Extending Partitions and File Systems for System Disks (Linux)

Scenarios

After a disk is expanded on the management console, the disk size is enlarged, but the additional space cannot be used directly.

In Linux, you must allocate the additional space to an existing partition or a new partition.

If the disk capacity is expanded when its server is stopped, the additional space of a Linux system disk will be automatically added to the partition at the disk end upon the server startup. In this case, the additional space can be used directly.

This section uses CentOS 7.4 64bit to describe how to extend the disk partition using growpart and fdisk. The method is used for ECS only. The method for allocating the additional space varies with the server OS. This section is used for reference only. For detailed operations and differences, see the corresponding OS documents.

For how to extend partitions and file systems on a BMS system disk, see How Do I Increase the Size of the Root Partition of a BMS Which Is Quickly Provisioned?

• Extending an Existing MBR Partition
• Creating a New MBR Partition

Incorrect operations may lead to data loss or exceptions. So you are advised to back up the disk data using CBR or snapshots before expansion. Incorrect operations may lead to data loss or exceptions. So you are advised to back up the disk data using VBS or snapshots before expansion.

Notes and Constraints

• The additional space of a data disk cannot be added to the root partition. To extend the root partition, expand the system disk instead.
• During an expansion, the additional space is added to the end of the disk. If the disk has multiple partitions, the additional space can only be allocated to the partition at the disk end.
• If a disk uses MBR, the storage space in excess of 2 TiB cannot be used because the maximum capacity that MBR supports is 2 TiB. If your disk already uses MBR for partitioning and you require more than 2 TiB after the capacity expansion, do as follows:
  • (Recommended) Create a new EVS disk and use GPT.
  • Back up the disk data, perform the expansion, and then change the partition style from MBR to GPT. During this change, services will be interrupted and data on the disk will be erased.

Prerequisites

• You have expanded the disk capacity and attached the disk to a server on the management console.
• You have logged in to the ECS.
• The Linux kernel version is 3.6.0 or later.

  You can run uname -a to check the Linux kernel version. If your kernel version is earlier than 3.6.0, see Extending Disk Partitions and File Systems (Linux Kernel Earlier Than 3.6.0).

Extending an Existing MBR Partition

CentOS 7.4 64bit is used as the sample OS. Originally, system disk /dev/vda has 40 GiB and one partition (/dev/vda1), and then 60 GiB is added to the disk. The following procedure shows you how to allocate the additional 60 GiB to the existing MBR partition /dev/vda1.

1. Run the growpart command to check whether growpart has been installed.

   If the tool usage instructions are returned, the tool has been installed, and you do not need to install it again. Skip step 2.

2. Run the following command to install the growpart tool:

   yum install cloud-utils-growpart

3. Run the following command to view the total capacity of the /dev/vda system disk:

   fdisk -l

   Information similar to the following is displayed:

   [root@ecs-test-0001 ~]# fdisk -l
   
   Disk /dev/vda: 107.4 GiB, 107374182400 bytes, 209715200 sectors
   Units = sectors of 1 * 512 = 512 bytes
   Sector size (logical/physical): 512 bytes / 512 bytes
   I/O size (minimum/optimal): 512 bytes / 512 bytes
   Disk label type: dos
   Disk identifier: 0x000bcb4e
   
      Device Boot      Start         End      Blocks   Id  System
   /dev/vda1   *        2048    83886079    41942016   83  Linux

4. Run the following command to view the capacity of the /dev/vda1 partition:

   df -TH

   Information similar to the following is displayed:

   [root@ecs-test-0001 ~]# df -TH
   Filesystem     Type      Size  Used Avail Use% Mounted on
   /dev/vda1      ext4       43G  2.0G   39G   5% /
   devtmpfs       devtmpfs  2.0G     0  2.0G   0% /dev
   tmpfs          tmpfs     2.0G     0  2.0G   0% /dev/shm
   tmpfs          tmpfs     2.0G  9.0M  2.0G   1% /run
   tmpfs          tmpfs     2.0G     0  2.0G   0% /sys/fs/cgroup
   tmpfs          tmpfs     398M     0  398M   0% /run/user/0

