1 00:00:00,480 --> 00:00:02,280 In this lesson, you're going to learn 2 00:00:02,280 --> 00:00:03,840 how to use logical volumes 3 00:00:03,840 --> 00:00:05,400 to create a more flexible structure 4 00:00:05,400 --> 00:00:08,100 for organizing data on storage devices, 5 00:00:08,100 --> 00:00:10,890 because after all, partitioning is not the only way 6 00:00:10,890 --> 00:00:13,740 to divide up a storage device logically. 7 00:00:13,740 --> 00:00:15,210 As we discussed before, 8 00:00:15,210 --> 00:00:17,730 we've made the use of physical storage devices, 9 00:00:17,730 --> 00:00:21,000 but device mapping turns these physical storage devices 10 00:00:21,000 --> 00:00:23,100 into virtual storage devices. 11 00:00:23,100 --> 00:00:26,250 In Linux, the device mapper creates the virtual device 12 00:00:26,250 --> 00:00:28,440 and passes data from that virtual device 13 00:00:28,440 --> 00:00:30,810 to one or more physical devices. 14 00:00:30,810 --> 00:00:33,750 DM-multipath is a feature of the Linux Kernel 15 00:00:33,750 --> 00:00:35,970 that provides redundancy and improved performance 16 00:00:35,970 --> 00:00:37,770 for block storage devices. 17 00:00:37,770 --> 00:00:40,800 If one path fails, DM-multipath will switch 18 00:00:40,800 --> 00:00:42,810 to one of the other paths that remains, 19 00:00:42,810 --> 00:00:44,610 keeping that storage device available 20 00:00:44,610 --> 00:00:46,200 for reading and writing. 21 00:00:46,200 --> 00:00:48,031 The multipath tools package enables you 22 00:00:48,031 --> 00:00:51,060 to manage DM-multipath for storage devices 23 00:00:51,060 --> 00:00:53,700 and typically, you're going to find the configuration file 24 00:00:53,700 --> 00:00:58,700 located at /etc/multipath.conf. 25 00:00:59,070 --> 00:01:01,766 The mdadm command is a tool that's going to be used 26 00:01:01,766 --> 00:01:04,470 to manage software-based RAID arrays. 27 00:01:04,470 --> 00:01:06,360 In a RAID array, data is stored 28 00:01:06,360 --> 00:01:08,700 across multiple physical storage devices 29 00:01:08,700 --> 00:01:10,350 and those devices are combined 30 00:01:10,350 --> 00:01:12,900 into a single virtual storage device. 31 00:01:12,900 --> 00:01:15,270 This type of software-based RAID configuration 32 00:01:15,270 --> 00:01:17,490 is an alternative to using device mapper 33 00:01:17,490 --> 00:01:19,260 and the DM-multipath. 34 00:01:19,260 --> 00:01:22,080 The mdadm tool enables you to create, 35 00:01:22,080 --> 00:01:24,780 manage and monitor RAID arrays. 36 00:01:24,780 --> 00:01:26,280 Now, before we dive too much further 37 00:01:26,280 --> 00:01:28,140 into logical volumes and RAIDs, 38 00:01:28,140 --> 00:01:29,803 it's important for us to take a quick detour 39 00:01:29,803 --> 00:01:31,830 and do a quick review of RAIDs 40 00:01:31,830 --> 00:01:34,080 from your previous A plus studies. 41 00:01:34,080 --> 00:01:35,550 A RAID is a redundant array 42 00:01:35,550 --> 00:01:38,250 of independent or inexpensive disks. 43 00:01:38,250 --> 00:01:40,200 RAIDs are classified based on their level, 44 00:01:40,200 --> 00:01:45,200 such as RAID 0, RAID 1, RAID 5, RAID 6 and RAID 10. 45 00:01:45,480 --> 00:01:47,880 Each RAID has either striping, mirroring, 46 00:01:47,880 --> 00:01:50,190 parity, or a combination of these features, 47 00:01:50,190 --> 00:01:52,170 depending on the RAID level. 48 00:01:52,170 --> 00:01:54,690 Striping refers to a technique used by RAIDs 49 00:01:54,690 --> 00:01:56,940 in which multiple smaller physical disks 50 00:01:56,940 --> 00:02:00,540 can be combined to logically act as a single larger disk. 51 00:02:00,540 --> 00:02:03,600 For example, if I wanted to have a four terabyte hard drive 52 00:02:03,600 --> 00:02:05,310 for storage in my Linux system, 53 00:02:05,310 --> 00:02:08,460 but I only have two two terabyte hard drives available, 54 00:02:08,460 --> 00:02:11,039 I can use striping to combine them logically 55 00:02:11,039 --> 00:02:12,269 and my system is going to treat them 56 00:02:12,269 --> 00:02:15,660 as a single four terabyte hard disk drive. 