Disk Management is important for an OPS engineer. How to manage the data, you can not do without raid.
Content:
1. Common Types of RAID:
RAID0 (Strip)
650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/4C/2D/wKioL1Q3uY-j5MMgAACbfrJJtok640.jpg "title=" 1.PNG " alt= "Wkiol1q3uy-j5mmgaacbfrjjtok640.jpg"/>
RAID1 (Mirror)
650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M01/4C/2D/wKioL1Q3ufyiOkzaAACQkt6mi1k438.jpg "title=" 2.PNG " alt= "Wkiol1q3ufyiokzaaacqkt6mi1k438.jpg"/>
RAID5 (Parity check)
650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M00/4C/2D/wKioL1Q3ujXATegRAACwzkxBu1E313.jpg "title=" Capture 5. PNG "alt=" Wkiol1q3ujxategraacwzkxbu1e313.jpg "/>
-
Raid1 + Raid0 (first 1 after 0) &NB Sp , &NB Sp
RAID0+RAID1 (first 0 after 1)
650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/4C/2D/wKioL1Q3us7zDgV7AADkgcpAuw0861.jpg "title=" Capture 01. PNG "alt=" Wkiol1q3us7zdgv7aadkgcpauw0861.jpg "/>
Raid5+raid0
RAID0+RAID5 Ibid.
Jbod is also commonly used, which can be expanded by increasing the number of hard drives.
Summarize:
Raid0
|
Performance improvement: Read, write Redundancy capability: None Space utilization: NS At least 2 pieces of disk |
Raid1
|
Performance gains: Read performance gains, write performance drops Redundancy capability: Available Space utilization: 1/2 At least 2 pieces of disk |
Raid5
|
Performance improvement: Read, write promotion Redundancy capability: Available Space utilization: (n-1)/n At least 3 plates required |
Raid1+raid0
|
Performance improvement: Read, write promotion Redundancy capability: Available Space utilization: 1/2 At least 4 plates required |
Raid0+raid1
|
Performance improvement: Read, write promotion Redundancy capability: Available Space utilization: 1/2 At least 4 plates required |
Raid5+raid0
|
Performance improvement: Read, write promotion Redundancy capability: Available Space utilization: (n-2)/n At least 6 plates required |
Jbod |
Performance improvement: None Redundancy capability: None Space utilization: 100% Need at least 2 bucks. |
2. Implement raid
RAID can be done through software simulations, where there is an environment for hardware RAID.
Logical raid:/dev/dm# #代表着编号
Some soft raid software uses:
Md:raid Module
Modal Commands:
Create a pattern
-C
Dedicated options:
-L: Level
-N #: Number of devices
-a {Yes|no}: whether to automatically create a device file for it
-c:chunk Size (chunk size), 2^n, default is 64K
-X #: Specify the number of free disks (after a disk is broken, it can be on top)
Management Mode
--add,--remove,--fail
mdadm/dev/md#--fail/dev/sda7
Monitoring mode
-F
Growth Model
-G
Assembly Mode
-A
View details of a RAID array
Mdadm-d/dev/md#
--detail
Stop array:
mdadm-s/dev/md#
--stop
Practice:
Create a RAID5 device with a space size of 10G and a chuck size of 32k, requiring the device to automatically mount to the/backup directory when it is powered on;
Create a 2G size RAID0 can be 4 x 512MB or 2 1G
Create a 2G size RAID1 can be 2 x 2G
Create a 2G size RAID5 can be 3 x 1G
-N #: Specifies the period length in seconds and defaults to 2
Format: Watch-n # ' COMMAND ' (double quotes also line)
Example: Watch "Cat/proc/mdstat"
mdamd-d--scan >/etc/mdadm.conf
Mdamd-a/dev/md# can be assembled directly.
Attention:
If a chunk is 64k a disk block is 4K (4096) then this means: a chunk is 16 disk blocks.
The MKE2FS has the-e stride function to adjust the chunk by a multiple of the block size, which optimizes the soft raid
3.rm-f/dev/md# should be mdadm-s/dev/md# (stop soft raid) before.
3 Specific processes: (Take Raid0 as an example)
1. Create/dev/sda{5,6,7}, and the type is RAID type
2.PARTPORBE/DEV/SDA (update partition information)
3.cat/proc/partitions (view partition)
4.mdadm-c/dev/md0-n 2-l 0-x 1-a yes/dev/sda{5,6,7}
5.mdadm-d/dev/md0 (can see the running situation)
This article from "Technology first" blog, declined reprint!
Linux Learning Note 10 (RAID Basics)