LVM管理 | Shen Peng's Blog

环境

Oracle Linux 7.x
服务器 DL380 Gen8
光纤卡 HPE 82Q 8Gb Dual Port PCI-e FC HBA
阵列柜 MSA2040

理论

物理存储介质(The physical media)
如磁盘或LUN等，比如系统认出来的/dev/sda

物理卷(physical volume)
物理卷就是指磁盘分区或从逻辑上与磁盘分区具有同样功能的设备（如RAID），是LVM的基本存储逻辑块，包含LVM的管理信息

卷组(Volume Group)
LVM卷组由一个或多个物理卷组成，可以在卷组上创建一个或多个“LVM分区”（逻辑卷）
卷组的作用类似于非LVM系统中的磁盘

逻辑卷(Logical Volume)
在逻辑卷之上可以建立文件系统（比如/home或者/usr等）
逻辑卷类似于非LVM系统中的硬盘分区

PE(physical extent)
PE是物理卷的基本单位，默认为4MB

LE(logical extent)
LE是逻辑卷的基本单位，在同一个卷组中，LE的大小和PE是相同的，并且一一对应

配置

本例中做完多路径映射的LUN为/dev/mapper/dblun01

创建物理卷

[root@labdb db12ccdb]# pvcreate /dev/mapper/dblun01
  Physical volume "/dev/mapper/dblun01" successfully created.
[root@labdb db12ccdb]# pvdisplay
  "/dev/mapper/dblun01" is a new physical volume of "6.54 TiB"
  --- NEW Physical volume ---
  PV Name               /dev/mapper/dblun01
  VG Name               
  PV Size               6.54 TiB
  Allocatable           NO
  PE Size               0   
  Total PE              0
  Free PE               0
  Allocated PE          0
  PV UUID               c3mo3X-Z1re-8uwq-mc2f-vTBN-JPsT-HDPkhg

创建卷组

[root@labdb db12ccdb]# vgcreate dblun01 /dev/mapper/dblun01
  Volume group "dblun01" successfully created
[root@labdb db12ccdb]# vgdisplay   
  --- Volume group ---
  VG Name               dblun01
  System ID             
  Format                lvm2
  Metadata Areas        1
  Metadata Sequence No  1
  VG Access             read/write
  VG Status             resizable
  MAX LV                0
  Cur LV                0
  Open LV               0
  Max PV                0
  Cur PV                1
  Act PV                1
  VG Size               6.54 TiB
  PE Size               4.00 MiB
  Total PE              1714705
  Alloc PE / Size       0 / 0   
  Free  PE / Size       1714705 / 6.54 TiB
  VG UUID               ew0syV-ncdU-gta0-4KRX-iDyS-11r2-8xP2yf

创建逻辑卷

创建固定大小的逻辑卷

在卷组dblun01上创建名为oradata 大小为100G的逻辑卷

1	lvcreate -L 100G -n oradata dblun01

创建固定百分比大小的逻辑卷

在卷组dblun01上创建名为oradata 大小卷组50%的逻辑卷

1	lvcreate -l 50%VG -n oradata dblun01

创建使用了卷组所有剩余空间的逻辑卷

在卷组dblun01上创建名为oradata 大小为卷组所有剩余空间的逻辑卷

1	lvcreate -l 100%FREE -n oradata dblun01

删除逻辑卷

首先要umount分区
如果不能删除，需要用fuser先找到哪些线程在用，然后删除

[root@labdb db12cn2]# fuser -v /u05/oradata/db12cn2
                     USER        PID ACCESS COMMAND
/u05/oradata/db12cn2:
                     root     kernel mount /u05/oradata/db12cn2
                     root      24330 ..c.. bash
                     root      31344 ..c.. bash

去掉/etc/fstab里面的开机挂载项

删除不需要的逻辑卷

[root@labdb ~]# lvdisplay
  --- Logical volume ---
  LV Path                /dev/db/db-u09
  LV Name                db-u09
  VG Name                db
  LV UUID                1eEoLq-Svkg-HdUY-yvUU-g599-kVDo-ekJJxE
  LV Write Access        read/write
  LV Creation host, time labdb, 2019-09-24 11:41:57 +0800
  LV Status              available
  # open                 0
  LV Size                10.00 GiB
  Current LE             2560
  Segments               1
  Allocation             inherit
  Read ahead sectors     auto
  - currently set to     8192
  Block device           249:12
   
