Implementation of blog 8:raid array

The composition principle of RAID arrays:

1.raid:redundant Arrays of independent Disks independent redundant disk array
1) Improved IO capability and increased durability
2) Level: Multiple disk organizations work together in different ways
3) How the raid is implemented:
External disk array: Provides adaptive capability through an extended adapter
Internal raid (soft RAID): Motherboard integrated RAID controller
4) Level of RAID:
RAID-0: Striped Reel
RAID-1: Mirrored volumes
...
RAID-5: Check code disk
RAID10: First mirror from the bottom, then strip (Enterprise-specific)
RAID01: First strip from bottom, rear mirror
JBOD: A simple technology that connects multiple disks into a single market
2.raid-0:
Read and write performance improvement;
Free space: n*min (S1,s2,...)
No fault-tolerant capability
Minimum number of disks: 2,

RAID-1:
Read performance improvement, write performance slightly decreased;
Free space: 1*min (S1,s2,...)
Have redundancy capability
Minimum number of disks: 2,

RAID-4:
1101, 0110, 1011

RAID-5:
Improved read and write performance
Free space: (N-1) *min (s1,s2,...)
Fault tolerance: 1 disks
Minimum number of disks: 3,

RAID-6:
Improved read and write performance
Free space: (N-2) *min (s1,s2,...)
Fault Tolerance: 2 disks
Minimum number of disks: 4, 4+

Mixed type
RAID-10:
Improved read and write performance
Free space: n*min (S1,s2,...) /2
Fault tolerance: Each group of images can only be broken one piece;
Minimum number of disks: 4, 4+

Jbod:just a Bunch of Disks
Function: The space of multiple disks is combined with a large continuous space;
Free space: sum (s1,s2,...)

The implementation principle of soft raid and the introduction of related commands
1. There is a module md:multidisks in the kernel, which realizes the soft raid
2. This module has its own management tools: Mdadm.
This is a modular tool, first of all it has the syntax format:

mdadm [mode] [Raiddevice] [options] <component-devices>

Supported RAID levels: LINEAR, RAID0, RAID1, RAID4, RAID5, RAID6, RAID10;

Mode:
Create:-C
Assembly:-A
Monitoring:-F
Management:-F,-R,-a

<raiddevice>:/dev/md#
<component-devices>: any block device

-C: Create pattern
-N #: Create this raid with # blocks of devices;
-L #: Indicates the level of RAID to be created;
-A {Yes|no}: Automatically create device files for target RAID devices;
-C Chunk_size: Indicates the block size;
-X #: Indicates the number of free disks;

-D: Displays details of the raid;
Mdadm-d/dev/md#

Management mode:
-F: flag specifies that the disk is damaged;
-A: Adding disks
-R: Remove disk

Observe the status of MD:
Cat/proc/mdstat

To stop the MD device:
Mdadm-s/dev/md#

Watch command:
-N #: Refresh interval, unit is seconds;

watch-n# ' COMMAND '

Soft RAID Practice Operation:

1. First we have to prepare 4 partitions, of which three are used for RAID5 components, the remaining one for the idle RAID disk, to prevent RAID5 in a hard disk in the case of errors, can be replaced in a timely manner.
[Email protected] ~]# FDISK/DEV/SDA
Command (M for help): N
First cylinder (8513-15665, default 8513): 5
Value out of range.
First cylinder (8513-15665, default 8513):
Using Default Value 8513
Last cylinder, +cylinders or +size{k,m,g} (8513-15665, default 15665): +5g

Command (M for help): t
Partition number (1-7): 5
Hex code (type L to list codes): FD
Changed system type of partition 5 to FD (Linux raid AutoDetect)

Command (M for help): P

disk/dev/sda:128.8 GB, 128849018880 bytes
255 heads, Sectors/track, 15665 cylinders
Units = Cylinders of 16065 * 8225280 bytes
Sector size (logical/physical): bytes/512 bytes
I/O size (minimum/optimal): bytes/512 bytes
Disk identifier:0x0008ea5a

Device Boot Start End Blocks Id System
/DEV/SDA1 * 1 204800 Linux
Partition 1 does not end on cylinder boundary.
/dev/sda2 7859 62914560 8e Linux LVM
/dev/sda3 7859 8512 5252256 fd Linux raid AutoDetect
/DEV/SDA4 8513 15665 57456472+ 5 Extended
/dev/sda5 8513 9166 5253223+ fd Linux raid AutoDetect
/dev/sda6 9167 9820 5253223+ fd Linux raid AutoDetect
/dev/sda7 9821 10474 5253223+ fd Linux raid AutoDetect

Prepare the partition we will use with the above command.

[Email protected] ~]# partx-a/DEV/SDA

2. Then we're going to create the RAID5.
[Email protected] ~]# mdadm-c/dev/md0-a yes-n 3-x 1-l 5/dev/sda{3,5,6,7}

View MD devices in the system
[Email protected] ~]# Cat/proc/mdstat
In this process will do the recovery process, the effect is
To do a bitwise alignment of the RAID disks, which makes the raid redundant, capable of bitwise XOR or processing

So our hardware will be processed, and then we'll have to format it.
[Email protected] ~]# mke2fs-t ext4/dev/md0

to mount
[Email protected] ~]# Mkdir/mydata
[Email protected] ~]# Mount/dev/md0/mydata

View information about the creation
[Email protected] ~]# DF-LH
[Email protected] ~]# blkid/dev/md0

Such a complete, able-to-work RAID5 is done.

2. What happens if we artificially destroy one of the partitions?
[Email protected] ~]# mdadm/dev/md0-f/DEV/SDA7

At this point, the raid is automatically synchronized, and the View command is:
[Email protected] ~]# Cat/proc/mdstat
If you want a dynamic query, you can use the Watch command
[[email protected] ~]# watch-n 1 ' cat/proc/mdstat '
When you check this raid again, you will find that the broken hard drive has been replaced by the free disk.
[Email protected] ~]# mdadm-d/dev/md0
Note that at this point the RAID5 can still hold a piece of hard disk damage, at which point a RAID device is used for demotion
[Email protected] ~]# mdadm/dev/md0-r/dev/sda6

3. In the already set raid on the home hard disk, then can only be used as a free disk, to make it a part of the RAID5, need to use grow self-command, but here will not repeat, after all, not very important!

This article is from the "Fante" blog, make sure to keep this source http://8755097.blog.51cto.com/8745097/1690279

Implementation of blog 8:raid array