Hi all. In anticipation of the start of a new course group
This article will consider 2 cases of replacing disks and transferring information to new larger disks with further expansion of the array and file system. The first case will concern the replacement of disks with the MBR/MBR or GPT/GPT layout of the same name, the second case will concern the replacement of disks with MBR partitions for disks larger than 2 TB, which will require GPT partitioning with a biosboot partition. In both cases, the disks to which we transfer the data are already installed on the server. The file system used for the root partition is ext4.
Case 1: Replacing smaller drives with larger drives (up to 2TB)
Problem: Replace current disks with larger disks (up to 2 TB) with information transfer. In this case, we have 2 x 240 GB SSD (RAID-1) drives with the system installed and 2 x 1 TB SATA drives to which the system needs to be transferred.
Consider the current disk layout.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sda2 8:2 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdb2 8:18 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdc 8:32 0 931,5G 0 disk
sdd 8:48 0 931,5G 0 disk
Let's check the currently used file system space.
[root@localhost ~]# df -h
Файловая система Размер Использовано Дост Использовано% Cмонтировано в
devtmpfs 32G 0 32G 0% /dev
tmpfs 32G 0 32G 0% /dev/shm
tmpfs 32G 9,6M 32G 1% /run
tmpfs 32G 0 32G 0% /sys/fs/cgroup
/dev/mapper/vg0-root 204G 1,3G 192G 1% /
/dev/md126 1007M 120M 837M 13% /boot
tmpfs 6,3G 0 6,3G 0% /run/user/0
The size of the file system before replacing the disks is 204 GB, 2 md126 software arrays are used, which is mounted in /boot
и md127
, which is used as physical volume for VG group vg0.
1. Removing disk partitions from arrays
Check the state of the array
[root@localhost ~]# cat /proc/mdstat
Personalities : [raid1]
md126 : active raid1 sda1[0] sdb1[1]
1047552 blocks super 1.2 [2/2] [UU]
bitmap: 0/1 pages [0KB], 65536KB chunk
md127 : active raid1 sda2[0] sdb2[1]
233206784 blocks super 1.2 [2/2] [UU]
bitmap: 0/2 pages [0KB], 65536KB chunk
unused devices: <none>
The system uses 2 arrays: md126
(mount point /boot
) - consists of a section /dev/sda1
и /dev/sdb1
, md127
(LVM for swap and the root of the file system) - consists of /dev/sda2
и /dev/sdb2
.
We mark the partitions of the first disk that are used in each array as bad.
mdadm /dev/md126 --fail /dev/sda1
mdadm /dev/md127 --fail /dev/sda2
Remove the /dev/sda block device partitions from the arrays.
mdadm /dev/md126 --remove /dev/sda1
mdadm /dev/md127 --remove /dev/sda2
After we have removed the disk from the array, the block device information will look like this.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
└─sda2 8:2 0 222,5G 0 part
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdb2 8:18 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdc 8:32 0 931,5G 0 disk
sdd 8:48 0 931,5G 0 disk
The state of the arrays after removing disks.
[root@localhost ~]# cat /proc/mdstat
Personalities : [raid1]
md126 : active raid1 sdb1[1]
1047552 blocks super 1.2 [2/1] [_U]
bitmap: 0/1 pages [0KB], 65536KB chunk
md127 : active raid1 sdb2[1]
233206784 blocks super 1.2 [2/1] [_U]
bitmap: 1/2 pages [4KB], 65536KB chunk
unused devices: <none>
2. Copy the partition table to a new drive
You can check the used partition table on the disk with the following command.
fdisk -l /dev/sdb | grep 'Disk label type'
The output for the MBR will be:
Disk label type: dos
for GPT:
Disk label type: gpt
Copying the partition table for MBR:
sfdisk -d /dev/sdb | sfdisk /dev/sdc
In this command first disk is indicated с whom markup is copied second - where copy.
ATTENTION: For GPT first disk is indicated on which copy markup, second disk indicates disk from which copy markup. If you mix up the disks, then the initially good partitioning will be overwritten and destroyed.