5. Run the following command to extend the partition using growpart:

   growpart System disk Partition number

   In this example, run the following command:

   growpart /dev/vda 1

   Information similar to the following is displayed:

   [root@ecs-test-0001 ~]# growpart /dev/vda 1
   CHANGED: partition=1 start=2048 old: size=83884032 end=83886080 new: size=209713119,end=209715167

   

   If error message "NOCHANGE:partition 1 is size xxxxxxx. it cannot be grown" is returned, the expansion may be failed because the server disk is full (100% usage). Back up the disk data and clear necessary files or programs.

6. Run the following command to extend the file system of the partition:

   resize2fs Disk partition

   In this example, run the following command:

   resize2fs /dev/vda1

   Information similar to the following is displayed:

   [root@ecs-test-0001 ~]# resize2fs /dev/vda1
   resize2fs 1.42.9 (28-Dec-2013)
   Filesystem at /dev/vda1 is mounted on /; on-line resizing required
   old_desc_blocks = 5, new_desc_blocks = 13
   The filesystem on /dev/vda1 is now 26214139 blocks long.

   

   If the error message "open: No such file or directory while opening /dev/vdb1" is returned, an incorrect partition is specified. Run df -TH to view the disk partitions.

7. Run the following command to view the new capacity of the /dev/vda1 partition:

   df -TH

   Information similar to the following is displayed:

   [root@ecs-test-0001 ~]# df -TH
   Filesystem     Type      Size  Used Avail Use% Mounted on
   /dev/vda1      ext4      106G  2.0G   99G   2% /
   devtmpfs       devtmpfs  2.0G     0  2.0G   0% /dev
   tmpfs          tmpfs     2.0G     0  2.0G   0% /dev/shm
   tmpfs          tmpfs     2.0G  9.0M  2.0G   1% /run
   tmpfs          tmpfs     2.0G     0  2.0G   0% /sys/fs/cgroup
   tmpfs          tmpfs     398M     0  398M   0% /run/user/0

Creating a New MBR Partition

Originally, system disk /dev/vda has 40 GiB and one partition (/dev/vda1), and then 40 GiB is added to the disk. The following procedure shows you how to create a new MBR partition /dev/vda2 with this 40 GiB.

1. Run the following command to view the disk partition information:

   fdisk -l

   Information similar to the following is displayed:

   [root@ecs-2220 ~]# fdisk -l
   
   Disk /dev/vda: 85.9 GiB, 85899345920 bytes, 167772160 sectors
   Units = sectors of 1 * 512 = 512 bytes
   Sector size (logical/physical): 512 bytes / 512 bytes
   I/O size (minimum/optimal): 512 bytes / 512 bytes
   Disk label type: dos
   Disk identifier: 0x0008d18f
   
      Device Boot      Start         End      Blocks   Id  System
   /dev/vda1   *        2048    83886079    41942016   83  Linux

   In the command output, the capacity of the dev/vda system disk is 80 GiB, in which the in-use dev/vda1 partition takes 40 GiB and the additional 40 GiB has not been allocated.

2. Run the following command to enter fdisk:

   fdisk /dev/vda

   Information similar to the following is displayed:

   [root@ecs-2220 ~]# fdisk /dev/vda
   Welcome to fdisk (util-linux 2.23.2).
   
   Changes will remain in memory only, until you decide to write them.
   Be careful before using the write command.
   
   
   Command (m for help): 

3. Enter n and press Enter to create a new partition.
   Information similar to the following is displayed:

   Command (m for help): n
   Partition type:
      p   primary (1 primary, 0 extended, 3 free)
      e   extended

   There are two types of disk partitions:

   • Choosing p creates a primary partition.
   • Choosing e creates an extended partition.

   

   If MBR is used, a maximum of four primary partitions, or three primary partitions plus one extended partition can be created. The extended partition must be divided into logical partitions before use.

   Disk partitions created using GPT are not categorized.