[root@labdb ~]# lvremove /dev/db/db-u09
Do you really want to remove active logical volume db/db-u09? [y/n]: y
  Logical volume "db-u09" successfully removed

格式化逻辑卷

将逻辑卷/dev/mapper/dblun01-oradata看作一个磁盘，使用前需要格式化

[root@labdb ~]# mkfs.ext4 /dev/mapper/dblun01-oradata
mke2fs 1.42.9 (28-Dec-2013)
Discarding device blocks: done                            
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
Stride=0 blocks, Stripe width=256 blocks
219484160 inodes, 1755857920 blocks
87792896 blocks (5.00%) reserved for the super user
First data block=0
Maximum filesystem blocks=3904897024
53585 block groups
32768 blocks per group, 32768 fragments per group
4096 inodes per group
Superblock backups stored on blocks: 
	32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 
	4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968, 
	102400000, 214990848, 512000000, 550731776, 644972544

Allocating group tables: done                            
Writing inode tables: done                            
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done

挂载并使用

挂载逻辑卷到某目录，并使用

1	[root@labdb ~]# mount /dev/mapper/dblun01-oradata /u03/oradata/db12ccdb/

查看挂载情况

[root@labdb ~]# df -h
Filesystem                   Size  Used Avail Use% Mounted on
devtmpfs                      63G     0   63G   0% /dev
tmpfs                         63G     0   63G   0% /dev/shm
tmpfs                         63G   27M   63G   1% /run
tmpfs                         63G     0   63G   0% /sys/fs/cgroup
/dev/mapper/os-root           20G  4.1G   16G  21% /
/dev/sda2                   1014M  204M  811M  21% /boot
tmpfs                         13G   12K   13G   1% /run/user/42
tmpfs                         13G     0   13G   0% /run/user/0
/dev/mapper/dblun01-oradata  6.5T   89M  6.2T   1% /u03/oradata/db12ccdb

挂载的时候用fdisk -l查出来的/dev/mapper/dblun01-oradata
或lvdisplay查出来的/dev/dblun01/oradata都是可以的

扩展逻辑卷

使用lvextend -L来扩展逻辑卷的大小，+号代表增加

1
2
3

[root@labdb oradata]# lvextend -L +60G /dev/db/db-oracle
  Size of logical volume db/db-oracle changed from 20.00 GiB (5120 extents) to 80.00 GiB (20480 extents).
  Logical volume db/db-oracle successfully resized.

查看发现依然还是扩容前的大小

[root@labdb oradata]# df -h
...
/dev/mapper/db-db--oracle     20G   13G  5.9G  69% /u01
...

此时需要用resize2fs使扩容生效，或者在lvextend时使用-r参数自动使配置生效，如lvextend -r -L +60G /dev/db/db-oracle

[root@labdb oradata]# resize2fs /dev/db/db-oracle
resize2fs 1.42.9 (28-Dec-2013)
Filesystem at /dev/db/db-oracle is mounted on /u01; on-line resizing required
old_desc_blocks = 2, new_desc_blocks = 5
Performing an on-line resize of /dev/db/db-oracle to 20971520 (4k) blocks.
The filesystem on /dev/db/db-oracle is now 20971520 blocks long.

要将vg上剩余的空间都用于lv扩展，如下

[root@labdb oradata]# lvextend -r -l +100%FREE /dev/db/db-arch
  Size of logical volume db/db-arch changed from 50.00 GiB (12800 extents) to <138.36 GiB (35419 extents).
  Logical volume db/db-arch successfully resized.
resize2fs 1.42.9 (28-Dec-2013)
Filesystem at /dev/mapper/db-db--arch is mounted on /u99/oradata/archive; on-line resizing required
old_desc_blocks = 7, new_desc_blocks = 18
The filesystem on /dev/mapper/db-db--arch is now 36269056 blocks long.

History

v1.0 2020.03.12~2020.03.23，初始版本