Thursday 20 December 2018

Understanding how a hardware RAID controller works


I am currently looking at a specific product:


https://www.amazon.com/gp/product/B004JPUZWU/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1


enter image description here


This is a hardware RAID controller and I just want to understand how to use it very well before actually implementing it and risking any data loss. So I will be using a RAID 10 setup. I see that I just put the dial to RAID 10 and press the change mode button to create the initial RAID. Once the RAID is going, if I take a drive out ever and have to replace it it seems I would press this button again and it should rebuild the RAID.


How does the controller know to copy the data over to the new drive as opposed to just building a new RAID setup and eliminating all of my data?


I did look at the documentation for this device but it doesn't appear to answer how this is done.



Answer



Per talking to someone from the company that makes this device, this seems to be how this particular RAID controller works:


Initial setup:



  • Plug in all drives, connections, and power (separate power supply) select RAID mode on dial, press Change Mode button. This simply configures the mode so anything viewing will see the drives as one volume. Drives still need to be formatted to get a file system on the volume.


If a drive fails:




  • The buzzer onboard will beep and red LED on front will be solid indicating there is a drive failure. If leds are hooked up individually to each header for each drive, the specified LED will light up showing which drive has failed.




  • Power down unit and replace bad drive with new one.




  • Turn unit back on. Red light on front will now be blinking, rather than on solid. This indicates that the new drive is being written to and being rebuilt. No pressing of buttons or use of software needed.




Doing offsite backups with this unit:




  • Following the same procedure for if a drive fails, you can unplug a drive and store it in another location. The unit will see this as a missing drive.




  • Power unit off and install a new drive in place. Red light will be blinking indicating that the new drive is being written to, rebuilding the RAID.




I confirmed with the company and there is a small partition that exists on each drive when it is built that is hidden from the rest of the drive that contains the RAID volume. When the device sees a drive missing this section, and the data doesn't match what it is looking for, then it assumes that drive is new.


Different RAID Types with this device


The different RAID types are slightly different from their normal definitions for this device.




RAID 10


enter image description here


According to the company, the drives are not "paired" like in a traditional RAID 10 setup (where each pair would be a RAID 1 and then a RAID 0 layer on top of that). They are all "equal" drives. In this setup you can need 4 drives and 2 can fail at max and still allow data to be safe. It does not matter which 2 in combination. Capacity seen from the resulting RAID = about half of the total capacity of the drives put together.




RAID 5


enter image description here


So on the RAID 5 configurations, data is stored on a minimum of 3 drives (in theory can be an infinite maximum). For this particular unit, it was confirmed that each drive contains the parity and the data is striped across them all. In this configuration, you can lose one drive and still have your data backed up. It does not matter which drive is lost. Capacity = approximately 90% of the total capacity. The parity on a RAID 5 = about 10% of the total drive capacity.




RAID 3


enter image description here


Raid 3 stripes its data and has parity data like RAID 5. Instead of the parity scattered among the drives, it has one drive that contains the parity. This makes it a slower option in general than a RAID 5 or RAID 10 but still is an option for this unit. Raid 3 needs a minimum of 3 drives as well and will allow one drive failure with keeping data safe. Capacity on a RAID 3 is the total drive capacity of all of the drives minus the one being used for parity.




"Clone"


Clone on this device is the same thing as a RAID1


enter image description here


In a RAID 1 setup, every drive connected is a copy of the other. Since this unit supports 5 drives, you would have 5 times the redundancy. Your RAID storage space would = 1/5 of the total capacity. You could lose a maximum of 4 drives and still have your data backed up safely.




RAID 0


enter image description here


RAID 0 only does striping. This will improve performance as data read / writes are spread across multiple drives but no data is backed up. Minimum drives = 2. Maximum = unlimited. Data capacity = total drive capacity.




Large


On this device the "Large" mode will make all of the drives just show up by themselves as they are. In this mode the device just acts like a HUB and does no RAID at all.


No comments:

Post a Comment

Where does Skype save my contact's avatars in Linux?

I'm using Skype on Linux. Where can I find images cached by skype of my contact's avatars? Answer I wanted to get those Skype avat...