Copying the markup table for GPT:
sgdisk -R /dev/sdс /dev/sdb
Next, assign a random UUID to the disk (for GPT).
sgdisk -G /dev/sdc
After the command is executed, the partitions should appear on the disk /dev/sdc
.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
└─sda2 8:2 0 222,5G 0 part
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdb2 8:18 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdc 8:32 0 931,5G 0 disk
├─sdc1 8:33 0 1G 0 part
└─sdc2 8:34 0 222,5G 0 part
sdd 8:48 0 931,5G 0 disk
If after the action the partitions in the system on the disk /dev/sdc
undecided, then we execute the command to reread the partition table.
sfdisk -R /dev/sdc
If current disks use the MBR table and information needs to be migrated to disks larger than 2 TB, then new disks will need to manually create a GPT partition using the biosboot partition. This case will be considered in part 2 of this article.
3. Adding new disk partitions to the array
Let's add disk partitions to the appropriate arrays.
mdadm /dev/md126 --add /dev/sdc1
mdadm /dev/md127 --add /dev/sdc2
Check that the sections have been added.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
└─sda2 8:2 0 222,5G 0 part
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdb2 8:18 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdc 8:32 0 931,5G 0 disk
├─sdc1 8:33 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdc2 8:34 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdd 8:48 0 931,5G 0 disk
After that, we wait for the synchronization of the arrays.
[root@localhost ~]# cat /proc/mdstat
Personalities : [raid1]
md126 : active raid1 sdc1[2] sdb1[1]
1047552 blocks super 1.2 [2/2] [UU]
bitmap: 0/1 pages [0KB], 65536KB chunk
md127 : active raid1 sdc2[2] sdb2[1]
233206784 blocks super 1.2 [2/1] [_U]
[==>..................] recovery = 10.6% (24859136/233206784) finish=29.3min speed=118119K/sec
bitmap: 2/2 pages [8KB], 65536KB chunk
unused devices: <none>
You can continuously monitor the synchronization process using the utility watch
.
watch -n 2 cat /proc/mdstat
Parameter -n
indicates at what intervals in seconds the command must be executed to check progress.
Repeat steps 1 - 3 for the next disk to be replaced.
We mark the partitions of the second disk that are used in each array as bad.
mdadm /dev/md126 --fail /dev/sdb1
mdadm /dev/md127 --fail /dev/sdb2
Delete block device partitions /dev/sdb
from arrays.
mdadm /dev/md126 --remove /dev/sdb1
mdadm /dev/md127 --remove /dev/sdb2
After we have removed the disk from the array, the block device information will look like this.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
└─sda2 8:2 0 222,5G 0 part
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
└─sdb2 8:18 0 222,5G 0 part
sdc 8:32 0 931,5G 0 disk
├─sdc1 8:33 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdc2 8:34 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdd 8:48 0 931,5G 0 disk
The state of the arrays after removing disks.
[root@localhost ~]# cat /proc/mdstat
Personalities : [raid1]
md126 : active raid1 sdc1[2]
1047552 blocks super 1.2 [2/1] [U_]
bitmap: 0/1 pages [0KB], 65536KB chunk
md127 : active raid1 sdc2[2]
233206784 blocks super 1.2 [2/1] [U_]
bitmap: 1/2 pages [4KB], 65536KB chunk
unused devices: <none>
Copy the MBR partition table from disk /dev/sdс
to disk /dev/sdd
.
sfdisk -d /dev/sdс | sfdisk /dev/sdd
After the command is executed, the partitions should appear on the disk /dev/sdd
.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
└─sda2 8:2 0 222,5G 0 part
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
└─sdb2 8:18 0 222,5G 0 part
sdc 8:32 0 931,5G 0 disk
├─sdc1 8:33 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdc2 8:34 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdd 8:48 0 931,5G 0 disk
├─sdd1 8:49 0 1G 0 part
└─sdd2 8:50 0 222,5G 0 part
Add disk partitions to arrays.
mdadm /dev/md126 --add /dev/sdd1
mdadm /dev/md127 --add /dev/sdd2
Check that the sections have been added.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
└─sda2 8:2 0 222,5G 0 part
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
└─sdb2 8:18 0 222,5G 0 part
sdc 8:32 0 931,5G 0 disk
├─sdc1 8:33 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdc2 8:34 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdd 8:48 0 931,5G 0 disk
├─sdd1 8:49 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdd2 8:50 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
After that, we wait for the synchronization of the arrays.
