我们前面计算的100M=12个柱面相吻合
那我问一下：
/dev/sda1   *           1          13
/dev/sda2              14        1057
/dev/sda3            1058        1122
/dev/sda4            1123        2088
/dev/sda5            1123        1159
中的第13柱面和第1123柱面去哪了？

marsaber

我觉得sda1应该是13-1+1=13个柱面
sda5应该说是1159-1123+1=37个柱面

Seker

原帖由 marsaber 于 2009-7-1 09:27 发表
/dev/sda1   *           1          13      104391   83  Linux
/dev/sda2              14        1057     8385930   83  Linux
/dev/sda3            1058        1122      522112+  82  Linux swap /  ...

约值而已.

Seker

原帖由 webyuhang 于 2009-7-1 08:02 发表
矛盾;
因为收缩或放大的命令式resize2fs只针对ext2文件系统可用
The resize2fs program will resize ext2 or ext3 file systems.


The resize2fs program will resize ext2 or ext3 file systems.   It  ...

恩,确实是支持ext3的.即使不转化成EXT2的也可以做. 我修改一下.
后面的ext3 to ext2也可以省略.

课时后面的在线的 回头我再去尝试了.



谢谢...多此一举了...呵呵.


后面说的在线 我没实现,修改后还是原来的大小,回头再去尝试一下.

[ 本帖最后由 Seker 于 2009-7-1 13:43 编辑 ]