4. In this example, a primary partition is created. Therefore, enter p and press Enter to create a primary partition.
   Information similar to the following is displayed:

   Select (default p): p
   Partition number (2-4, default 2): 

5. Enter the serial number of the primary partition and press Enter. Partition number 2 is used in this example. Therefore, enter 2 and press Enter.
   Information similar to the following is displayed:

   Partition number (2-4, default 2): 2
   First sector (83886080-167772159, default 83886080):

6. Enter the new partition's start sector and press Enter. In this example, the default start sector is used.

   The system displays the start and end sectors of the partition's available space. You can customize the value within this range or use the default value. The start sector must be smaller than the partition's end sector.

   Information similar to the following is displayed:

   First sector (83886080-167772159, default 83886080):
   Using default value 83886080
   Last sector, +sectors or +size{K,M,G} (83886080-167772159,default 167772159):

7. Enter the new partition's end sector and press Enter. In this example, the default end sector is used.

   The system displays the start and end sectors of the partition's available space. You can customize the value within this range or use the default value. The start sector must be smaller than the partition's end sector.

   Information similar to the following is displayed:

   Last sector, +sectors or +size{K,M,G} (83886080-167772159,
   default 167772159):
   Using default value 167772159
   Partition 2 of type Linux and of size 40 GiB is set
   
   Command (m for help): 

8. Enter p and press Enter to view the new partition.
   Information similar to the following is displayed:

   Command (m for help): p
   
   Disk /dev/vda: 85.9 GiB, 85899345920 bytes, 167772160 sectors
   Units = sectors of 1 * 512 = 512 bytes
   Sector size (logical/physical): 512 bytes / 512 bytes
   I/O size (minimum/optimal): 512 bytes / 512 bytes
   Disk label type: dos
   Disk identifier: 0x0008d18f
   
      Device Boot      Start         End      Blocks   Id  System
   /dev/vda1   *        2048    83886079    41942016   83  Linux
   /dev/vda2        83886080   167772159    41943040   83  Linux
   Command (m for help): 

9. Enter w and press Enter to write the changes to the partition table.
   Information similar to the following is displayed:

   Command (m for help): w
   The partition table has been altered!
   
   Calling ioctl() to re-read partition table.
   
   WARNING: Re-reading the partition table failed with error 16: Device or resource busy.
   The kernel still uses the old table. The new table will be used at
   the next reboot or after you run partprobe(8) or kpartx(8)
   Syncing disks.

   The partition is created.

   

   In case that you want to discard the changes made before, you can exit fdisk by entering q.

10. Run the following command to synchronize the new partition table to the OS:

    partprobe

11. Run the following command to set the file system format for the new partition:

    mkfs -t File system Disk partition

    • Sample command of the ext* file system:

      (The ext4 file system is used in this example.)

      mkfs -t ext4 /dev/vda2

    • Sample command of the xfs file system:

      mkfs -t xfs /dev/vda2

      Information similar to the following is displayed:

      [root@ecs-2220 ~]# mkfs -t xfs /dev/vda2 
      meta-data=/dev/vda2              isize=512     agcount=4, agsize=2621440 blks
               =                       sectsz=512    attr=2, projid32bit=1
               =                       crc=1         finobt=0, sparse=0
      data     =                       bsize=4096    blocks=10485760, imaxpct=25
               =                       sunit=0       swidth=0 blks
      naming   =version2               bsize=4096    ascii-ci=0 ftype=1
      log      =internal log           bsize=4096    blocks=5120, version=2
               =                       sectsz=512    sunit=0 blks, lazy-count=1
      realtime =none                   extsz=4096    blocks=0, rtextents=0

    The formatting takes a while, and you need to observe the system running status. Once done is displayed in the command output, the formatting is complete.

12. (Optional) Run the following command to create a mount point:

    Perform this step if you want to mount the partition on a new mount point.

    mkdir Mount point

    In this example, run the following command to create the /opt mount point:

    mkdir /opt

13. Run the following command to mount the new partition:

    mount Disk partition Mount point

    In this example, run the following command to mount the new partition /dev/vda2 on /opt:

    mount /dev/vda2 /opt

    

    If the new partition is mounted on a directory that is not empty, the subdirectories and files in the directory will be hidden. Therefore, you are advised to mount the new partition on an empty directory or a new directory. If the new partition must be mounted on a directory that is not empty, move the subdirectories and files in this directory to another directory temporarily. After the partition is successfully mounted, move the subdirectories and files back.

