1 00:00:00,000 --> 00:00:11,780 In this video, you'll learn about RAID devices. RAID configurations are pretty common on enterprise 2 00:00:11,780 --> 00:00:18,940 hardware. The goal of RAID is to offer an array of independent disks. That's also where 3 00:00:18,940 --> 00:00:26,219 it stands for Redundant Array of Independent Disks. And that is for increased redundancy. 4 00:00:26,219 --> 00:00:31,040 If one of the disks fails, you still have storage left to take care of it. There are 5 00:00:31,040 --> 00:00:36,880 different types of RAID. RAID 0 is striping. Striping means that you create one big disk 6 00:00:36,880 --> 00:00:42,599 device out of multiple disk devices. In striping, you don't take care of redundancy. So if one 7 00:00:42,599 --> 00:00:49,279 of the disks breaks, you're in trouble. RAID 1 is mirroring. So if one of the disks breaks, 8 00:00:49,279 --> 00:00:56,200 then the other one takes over. RAID 5 is striping with distributed parity. And RAID 6 is striping 9 00:00:56,660 --> 00:01:01,939 with dual distributed parity. RAID 5 and RAID 6 both are using different disk devices in 10 00:01:01,939 --> 00:01:06,540 such a way that if one of the disk devices breaks, you still can continue using your 11 00:01:06,540 --> 00:01:12,980 storage. Then there is RAID 10, which is a combination between mirrored and striped. 12 00:01:12,980 --> 00:01:18,220 RAID devices can be purchased hardware-wise. You can also create software RAID devices. 13 00:01:18,220 --> 00:01:23,180 That's what we are going to do here. If you want to, you can use either an entire disk 14 00:01:23,199 --> 00:01:28,760 or you can put a partition in your RAID device. If you put a partition inside, then make sure 15 00:01:28,760 --> 00:01:36,000 that you are setting the RAID device type. Next, you would use the mdadm command. MD 16 00:01:36,000 --> 00:01:41,839 stands for multiple device and ADM, obviously, for administrator. You define the device that 17 00:01:41,839 --> 00:01:46,919 you want to create, as well as the level, as well as the number of disks. Next, you 18 00:01:46,919 --> 00:01:52,919 put a file system on it. You make the configuration persistent and you mount it. And finally, 19 00:01:52,919 --> 00:01:58,160 you put everything in etcfsstep. And you can monitor the status of your RAID devices 20 00:01:58,160 --> 00:02:06,440 through slashproc slash mdstat or using mdadm. Let me show you. As you can see in LSBOK, 21 00:02:06,440 --> 00:02:12,720 I have added two more disks, NVMe 0, N4, and N5. And that is what I'm going to use in my 22 00:02:12,720 --> 00:02:20,839 RAID configuration. So do we have mdadm? Yeah, we do have mdadm. And I am going to use it 23 00:02:20,839 --> 00:02:28,440 to create the device md0. The level is one. RAID level one is a simple RAID mirroring. 24 00:02:28,440 --> 00:02:36,800 So one of the disks is used as the backup of the other disk. RAID disks is two. And 25 00:02:36,800 --> 00:02:47,600 then I need to mention my devices. So dev NVMe 0, N4, and dev NVMe 0, N5. So it is telling 26 00:02:47,600 --> 00:02:53,240 us that this RAID device might not be suitable for booting. We don't talk about booting. 27 00:02:53,240 --> 00:02:59,320 So that is OK. And I'm continuing to create the array. And you can see the device is already 28 00:02:59,320 --> 00:03:07,320 started. So next, you put a file system on it. And then you create the configuration 29 00:03:07,320 --> 00:03:18,679 mdadm, minus minus detail, minus minus scan. And you write that to the file etc mdadm.conf. 30 00:03:18,679 --> 00:03:23,399 That will make sure that your configuration is persistent. And you can reinitialize your 31 00:03:23,399 --> 00:03:29,160 RAID device. Then we need a mount point, so the RAID directory. And I'm going to mount 32 00:03:29,160 --> 00:03:38,639 dev md0 on slash RAID. And now if I copy stuff to this RAID device, it will be stored 33 00:03:38,639 --> 00:03:50,199 there. So if you want to make it persistent, you need to put it in etc fsstep. And in etc 34 00:03:50,199 --> 00:03:57,720 fsstep, you are going to define your device. This time, dev md0 is OK, because the device 35 00:03:57,720 --> 00:04:03,320 will be detected based on the metadata information that has been written to the hard disk. So 36 00:04:03,320 --> 00:04:11,080 I'm mounting my dev md0 on slash RAID. The file system is just ext4. The mount options 37 00:04:11,080 --> 00:04:17,679 are default and 0 and 0. And that will make sure that it is persistent. You want to know 38 00:04:17,679 --> 00:04:24,839 what is going on? Well, have a look at the contents of proc mdstat, which is a proc file 39 00:04:24,839 --> 00:04:30,440 that is giving synchronization information about your RAID device. Or alternatively, 40 00:04:30,440 --> 00:04:36,600 use mdadm minus minus detail on your RAID device. Like this, this is giving a little 41 00:04:36,600 --> 00:04:42,799 bit more detail about the current state. Now, when you are working with RAID devices, occasionally 42 00:04:42,799 --> 00:04:49,200 you will have to replace a failing device. In order to do so efficiently, it's a smart 43 00:04:49,200 --> 00:04:55,959 idea to have an unused RAID member as a hot spare. So when you are creating, you can create 44 00:04:55,959 --> 00:05:02,640 this way with minus minus spare devices is one. And then you can fail a device. You can 45 00:05:02,640 --> 00:05:08,239 do it manually to test if it is working. And after failure, you remove the device, and 46 00:05:08,239 --> 00:05:13,239 then you add the device that was marked as a spare device previously. And that's how 47 00:05:13,239 --> 00:05:15,040 you replace failing devices.