RAID disk redundancy queue

RAID brief introduction

As servers are more and more widely used in enterprises , Data security , data storage , Data recovery , Storage speed and other aspects are becoming more and more important for enterprise users , and RAID Cards play a vital role in this process ,RAID The performance and characteristics of the card itself are often ignored by some users , In order to save costs, server manufacturers will also match some relatively elementary or even small brands RAID Card products , When there are problems in the use process, users just focus on the hard disk

RAID（ Independent redundant disk array ） It's a kind of multi independent hard disk （ Physical hard disk ） Combine them in different ways to form a hard disk group （ Logical hard disk ）, So as to provide higher storage performance and data backup capability than a single hard disk .

  soft RAID And hard RAID
We want to implement on the server RAID, You can use disk array cards （RAID card ） To form a RAID, That is, hard RAID.RAID There is a special chip on the card for RAID Mission , So the performance is much better , And does not occupy the system performance , The disadvantage is that RAID Cards are expensive
If we don't want to spend money and want to use RAID, Then only use soft RAID 了 . soft RAID It refers to the realization through software RAID function , There is no extra cost , But it costs more server system performance , And the writing speed of data is faster than that of hard RAID slow

RAID It features N Block hard disk simultaneously accelerates reading speed and provides fault tolerance . Redundant disk array technology was born in 1987 year , Proposed by the University of California, Berkeley . According to different combinations of disk display , Can be RAID Divided into different levels . The level does not represent the level of Technology , Choose which one RAIDlevel Our products depend solely on the user's operating environment and Application , It has nothing to do with the level

 RAID advantage ：
1、 Improve I/O Ability , Disk parallel read and write
2、 Improve durability , Disk redundancy
3、 Level ： Multiple disks work in different ways
4、raid How to achieve ：
External disk array ： Provide adaptability through expansion card
Built in raid： motherboard integrated raid controller , In the installation OS Before BIOS It's equipped with
Software raid： adopt OS To achieve , Hardware is used in production raid To achieve

 raid The shortcomings of ：
raid In the later stage of the array group, new members cannot be added to expand the overall capacity . for example ： use 2 Block hard disk creation raid 0, Later cannot be in raid 0 Add a new hard disk . And that's where LVM Logical volume management function

RAID It can be divided into levels 0 To level 6, Often referred to as ：RAID0、RAID1、RAID2、RAID3、RAID4、RAID5、RAID6
More classic RAID0、RAID1、RAID5、RAID6、RAID10 （RAID0+RAID1）

Grade	name	Number of hard disks	Total capacity	The number of bad sets allowed ( Maximum )	function
RAID0	Banding	>=2 block	2n+n	0 block ( redundancy )	High read-write performance , Capacity combination
RAID1	Mirror image	=2 block	n	1 block ( redundancy )	Synchronous backup , Data security
RAID5	Striping of distributed parity	>=3 block	3n-n	1 block ( redundancy )	High read-write performance
RAID6	Striping of two groups of distributed parity	>=4 block	4n-2n	2 block ( redundancy )	High read-write performance , Data security
RAID10	Mirror image + Banding	4+2n block	Total capacity 50%	The bad half ( redundancy )	Quantity combination + Synchronous backup

mdadm command

linux Software in the system raid Management tool software ：mdadm

Parameters	explain
-D	Show RAID Device details
-A	Add a previously defined RAID
-B	Create one without super blocks RAID equipment
-F	Option monitoring mode
-G	change RAID Size or shape of equipment
-I	Add device to RAID in , Or from RAID Delete device from
-z	To form a RAID1、RAID4、RAID5、RAID6 After each RAID Space capacity acquired by members
-s	Scan configuration files or /proc/mdstat Search for lost information
-C	establish RAID equipment , hold RAID Information is written to each RAID Member super block
-v	Show RAID Details of the creation process
-B	establish RAID Another way , Not put RAID Information is written to each RAID In the super block of members
-l	Appoint RAID The level of
-n	Appoint RAID Number of active devices in
-f	hold RAID Members are listed as having problems , To remove the member
-r	hold RAID Members move out RAID equipment
-a	towards RAID Add a member to the device
--re-add	Put the recently removed RAID Members are added back to RAID In the device
-E	see RAID Member details
-c	Appoint chunk size , Create a RAID It defaults to 512kb
-R	Start partial assembly RAID
-S	Discontinue use RAID equipment , Release all resources
-x	Specify initial RAID Number of spare members of the device
--zero-superblock	If RAID The device contains a valid superblock , This block uses zero coverage

Practice preparation

take linux Virtual machine shutdown , Add 5 block SCSI Interface 2GB The hard disk of , Create a snapshot “5 block 2G Hard disk ”, Then boot the virtual machine , perform lsblk The command displays a list of disk block devices

Raid practice ： The request will linux Of system virtual machine 5 block 2G Hard disk for raid 5 Conduct management , Before the 4 The hard disk is organized into a file named md5 Of Raid5, among raid 5 The active disk of is the first 3 Block hard disk , The first 4 Hard disk as hot standby ( Spare tire 、 As a substitute ) disk .

lsblk                                             # Display the list of disk devices 
mdadm -Dsv                                        # Show raid The equipment list 
mdadm -C md5 -l 5 -n 3 -x 1 /dev/sd{b..e}         # Create name as md5 Of raid 5 equipment 
mdadm -D /dev/md/md5                              # The display name is md5 Of raid 5 Device details 
mdadm -Dsv                                        # Show raid The equipment list 
mdadm -Dsv > /etc/mdadm.conf                      # Generate raid The configuration file 



——————————————————————————————————————
lsblk                                             # Display the list of disk devices 
mdadm -S /dev/md/md5                              # Discontinue use md5 This raid equipment 
mdadm -A /dev/md/md5                              # Enable md5 This raid equipment 


 Be careful ： To deactivate raid Enable the device later , Must ensure mdadm.conf The configuration file exists 

practice ： simulation md5 This Raid5 In the equipment sdc The hard disk is damaged , And then from md5 Remove from the array sdc This member , However, check md5 Device details . take sdc Remove the hard disk from the server , Install a new hard disk to Sdc On the slot of the hard disk . then sdf Hard disk added to md5 This raid 5 In the device group

mdadm -D /dev/md/md5                     # The display name is md5 Of raid 5 Device details 
mdadm /dev/md/md5 -f /dev/sdc            # Appoint sdc The hard disk is bad (fail)
mdadm -D /dev/md/md5                     # The display name is md5 Of raid 5 Device details 
mdadm /dev/md/md5 -r /dev/sdc            # take sdc Hard disk from md5 This raid Remove from group (remove)
mdadm /dev/md/md5 -a /dev/sdf            # take sdf Hard disk add (add) To md5 This raid In the group 
mdadm -D /dev/md/md5                     # The display name is md5 Of raid 5 Device details