[root@localhost ~]# cat /proc/mdstat
Personalities : [raid1]
md126 : active raid1 sdd1[3] sdc1[2]
1047552 blocks super 1.2 [2/2] [UU]
bitmap: 0/1 pages [0KB], 65536KB chunk
md127 : active raid1 sdd2[3] sdc2[2]
233206784 blocks super 1.2 [2/1] [U_]
[>....................] recovery = 0.5% (1200000/233206784) finish=35.4min speed=109090K/sec
bitmap: 2/2 pages [8KB], 65536KB chunk
unused devices: <none>
5. Installing GRUB on new drives
For CentOS:
grub2-install /dev/sdX
For Debian/Ubuntu:
grub-install /dev/sdX
where X
is the letter of the block device. In this case, you need to install GRUB on /dev/sdc
и /dev/sdd
.
6. Extending the file system (ext4) of the root partition
On new discs /dev/sdc
и /dev/sdd
931.5 GB available. Due to the fact that the partition table was copied from smaller disks, partitions /dev/sdc2
и /dev/sdd2
222.5 GB available.
sdc 8:32 0 931,5G 0 disk
├─sdc1 8:33 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdc2 8:34 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdd 8:48 0 931,5G 0 disk
├─sdd1 8:49 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdd2 8:50 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
It is necessary:
- Extend partition 2 on each of the drives,
- Expand array md127,
- Expand PV (physical volume),
- Expand LV (logical-volume) vg0-root,
- Expand the file system.
Using the utility parted expand the section /dev/sdc2
up to the maximum value. Execute the command parted /dev/sdc
(1) and view the current partition table with the command p
(2)
As you can see, the end of partition 2 ends at 240 GB. Let's expand the partition with the command resizepart
2
, where 2 is the section number (3). Specify the value in digital format, for example 1000 GB, or use the indication of the disk share - 100%. Check again that the partition has the new size (4).
Repeat the above steps for the disk /dev/sdd
. After Partition Expansion /dev/sdc2
и /dev/sdd2
became equal to 930.5 GB.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
└─sda2 8:2 0 222,5G 0 part
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
└─sdb2 8:18 0 222,5G 0 part
sdc 8:32 0 931,5G 0 disk
├─sdc1 8:33 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdc2 8:34 0 930,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdd 8:48 0 931,5G 0 disk
├─sdd1 8:49 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdd2 8:50 0 930,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
After that, we expand the array md127 up to the maximum.
mdadm --grow /dev/md127 --size=max
We check that the array has expanded. Now its size has become 930.4 GB.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
└─sda2 8:2 0 222,5G 0 part
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
└─sdb2 8:18 0 222,5G 0 part
sdc 8:32 0 931,5G 0 disk
├─sdc1 8:33 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdc2 8:34 0 930,5G 0 part
└─md127 9:127 0 930,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdd 8:48 0 931,5G 0 disk
├─sdd1 8:49 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdd2 8:50 0 930,5G 0 part
└─md127 9:127 0 930,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
Performing an extension physical volume. Before expanding, check the current state of the PV.
[root@localhost ~]# pvscan
PV /dev/md127 VG vg0 lvm2 [222,40 GiB / 0 free]
Total: 1 [222,40 GiB] / in use: 1 [222,40 GiB] / in no VG: 0 [0 ]
As you can see, PV /dev/md127
uses 222.4 GB of space.
Expand PV with the following command.
pvresize /dev/md127
Checking the result of the PV expansion.
[root@localhost ~]# pvscan
PV /dev/md127 VG vg0 lvm2 [930,38 GiB / 707,98 GiB free]
Total: 1 [930,38 GiB] / in use: 1 [930,38 GiB] / in no VG: 0 [0 ]
Expanding logical volume. Before expanding, check the current state of LV (1).
[root@localhost ~]# lvscan
ACTIVE '/dev/vg0/swap' [<16,00 GiB] inherit
ACTIVE '/dev/vg0/root' [<206,41 GiB] inherit
LV /dev/vg0/root
uses 206.41 GB.
We expand LV with the following command (2).
lvextend -l +100%FREE /dev/mapper/vg0-root
We check the performed action (3).