57 00:02:15,660 --> 00:02:16,980 Mirroring, on the other hand, 58 00:02:16,980 --> 00:02:19,320 is the second technique that's used in a RAID. 59 00:02:19,320 --> 00:02:21,900 With mirroring, two physical hard disk drives 60 00:02:21,900 --> 00:02:24,690 are going to be combined into a single logical volume. 61 00:02:24,690 --> 00:02:27,330 An identical copy of everything that's put on them, 62 00:02:27,330 --> 00:02:29,670 is going to be put on both drives. 63 00:02:29,670 --> 00:02:31,830 This way, if a single drive fails, 64 00:02:31,830 --> 00:02:34,470 the system can quickly access all the files it needs, 65 00:02:34,470 --> 00:02:36,330 because there's still a fully redundant copy 66 00:02:36,330 --> 00:02:39,540 on that second drive as part of that RAID. 67 00:02:39,540 --> 00:02:42,330 Parity is the third technique we use in RAIDs. 68 00:02:42,330 --> 00:02:44,400 Now, parity computations are going to be used 69 00:02:44,400 --> 00:02:46,650 in RAID drive arrays for fault tolerance 70 00:02:46,650 --> 00:02:48,960 by calculating the data in two drives 71 00:02:48,960 --> 00:02:51,450 and storing the results on a different drive. 72 00:02:51,450 --> 00:02:53,430 For example, in a RAID 5, 73 00:02:53,430 --> 00:02:55,770 there's going to be a minimum of three disk drives 74 00:02:55,770 --> 00:02:56,610 and these are going to be used 75 00:02:56,610 --> 00:02:58,740 so that data can be striped across the drives 76 00:02:58,740 --> 00:03:01,140 by splitting each file into two pieces, 77 00:03:01,140 --> 00:03:02,730 putting one piece on one drive 78 00:03:02,730 --> 00:03:04,860 and the other piece on the second drive, 79 00:03:04,860 --> 00:03:08,790 and then the calculated parity gets put on a third drive. 80 00:03:08,790 --> 00:03:11,400 All right, now that we've defined the term striping, 81 00:03:11,400 --> 00:03:13,950 mirroring and parity, let's take a quick look 82 00:03:13,950 --> 00:03:15,180 at the different types of RAIDs 83 00:03:15,180 --> 00:03:17,700 that you can configure on your Linux system. 84 00:03:17,700 --> 00:03:19,890 First, we have a RAID 0. 85 00:03:19,890 --> 00:03:21,930 RAID 0 relies on striping. 86 00:03:21,930 --> 00:03:24,240 With a RAID 0, we have two different disks 87 00:03:24,240 --> 00:03:25,170 that are working together 88 00:03:25,170 --> 00:03:27,330 and each of them holds half of the data. 89 00:03:27,330 --> 00:03:29,340 So as you can see here, 90 00:03:29,340 --> 00:03:32,340 part of the file A1 and part of the file A2 91 00:03:32,340 --> 00:03:34,530 are each put on each of the drives. 92 00:03:34,530 --> 00:03:38,010 So let's assume that I have a RAID 0 with my Linux system. 93 00:03:38,010 --> 00:03:39,960 If I've installed it and configured it, 94 00:03:39,960 --> 00:03:42,660 what does the operating system actually see? 95 00:03:42,660 --> 00:03:44,820 Do they see disk zero and disk one? 96 00:03:44,820 --> 00:03:45,750 Well, no. 97 00:03:45,750 --> 00:03:48,210 Instead it's just going to see one logical drive 98 00:03:48,210 --> 00:03:51,030 in the RAID, because both disks are working together 99 00:03:51,030 --> 00:03:53,700 to create a single larger logical volume. 100 00:03:53,700 --> 00:03:55,440 Now, if I'm going to copy this video 101 00:03:55,440 --> 00:03:58,170 over to the new striped RAID 0 logical drive, 102 00:03:58,170 --> 00:03:59,910 what do you think is going to happen? 103 00:03:59,910 --> 00:04:02,730 Well, it's going to put the first piece on disk zero 104 00:04:02,730 --> 00:04:04,590 and the second piece on disk one 105 00:04:04,590 --> 00:04:06,420 and the third piece on disk zero, 106 00:04:06,420 --> 00:04:08,370 and the fourth piece on disk one. 107 00:04:08,370 --> 00:04:09,600 And it keeps doing this, 108 00:04:09,600 --> 00:04:11,490 where all of the even parts and odd parts 109 00:04:11,490 --> 00:04:14,280 are being split up across the two different disks. 