1 00:00:06,510 --> 00:00:07,590 - [Instructor] In this segment, 2 00:00:07,590 --> 00:00:10,800 we're going to take another look at rate configurations, 3 00:00:10,800 --> 00:00:14,313 specifically disk mirroring and disk striping. 4 00:00:16,364 --> 00:00:18,720 The raid combines multiple disk drive components 5 00:00:18,720 --> 00:00:20,760 into one or more logical units 6 00:00:20,760 --> 00:00:24,150 for the purposes of fault tolerance, data redundancy, 7 00:00:24,150 --> 00:00:26,430 and/or performance improvement. 8 00:00:26,430 --> 00:00:31,430 The raid can be configured for mirroring, striping, or both. 9 00:00:32,130 --> 00:00:35,070 Disk mirroring is a process of writing data 10 00:00:35,070 --> 00:00:37,950 on two partitions on separate disks, 11 00:00:37,950 --> 00:00:41,040 that's referred to as Raid 1. 12 00:00:41,040 --> 00:00:43,020 And you can see from our illustration, 13 00:00:43,020 --> 00:00:45,240 we have two identical disks. 14 00:00:45,240 --> 00:00:48,690 We have Disk 0 and we have Disk 1. 15 00:00:48,690 --> 00:00:51,360 And everything that's being written to Disk 0 16 00:00:51,360 --> 00:00:54,150 is also being written to Disk 1. 17 00:00:54,150 --> 00:00:56,370 So we have complete redundancy. 18 00:00:56,370 --> 00:00:58,410 We can lose either drive 19 00:00:58,410 --> 00:01:01,860 and continue to be functional and operational. 20 00:01:01,860 --> 00:01:04,890 Matter of fact, our user will probably not even notice 21 00:01:04,890 --> 00:01:06,660 that a drive is gone. 22 00:01:06,660 --> 00:01:10,710 So that's referred to as raid level 1 disk mirroring. 23 00:01:10,710 --> 00:01:12,450 But we still have a single point of failure 24 00:01:12,450 --> 00:01:14,160 with traditional disk mirroring 25 00:01:14,160 --> 00:01:16,500 because we only have one drive controller. 26 00:01:16,500 --> 00:01:19,050 One drive controller, two drives, 27 00:01:19,050 --> 00:01:22,000 our single point of failure is going to be that controller. 28 00:01:22,980 --> 00:01:26,400 So best practices would dictate that we add a controller 29 00:01:26,400 --> 00:01:28,230 to mitigate the risk associated 30 00:01:28,230 --> 00:01:31,590 with a single point of failure or an SPOF. 31 00:01:31,590 --> 00:01:33,870 So in the configuration where each drive 32 00:01:33,870 --> 00:01:38,870 has its own controller, we refer to that as disk duplexing. 33 00:01:39,000 --> 00:01:42,000 So we went with disk mirroring, two drives, one controller, 34 00:01:42,000 --> 00:01:44,490 to disk duplexing, two drives, 35 00:01:44,490 --> 00:01:46,140 each having their own controller. 36 00:01:47,250 --> 00:01:50,490 Now let's take a look at disk striping in Raid 5. 37 00:01:50,490 --> 00:01:53,490 Disk striping is a process dividing data into blocks 38 00:01:53,490 --> 00:01:54,900 and spreading the data blocks 39 00:01:54,900 --> 00:01:57,240 across multiple storage devices. 40 00:01:57,240 --> 00:02:00,060 Disk striping is a performance enhancement. 41 00:02:00,060 --> 00:02:02,160 So where do we get the redundancy? 42 00:02:02,160 --> 00:02:04,920 We add in parity bits. 43 00:02:04,920 --> 00:02:07,530 So what we're doing is we're writing a parity bit 44 00:02:07,530 --> 00:02:08,820 with each stripe, 45 00:02:08,820 --> 00:02:12,240 and that parity bit can be used to recreate the data 46 00:02:12,240 --> 00:02:16,500 if any one drive is lost or fails. 47 00:02:16,500 --> 00:02:18,900 So let's look at our Raid 5 configuration. 48 00:02:18,900 --> 00:02:21,330 We've got four drives here. 49 00:02:21,330 --> 00:02:22,710 Looking at four drives. 50 00:02:22,710 --> 00:02:24,540 And watch our striping across. 51 00:02:24,540 --> 00:02:27,480 Stripe A1, A2, A3. 52 00:02:27,480 --> 00:02:29,220 Oh, there's our parity bit. 53 00:02:29,220 --> 00:02:34,220 Our next stripe, B1, B2, parity bit, then B3. 54 00:02:34,830 --> 00:02:38,970 C1, parity bit, C2, C3. 55 00:02:38,970 --> 00:02:42,840 And then parity bit, D1, D2, D3. 56 00:02:42,840 --> 00:02:45,600 We can lose any one of these drives, 57 00:02:45,600 --> 00:02:48,330 and the data on that failed or lost drive 58 00:02:48,330 --> 00:02:53,330 will be recreated from each of the remaining parity bits. 59 00:02:55,380 --> 00:02:57,180 So disk striping with parity 60 00:02:57,180 --> 00:02:59,940 gives us both the performance enhancement 61 00:02:59,940 --> 00:03:01,683 and fault tolerance. 62 00:03:02,610 --> 00:03:04,560 Now, what do we do if we lose a component? 63 00:03:04,560 --> 00:03:06,090 Well, we're going to have to replace 64 00:03:06,090 --> 00:03:08,040 that component or that drive. 65 00:03:08,040 --> 00:03:09,480 And just as a refresher, 66 00:03:09,480 --> 00:03:11,790 we have different ways that we do replacements 67 00:03:11,790 --> 00:03:13,500 referred to as swaps. 68 00:03:13,500 --> 00:03:16,410 A cold swap is when we need to power down a system 69 00:03:16,410 --> 00:03:18,330 in order to replace a component. 70 00:03:18,330 --> 00:03:21,420 A warm swap is the ability to insert and remove hardware 71 00:03:21,420 --> 00:03:24,030 while the system is in a suspended state. 72 00:03:24,030 --> 00:03:26,550 And the hot swap is the ability to insert 73 00:03:26,550 --> 00:03:29,910 and remove hardware while the system is running. 74 00:03:29,910 --> 00:03:32,040 Now, drive redundancy is absolutely 75 00:03:32,040 --> 00:03:34,380 a key component of resiliency. 76 00:03:34,380 --> 00:03:37,650 It's an investment that recognizes the importance 77 00:03:37,650 --> 00:03:40,230 of system availability. 78 00:03:40,230 --> 00:03:44,793 And that, my friends, is a closer look at Raid 1 and Raid 5.