[root@localhost ~]# lvscan
ACTIVE '/dev/vg0/swap' [<16,00 GiB] inherit
ACTIVE '/dev/vg0/root' [<914,39 GiB] inherit
As you can see, after the LV expansion, the amount of disk space occupied became 914.39 GB.
The LV size has increased (4), but the file system is still 204 GB (5).
1. Let's expand the file system.
resize2fs /dev/mapper/vg0-root
We check the size of the file system after the executed command.
[root@localhost ~]# df -h
Файловая система Размер Использовано Дост Использовано% Cмонтировано в
devtmpfs 32G 0 32G 0% /dev
tmpfs 32G 0 32G 0% /dev/shm
tmpfs 32G 9,5M 32G 1% /run
tmpfs 32G 0 32G 0% /sys/fs/cgroup
/dev/mapper/vg0-root 900G 1,3G 860G 1% /
/dev/md126 1007M 120M 837M 13% /boot
tmpfs 6,3G 0 6,3G 0% /run/user/0
The size of the root file system will increase to 900 GB. After the steps have been taken, you can remove the old disks.
Case 2: Replacing smaller disks with larger disks (more than 2TB)
The task: Replace current disks with larger disks (2 x 3TB) while preserving information. In this case, we have 2 x 240 GB SSD (RAID-1) drives with the system installed and 2 x 3 TB SATA drives to which the system needs to be transferred. The current disks use the MBR partition table. Since the new disks are larger than 2 TB, they will need to use the GPT table, since the MBR cannot handle disks larger than 2 TB.
Let's look at the current disk layout.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sda2 8:2 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdb2 8:18 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdc 8:32 0 2,7T 0 disk
sdd 8:48 0 2,7T 0 disk
Check the used partition table on the disk /dev/sda
.
[root@localhost ~]# fdisk -l /dev/sda | grep 'Disk label type'
Disk label type: dos
On disk /dev/sdb
the same partition table is used. Let's check the used disk space in the system.
[root@localhost ~]# df -h
Файловая система Размер Использовано Дост Использовано% Cмонтировано в
devtmpfs 16G 0 16G 0% /dev
tmpfs 16G 0 16G 0% /dev/shm
tmpfs 16G 9,5M 16G 1% /run
tmpfs 16G 0 16G 0% /sys/fs/cgroup
/dev/mapper/vg0-root 204G 1,3G 192G 1% /
/dev/md126 1007M 120M 837M 13% /boot
tmpfs 3,2G 0 3,2G 0% /run/user/0
As you can see, the root of the file system is 204 GB. Let's check the current state of the software RAID array.
1. Install GPT partition table and disk layout
Let's check the partitioning of disks by sectors.
[root@localhost ~]# parted /dev/sda print
Модель: ATA KINGSTON SVP200S (scsi)
Диск /dev/sda: 240GB
Размер сектора (логич./физич.): 512B/512B
Таблица разделов: msdos
Disk Flags:
Номер Начало Конец Размер Тип Файловая система Флаги
1 1049kB 1076MB 1075MB primary загрузочный, raid
2 1076MB 240GB 239GB primary raid
On the new 3TB disk, we will need to create 3 partitions:
- Section
bios_grub
2MiB size for GPT BIOS compatibility, - Partition for the RAID array to be mounted in
/boot
. - A partition for a RAID array that will contain LV root и LV swap.
Installing the utility parted the team yum install -y parted
(for CentOS), apt install -y parted
(for Debian/Ubuntu).
Using parted run the following commands to partition the disk.
Execute the command parted /dev/sdc
and go to the disk layout editing mode.
Create a GPT partition table.
(parted) mktable gpt
Create 1 section bios_grub
section and set a flag for it.
(parted) mkpart primary 1MiB 3MiB
(parted) set 1 bios_grub on
Create partition 2 and set a flag for it. The partition will be used as a block for the RAID array and mounted in /boot
.
(parted) mkpart primary ext2 3MiB 1028MiB
(parted) set 2 boot on
We create a 3rd partition, which will also be used as an array block in which there will be LVM.
(parted) mkpart primary 1028MiB 100%
In this case, it is not necessary to set the flag, but if necessary, it is possible to set it with the following command.
(parted) set 3 raid on
Check the created table.
