Explanation on CentOS 6.5 mounting over 16T of large storage space


Recently, a large-capacity storage machine has been configured for database data backup.

Type: 2U Storage Server

CPU: 2 Intel INTEL G6950 2.8G/dual core/L3 3M/32NM;

System memory: 96GB DDR3 ECC 800/1066/1333 MHz, supports 3-channel communication mode, provides 6 DIMMs

System hard disk: 300G SSD (raid1 does not support hot swap);

Storage hard disk: 16*2000G enterprise version SATAII 7200 RPM, 3.5″hard disk; up to 16 3.5 inch hard disk hot plug installation bits; disk array: Adaptec 16 channel array control module (WZ51645), 512M Cache, supporting RAID 0, 1, 1E, 5, 5EE, 6, 10, 50, 60, JBOD.

Network: 2 10/100/1000M adaptive Ethernet cards;

Power supply: 760W 2+1 server redundant power supply;

The system with CentOS 6.5 has eight 3T storage disks and RAID5 (raid5 + 1) system has an effective capacity space of about 22T.

Prepare to start partitioning and formatting mounts for mass storage 

Because MBR partitions support a maximum capacity of only 2T, if more than 2T, use GPT partitions.

First use the parted command to convert the hard disk into a GPT partition

If I need to install without the parted command, I can install it through the yum-y install parted command

[[email protected] ~]$ parted /dev/sdb 

GNU Parted 2.1 

Using /dev/sdb 

Welcome to GNU Parted! Type ‘help’ to view a list of commands.


From MBR to GPT Disk

(parted) mklabel gpt 

Warning: The existing disk label on /dev/sdb will be destroyed and all data on this disk will be lost. Do you want to continue? 

Yes/No? Yes 


Set the type and size of the partition you want to do (according to the actual situation) 0, -1 represents the starting capacity and the end capacity, the difference is the partition capacity, the unit is M mkpart primary 0-1 represents the partition of the total capacity.

(parted) mkpart primary 0 -1 

Warning: The resulting partition is not properly aligned for best performance. 

Ignore/Cancel? Ignore 



(parted) p 

Model: Adaptec RAID 6 (scsi) 

Disk /dev/sdb: 28.0TB 

Sector size (logical/physical): 512B/512B 

Partition Table: gpt

Number Start End Size File system Name Flags 

1 17.4kB 28.0TB 28.0TB primary 

(parted) quit 

Information: You may need to update /etc/fstab.

After partitioning, format the hard disk and mount it 

Centos6 already supports ext4 by default, so use ext4 filesystem

[[email protected] ~]$ mkfs.ext4 /dev/sdb1 

mke2fs 1.41.12 (17-May-2010) 

mkfs.ext4: Size of device /dev/sdb1 too big to be expressed in 32 bits 

using a blocksize of 4096.

Attempt to change blocksize to 4096 according to the error notification. The error is still reported.

Looking at the relevant information, it seems that ext4 does not really support single partition space above 16TB at present.

I decided to switch to XFS instead of entanglement.

Start installing XFS 

[[email protected] ~]$ yum install kmod-xfs xfsprogs 

[root @ ~]$modprobe XFS // load XFS file system module

[root @ ~]$lsmod | grep XFS // / Check if the XFS module is loaded

Format and mount with XFS 

[[email protected] ~]$ mkfs.xfs /dev/sdb1 

mount /dev/sdb1 /data/

Add to fstab to enable the system to start automatic mounting 

[root @ ~]$VI / etc / fstab // open fstab

UUID = c3749d2d-08b5-45a3-9aa4-312b61d471/data XFS defaults 12// Add mount information to specify file system XFS (uuid is obtained by blkid/dev/sdb1 command)

So far, great success has been achieved! 

[[email protected] ~]$ df -lTh 

Filesystem Type Size Used Avail Use% Mounted on 

/dev/sdb2 ext4 30G 5.2G 23G 19% / 

tmpfs tmpfs 1.8G 0 1.8G 0% /dev/shm 

/dev/sdb1 ext4 124M 31M 88M 26% /boot 

/dev/sda1 xfs 26T 33M 26T 1% /data