RAID implementation under Linux

First, the purpose of the experiment

1. Mastering the implementation method of soft raid in Linux system;

2. Master the configuration process of RAID5;

3. Be familiar with RAID, 5 characteristics through experiment.

Second, the contents and steps of the experiment

1. Create a Linux in VMware. 2. Add the Linux 4 virtual disk (select SCSI type, 2G). 3, the use of Madam in Linux to create RAID5, three disk to do RAID5, a disk to do spare. 4. Format and mount the RAID5. 5. Create some files and folders in RAID5 for fault detection. 6, modify the configuration file, let RAID5 boot automatically loaded. 7, shut down the system.

8. Delete a hard drive (a second hard drive that can be added after deletion).

9. Restart the system to see if the data in the RAID5 volume is missing.

10, according to the experimental results to summarize the contents of the experiment.

Third, the experimental requirements

1. The experimental results are carefully observed, recorded and compared, and the reasons should be found if the inconsistency is inconsistent.

2. All the places that need to be named in the experiment are named by their own names, and can be distinguished by different suffixes. Two servers, such as Zhang San, can be named: Zhangsans1,zhangsans2.

Experimental steps and experimental process:

1. Install red Hat Enterprise Linux 6 in VMware first.

2. Add 4 virtual disks (SCSI,2G) to red Hat Enterpriselinux 6, three blocks for RAID5, one for spare disks.

3. Start Red Hat enterpriselinux 6.

4, to configure the RAID5.

5. Close Red Hat enterpriselinux 6.

6. Delete the added disk (you can delete the second disk you added).

7. Restart Red hatenterprise Linux 6.

8, to test the RAID5 after a broken disk, you can find that the data will still be able to access the normal.

Number of disks required: Three blocks or more we add six hard drives here, four of them do disk arrays, 1 are prepared (spare) disks, and one is reserved for standby Note: The RAID5 disk usage is N-1, which means that four 100G of hard disk free space is 300G.

Experimental steps:

1 Check the disk devices in the system fdisk–l

2 Next Create raid the command used is mdadm, if not please install the MDADM package first

Rhel5 in the CD!

Mdadm--create--auto=yes/dev/md0--level=5--raid-devices=4--spare-devices=1/dev/sd[b-f]

Parameter explanation:

--create//indicates to create raid

--AUTO=YES/DEV/MD0//Newly established software disk display device for MD0,MD serial number can be 0-9

--LEVEL=5//disk array level, here is REID5

--raid-devices//Add the number of blocks as a prestaged (spare) disk

/DEV/SD[B-F]//disk array used by the device, also can be written as "/DEV/SDB/DEV/SDD/DEV/SDE/DEV/SDF" can also be abbreviated as: Mdadm–c/dev/md0–l5–n4–x1/de V/SD[B-F]

There are two ways to see if the raid was successful and whether the creation is working properly

View more information: Mdadm--detail/dev/md0 command to view raid details

More simple view: You can view the/proc/mdstat files directly to see the raid running situation

Cat/proc/mdstat

Format and mount and use raid created

MKFS.EXE3/DEV/MD0//Format RAID5 disk as Mkfs.exe

MIKDIR/MNT/RAID5//Create/raid5 folder under/mnt to mount the md0

MOUNT/DEV/MD0/MNT/RAID5//Mount the md0 to RAID5

Check the Mount status

Df–ht

Try a new raid is not available, write files to it

Set boot auto-start raid and auto Mount

Let raid boot up, riad configuration file name is mdadm.conf, this file

It does not exist, it should be established by itself. The primary role of this configuration file is when the system starts to

Automatic loading of soft raid, but also easy to manage later. Note, the mdadm.conf file is mainly made up of

The next section consists of:

The Devices option is formulated to compose the raid all devices, and the array option specifies the

Device name, RAID level, number of active devices in the array, and the device's UUID number.

Auto-start raid

First set up/etc/mdadm.conf this file

Mdadm--detail--scan >/etc/mdadm.conf

To make a change to this file: vi/etc/mdadm.conf

When setting up auto mount for raid

Modify File Vi/etc/fstab

Add a line/dev/md0/mnt/raid5 ext3 defaults 0 0

Simulate disk corruption in RAID5, verify the functionality of the spare disk (a disk is allowed in RAID5

Damage, that is, we set the spare disk to replace the broken disk immediately to raid the heavy

Security of the data):

MDADM–MANAGE/DEV/MD0–FAIL/DEV/SDD//Use this command to set the status of SDD as an error

Check out the disk information: mdadm–-detail/dev/md0

Simply look at the raid scenario:

The completion of the creation means that RAID5 has been automatically restored.