(parted) p
Модель: ATA TOSHIBA DT01ACA3 (scsi)
Диск /dev/sdc: 3001GB
Размер сектора (логич./физич.): 512B/4096B
Таблица разделов: gpt
Disk Flags:
Номер Начало Конец Размер Файловая система Имя Флаги
1 1049kB 3146kB 2097kB primary bios_grub
2 3146kB 1077MB 1074MB primary загрузочный
3 1077MB 3001GB 3000GB primary
Assign a new random GUID to the disk.
sgdisk -G /dev/sdd
2. Removing partitions of the first disk from the arrays
Check the state of the array
[root@localhost ~]# cat /proc/mdstat
Personalities : [raid1]
md126 : active raid1 sda1[0] sdb1[1]
1047552 blocks super 1.2 [2/2] [UU]
bitmap: 0/1 pages [0KB], 65536KB chunk
md127 : active raid1 sda2[0] sdb2[1]
233206784 blocks super 1.2 [2/2] [UU]
bitmap: 0/2 pages [0KB], 65536KB chunk
unused devices: <none>
The system uses 2 arrays: md126 (mount point /boot) - consists of /dev/sda1
и /dev/sdb1
, md127
(LVM for swap
and the root of the file system) - consists of /dev/sda2
и /dev/sdb2
.
We mark the partitions of the first disk that are used in each array as bad.
mdadm /dev/md126 --fail /dev/sda1
mdadm /dev/md127 --fail /dev/sda2
Delete block device partitions /dev/sda
from arrays.
mdadm /dev/md126 --remove /dev/sda1
mdadm /dev/md127 --remove /dev/sda2
Checking the state of the array after removing the disk.
[root@localhost ~]# cat /proc/mdstat
Personalities : [raid1]
md126 : active raid1 sdb1[1]
1047552 blocks super 1.2 [2/1] [_U]
bitmap: 0/1 pages [0KB], 65536KB chunk
md127 : active raid1 sdb2[1]
233206784 blocks super 1.2 [2/1] [_U]
bitmap: 2/2 pages [8KB], 65536KB chunk
unused devices: <none>
3. Adding new disk partitions to the array
The next step is to add partitions of the new disk to arrays for synchronization. We look at the current state of the disk layout.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
└─sda2 8:2 0 222,5G 0 part
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdb2 8:18 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdc 8:32 0 2,7T 0 disk
├─sdc1 8:33 0 2M 0 part
├─sdc2 8:34 0 1G 0 part
└─sdc3 8:35 0 2,7T 0 part
sdd 8:48 0 2,7T 0 disk
Section /dev/sdc1
is bios_grub
section and does not participate in the creation of arrays. Arrays will only use /dev/sdc2
и /dev/sdc3
. We add these sections to the corresponding arrays.
mdadm /dev/md126 --add /dev/sdc2
mdadm /dev/md127 --add /dev/sdc3
Then we wait for the synchronization of the array.
[root@localhost ~]# cat /proc/mdstat
Personalities : [raid1]
md126 : active raid1 sdc2[2] sdb1[1]
1047552 blocks super 1.2 [2/2] [UU]
bitmap: 0/1 pages [0KB], 65536KB chunk
md127 : active raid1 sdc3[2] sdb2[1]
233206784 blocks super 1.2 [2/1] [_U]
[>....................] recovery = 0.2% (619904/233206784) finish=31.2min speed=123980K/sec
bitmap: 2/2 pages [8KB], 65536KB chunk
unused devices: <none>
Partitioning disks after adding partitions to the array.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
└─sda2 8:2 0 222,5G 0 part
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdb2 8:18 0 222,5G 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdc 8:32 0 2,7T 0 disk
├─sdc1 8:33 0 2M 0 part
├─sdc2 8:34 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdc3 8:35 0 2,7T 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdd 8:48 0 2,7T 0 disk
4. Removing partitions of the second disk from the arrays
We mark the partitions of the second disk that are used in each array as bad.
mdadm /dev/md126 --fail /dev/sdb1
mdadm /dev/md127 --fail /dev/sdb2
Delete block device partitions /dev/sda
from arrays.
mdadm /dev/md126 --remove /dev/sdb1
mdadm /dev/md127 --remove /dev/sdb2
5. Copy the GPT layout table and synchronize the array
To copy the GPT markup table, use the utility sgdisk
, which is included in the package for working with disk partitions and the GPT table - gdisk
.