14. Run the following command to view the mount result:

    df -TH

    Information similar to the following is displayed:

    [root@ecs-2220 ~]# df -TH
    Filesystem     Type      Size  Used Avail Use% Mounted on
    /dev/vda1      ext4       43G  2.0G   39G   5% /
    devtmpfs       devtmpfs  509M     0  509M   0% /dev
    tmpfs          tmpfs     520M     0  520M   0% /dev/shm
    tmpfs          tmpfs     520M  7.2M  513M   2% /run
    tmpfs          tmpfs     520M     0  520M   0% /sys/fs/cgroup
    tmpfs          tmpfs     104M     0  104M   0% /run/user/0
    /dev/vda2      ext4       43G   51M   40G   1% /opt

    

    If the server is restarted, the mounting will become invalid. You can modify the /etc/fstab file to configure automount at startup. See the following part for details.

Configuring Automatic Mounting at System Start

The fstab file controls what disks are automatically mounted at ECS startup. You can configure the fstab file of an ECS that has data. This operation will not affect the existing data.

The following example uses UUIDs to identify disks in the fstab file. You are advised not to use device names (like /dev/vdb1) to identify disks in the file because device names are assigned dynamically and may change (for example, from /dev/vdb1 to /dev/vdb2) after an ECS stop or start. This can even prevent your ECS from booting up.

UUIDs are the unique character strings for identifying partitions in Linux.

1. Query the partition UUID.

   blkid Disk partition

   In this example, the UUID of the /dev/vdb1 partition is queried.

   blkid /dev/vdb1

   Information similar to the following is displayed:

   [root@ecs-test-0001 ~]# blkid /dev/vdb1
   /dev/vdb1: UUID="0b3040e2-1367-4abb-841d-ddb0b92693df" TYPE="ext4"

   Carefully record the UUID, as you will need it for the following step.

2. Open the fstab file using the vi editor.

   vi /etc/fstab

3. Press i to enter editing mode.
4. Move the cursor to the end of the file and press Enter. Then, add the following information:

   UUID=0b3040e2-1367-4abb-841d-ddb0b92693df /mnt/sdc                ext4    defaults        0 2

   The preceding information is used for reference only. The line starting with UUID is the information added. Edit this line from left to right to match the following format:

   • UUID: The UUID obtained in 1.
   • Mount point: The directory on which the partition is mounted. You can query the mount point using df -TH.
   • Filesystem: The file system format of the partition. You can query the file system format using df -TH.
   • Mount option: The partition mount option. Usually, this parameter is set to defaults.
   • Dump: The Linux dump backup option.
     • 0: Linux dump backup is not used. Usually, dump backup is not used, and you can set this parameter to 0.
     • 1: Linux dump backup is used.
   • fsck: The fsck option, which means whether to use fsck to check the disk during startup.
     • 0: not use fsck.
     • If the mount point is the root partition (/), this parameter must be set to 1.

       If this parameter is set to 1 for the root partition, this parameter for other partitions must start with 2 because the system checks the partitions in the ascending order of the values.

5. Press Esc, enter :wq, and press Enter.

   The system saves the configurations and exits the vi editor.

6. Verify that the disk is auto-mounted at startup.
   1. Unmount the partition.

      umount Disk partition

      In this example, run the following command:

      umount /dev/vdb1

   2. Reload all the content in the /etc/fstab file.

      mount -a

   3. Query the file system mounting information.

      mount | grep Mount point

      In this example, run the following command:

      mount | grep /mnt/sdc

      If information similar to the following is displayed, automatic mounting has been configured:

      root@ecs-test-0001 ~]# mount | grep /mnt/sdc
      /dev/vdb1 on /mnt/sdc type ext4 (rw,relatime,data=ordered)