Check to see if RAID5 is working properly!

Write the data inside.

Remove the faulted disk and add a new disk to the backup spare disk

First delete the damaged disk SDD

MDADM–MANAGE/DEV/MD0--REMOVE/DEV/SDD//will be broken disk SDD from raid in addition to adding a new piece as a spare disk:

MDADM–MANAGE/DEV/MD0--ADD/DEV/SDG//Add a new disk for SDG

RAID5 expansion: Using Grow mode

Mdadm–manage/dev/md0--ADD/DEV/SDG Add a new hard drive

True members of active SYNC/DEV/SDB1 # raid

SPARE/DEV/SDC1 #raid备用成员

Mdadm–g/dev/md0–n "x" #-g is grow Mode "x" is the number of RAID true members

Command parameters

-A =--assemble active

-S =--stop stop

-D =--detail View raid details

-C =--Create a RAID device

-V =--verbose Show details of the build process

-L =--level RAID level

-n = number of--raid-devices RAID devices

-s =--scan Scan raid device

-F =--fail marked bad hard drive

-A =--add add hard drive

-r =-remove removing bad hard drive

Now look at the disk information situation: MDADM–-DETAIL/DEV/MD0

How to turn off raid:

> Uninstall the/dev/md0 directly and comment out the configuration in the/etc/fstab file

UMOUNT/DEV/MD0//Lift Hook up

Vi/etc/fstab//Fstab The inside of the boot automatically mount comments out

#/dev/md0/mnt/raid5 exit3 Defaults 0 0

Experimental results:

The following points were accomplished through experiments:

1. Mastering the implementation method of soft raid in Linux system;

2. Master the configuration process of RAID5; 3. Be familiar with RAID, 5 characteristics through experiment. Create a Linux in VMware. Add the Linux 4 virtual disk (select SCSI type, 2G). Use Madam in Linux to create RAID5, three disks to do RAID5, one disk to make the spare. Format and mount the RAID5. Create some files and folders in RAID5 for fault detection. Modify the configuration file, let RAID5 boot automatically load. Shut down the system. Delete a hard drive (a second hard drive that you can delete and add). Restart the system to see if the data in the RAID5 volume is missing. According to the experimental results, the contents of the experiment were summarized.

Experiment Summary:

Learned some basic operations and understood the basic principles of RAID implementations:

The steps to configure RAID1 are relatively RAID5, but when using mdadm, it is important to note that you do not partition multiple partitions on a hard disk, and then make multiple partitions into an array, which not only does not increase the speed of access to the hard drive, but reduces the overall system performance. The correct approach is to divide a hard disk into one or more partitions, and then make the partitions of several different hard disks into an array. In addition, the system directory such as/usr should not be placed in the array, because the system will not function properly once the array problem occurs.

Set boot auto-start raid and auto Mount

Let raid boot up, riad configuration file name is mdadm.conf, this file is not exist by default, to establish itself. The main function of this configuration file is to automatically load soft raid when the system starts up, and to manage it later.

The mdadm.conf file consists mainly of the following parts:

The Devices option is formulated to compose the raid all devices, and the array option specifies the

Device name, RAID level, number of active devices in the array, and the device's UUID number.

Auto-start raid

First set up/etc/mdadm.conf this file

Mdadm--detail--scan >/etc/mdadm.conf

To make a change to this file: vi/etc/mdadm.conf

Data in today's enterprise occupies an important position, the security of data storage is one of the important issues that people use computers to pay attention to. Often people use a variety of redundant disk array RAID technology on the server side to protect data, high-end servers generally provide expensive hardware RAID controller, but many small and medium-sized enterprises do not have sufficient funds to withstand the cost. Do we have a way to implement raid through software? In fact, under Linux can be software to implement the RAID function of the hardware, so as to save investment, but also achieve good results.

command to start and stop the RAID1 array. Start the direct execution of "mdadm-as/dev/md0". Executing mdadm-s/dev/md0 will stop the RAID1 array. In addition, after adding command mdadm-as/dev/md0 to the Rc.sysinit startup script file, the array will be set to start with the system boot.

Copyright NOTICE: This article for Bo Master original article, without Bo Master permission not reproduced.

RAID implementation under Linux