Installation gdisk
for CentOS:
yum install -y gdisk
Installation gdisk
for Debian/Ubuntu:
apt install -y gdisk
ATTENTION: For GPT first disk is indicated on which copy markup, second disk indicates disk from which copy markup. If you mix up the disks, then the initially good partitioning will be overwritten and destroyed.
Copy the GPT partition table.
sgdisk -R /dev/sdd /dev/sdc
Partitioning disks after transferring a table to disk /dev/sdd
.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
└─sda2 8:2 0 222,5G 0 part
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
└─sdb2 8:18 0 222,5G 0 part
sdc 8:32 0 2,7T 0 disk
├─sdc1 8:33 0 2M 0 part
├─sdc2 8:34 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdc3 8:35 0 2,7T 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdd 8:48 0 2,7T 0 disk
├─sdd1 8:49 0 2M 0 part
├─sdd2 8:50 0 1G 0 part
└─sdd3 8:51 0 2,7T 0 part
Next, we add each of the partitions participating in software RAID arrays.
mdadm /dev/md126 --add /dev/sdd2
mdadm /dev/md127 --add /dev/sdd3
We are waiting for the synchronization of the array.
[root@localhost ~]# cat /proc/mdstat
Personalities : [raid1]
md126 : active raid1 sdd2[3] sdc2[2]
1047552 blocks super 1.2 [2/2] [UU]
bitmap: 1/1 pages [4KB], 65536KB chunk
md127 : active raid1 sdd3[3] sdc3[2]
233206784 blocks super 1.2 [2/1] [U_]
[>....................] recovery = 0.0% (148224/233206784) finish=26.2min speed=148224K/sec
bitmap: 2/2 pages [8KB], 65536KB chunk
unused devices: <none>
After copying the GPT partition to the second new disk, the partition will look like this.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
├─sda1 8:1 0 1G 0 part
└─sda2 8:2 0 222,5G 0 part
sdb 8:16 0 223,6G 0 disk
├─sdb1 8:17 0 1G 0 part
└─sdb2 8:18 0 222,5G 0 part
sdc 8:32 0 2,7T 0 disk
├─sdc1 8:33 0 2M 0 part
├─sdc2 8:34 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdc3 8:35 0 2,7T 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdd 8:48 0 2,7T 0 disk
├─sdd1 8:49 0 2M 0 part
├─sdd2 8:50 0 1G 0 part
│ └─md126 9:126 0 1023M 0 raid1 /boot
└─sdd3 8:51 0 2,7T 0 part
└─md127 9:127 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
Next, install GRUB on the new drives.
Installation for CentOS:
grub2-install /dev/sdX
Installation for Debian/Ubuntu:
grub-install /dev/sdX
where X
- drive letter, in our case drives /dev/sdc
и /dev/sdd
.
Update information about the array.
For CentOS:
mdadm --detail --scan --verbose > /etc/mdadm.conf
For Debian/Ubuntu:
echo "DEVICE partitions" > /etc/mdadm/mdadm.conf
mdadm --detail --scan --verbose | awk '/ARRAY/ {print}' >> /etc/mdadm/mdadm.conf
Updating the image initrd
:
For CentOS:
dracut -f -v --regenerate-all
For Debian/Ubuntu:
update-initramfs -u -k all
Update the GRUB configuration.
For CentOS:
grub2-mkconfig -o /boot/grub2/grub.cfg
For Debian/Ubuntu:
update-grub
After the steps have been taken, the old disks can be removed.
6. Extending the file system (ext4) of the root partition
Partitioning disks to file system extension after system migration to 2 x 3TB disks (RAID-1).
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
sdb 8:16 0 223,6G 0 disk
sdc 8:32 0 2,7T 0 disk
├─sdc1 8:33 0 2M 0 part
├─sdc2 8:34 0 1G 0 part
│ └─md127 9:127 0 1023M 0 raid1 /boot
└─sdc3 8:35 0 2,7T 0 part
└─md126 9:126 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdd 8:48 0 2,7T 0 disk
├─sdd1 8:49 0 2M 0 part
├─sdd2 8:50 0 1G 0 part
│ └─md127 9:127 0 1023M 0 raid1 /boot
└─sdd3 8:51 0 2,7T 0 part
└─md126 9:126 0 222,4G 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
Now sections /dev/sdc3
и /dev/sdd3
occupy 2.7 TB. Since we created a new disk layout with the GPT table, the size of the 3rd partition was immediately set to the maximum possible disk space, in this case it is not required to expand the partition.
It is necessary:
- Expand array md126,
- Expand PV (physical volume),
- Expand LV (logical-volume) vg0-root,
- Expand the file system.
1. Expanding the array md126
up to the maximum.
mdadm --grow /dev/md126 --size=max
After expanding the array md126
the size of the occupied space has increased to 2.7 TB.
[root@localhost ~]# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 223,6G 0 disk
sdb 8:16 0 223,6G 0 disk
sdc 8:32 0 2,7T 0 disk
├─sdc1 8:33 0 2M 0 part
├─sdc2 8:34 0 1G 0 part
│ └─md127 9:127 0 1023M 0 raid1 /boot
└─sdc3 8:35 0 2,7T 0 part
└─md126 9:126 0 2,7T 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
sdd 8:48 0 2,7T 0 disk
├─sdd1 8:49 0 2M 0 part
├─sdd2 8:50 0 1G 0 part
│ └─md127 9:127 0 1023M 0 raid1 /boot
└─sdd3 8:51 0 2,7T 0 part
└─md126 9:126 0 2,7T 0 raid1
├─vg0-root 253:0 0 206,4G 0 lvm /
└─vg0-swap 253:1 0 16G 0 lvm [SWAP]
Expanding physical volume.
Before expanding, we check the current value of the occupied space PV /dev/md126
.
[root@localhost ~]# pvs
PV VG Fmt Attr PSize PFree
/dev/md126 vg0 lvm2 a-- 222,40g 0
We expand PV with the following command.
pvresize /dev/md126
We check the performed action.
[root@localhost ~]# pvs
PV VG Fmt Attr PSize PFree
/dev/md126 vg0 lvm2 a-- <2,73t 2,51t
Expanding logical volume vg0-root.
After expanding PV, check the space occupied by VG.
[root@localhost ~]# vgs
VG #PV #LV #SN Attr VSize VFree
vg0 1 2 0 wz--n- <2,73t 2,51t
Let's check the space occupied by LV.
[root@localhost ~]# lvs
LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert
root vg0 -wi-ao---- <206,41g
swap vg0 -wi-ao---- <16,00g
Volume vg0-root is 206.41 GB.
Expand LV to maximum disk space.
lvextend -l +100%FREE /dev/mapper/vg0-root
Checking the LV space after expansion.
[root@localhost ~]# lvs
LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert
root vg0 -wi-ao---- 2,71t
swap vg0 -wi-ao---- <16,00g
Expanding the file system (ext4).
Let's check the current size of the file system.
[root@localhost ~]# df -h
Файловая система Размер Использовано Дост Использовано% Cмонтировано в
devtmpfs 16G 0 16G 0% /dev
tmpfs 16G 0 16G 0% /dev/shm
tmpfs 16G 9,6M 16G 1% /run
tmpfs 16G 0 16G 0% /sys/fs/cgroup
/dev/mapper/vg0-root 204G 1,4G 192G 1% /
/dev/md127 1007M 141M 816M 15% /boot
tmpfs 3,2G 0 3,2G 0% /run/user/0
The volume /dev/mapper/vg0-root is 204 GB after the LV expansion.
Expanding the file system.
resize2fs /dev/mapper/vg0-root
We check the size of the file system after its expansion.
[root@localhost ~]# df -h
Файловая система Размер Использовано Дост Использовано% Cмонтировано в
devtmpfs 16G 0 16G 0% /dev
tmpfs 16G 0 16G 0% /dev/shm
tmpfs 16G 9,6M 16G 1% /run
tmpfs 16G 0 16G 0% /sys/fs/cgroup
/dev/mapper/vg0-root 2,7T 1,4G 2,6T 1% /
/dev/md127 1007M 141M 816M 15% /boot
tmpfs 3,2G 0 3,2G 0% /run/user/0
The size of the file system has been increased by the entire size of the volume.
Source: habr.com