110 00:04:14,280 --> 00:04:15,660 This is what's known as striping, 111 00:04:15,660 --> 00:04:18,540 because as I'm putting a piece on each of those two disks, 112 00:04:18,540 --> 00:04:20,910 if you look at it, it looks like a candy stripe 113 00:04:20,910 --> 00:04:22,050 or a peppermint stick, 114 00:04:22,050 --> 00:04:24,270 because it's swirling around those two RAIDs, 115 00:04:24,270 --> 00:04:27,510 with the evens on one side and the odds on the other. 116 00:04:27,510 --> 00:04:29,370 Now, RAID 0 is a great thing 117 00:04:29,370 --> 00:04:31,200 if you want to have faster speeds, 118 00:04:31,200 --> 00:04:34,200 because it stripes the data across two physical drives, 119 00:04:34,200 --> 00:04:37,380 but this provides no redundancy at all for your data. 120 00:04:37,380 --> 00:04:39,180 If one of these two disks fail, 121 00:04:39,180 --> 00:04:40,557 then you're going to lose half of the file. 122 00:04:40,557 --> 00:04:43,290 And if you lose half of the file, well, guess what? 123 00:04:43,290 --> 00:04:45,840 The computer isn't going to be able to read the file at all, 124 00:04:45,840 --> 00:04:48,750 so really, it's like you lost the whole file. 125 00:04:48,750 --> 00:04:50,670 Now, the good thing about using a stripe disk 126 00:04:50,670 --> 00:04:52,110 or a RAID 0 array, 127 00:04:52,110 --> 00:04:54,540 is the fact that you get this speed boost, right? 128 00:04:54,540 --> 00:04:56,940 And so as you're doing things like high speed gaming 129 00:04:56,940 --> 00:04:58,470 or high speed video editing, 130 00:04:58,470 --> 00:05:00,720 RAID 0s can be a great setup for this, 131 00:05:00,720 --> 00:05:03,150 because you can access two disks a lot faster 132 00:05:03,150 --> 00:05:04,710 than you can access one disk, 133 00:05:04,710 --> 00:05:05,910 because each of them is going to give you 134 00:05:05,910 --> 00:05:08,550 half of the information at the same time. 135 00:05:08,550 --> 00:05:11,010 Now, the other good thing about using a RAID 0, 136 00:05:11,010 --> 00:05:14,370 is the fact that there is no loss of space on those disks. 137 00:05:14,370 --> 00:05:16,710 If I take 2 one terabyte hard disks 138 00:05:16,710 --> 00:05:19,230 and I combine them logically into a RAID 0, 139 00:05:19,230 --> 00:05:21,750 this is going to appear as a single logical volume 140 00:05:21,750 --> 00:05:25,110 containing two terabytes of size for us to use. 141 00:05:25,110 --> 00:05:27,120 The next type of RAID we need to discuss, 142 00:05:27,120 --> 00:05:28,890 is known as a RAID 1. 143 00:05:28,890 --> 00:05:30,270 Whenever you hear RAID 1, 144 00:05:30,270 --> 00:05:34,290 I want you to think about two words, mirror and redundancy. 145 00:05:34,290 --> 00:05:37,620 When you have a RAID 1, you have a mirrored disk array. 146 00:05:37,620 --> 00:05:40,530 This means that every single thing that's put on disk zero, 147 00:05:40,530 --> 00:05:42,600 is also put on disk one. 148 00:05:42,600 --> 00:05:44,910 So if I have a file called A 149 00:05:44,910 --> 00:05:46,890 and I break it up into four pieces, 150 00:05:46,890 --> 00:05:49,320 both disk zero and disk one are going to have 151 00:05:49,320 --> 00:05:51,960 all of those four parts of file A. 152 00:05:51,960 --> 00:05:54,270 And as you can see here, one, two, three 153 00:05:54,270 --> 00:05:56,190 and four is on disk zero. 154 00:05:56,190 --> 00:05:59,160 One, two, three and four is also on disk one. 155 00:05:59,160 --> 00:06:02,550 So if I delete everything on disk one, guess what? 156 00:06:02,550 --> 00:06:05,250 I can still access it on disk zero. 157 00:06:05,250 --> 00:06:06,330 This is what we're talking about 158 00:06:06,330 --> 00:06:08,310 by having a mirror disk array. 159 00:06:08,310 --> 00:06:11,730 The fact is, both of these disks have an identical copy 160 00:06:11,730 --> 00:06:14,370 and this gives us a great amount of redundancy. 161 00:06:14,370 --> 00:06:16,230 In fact, it's fully redundant, 162 00:06:16,230 --> 00:06:17,460 because we have an entire disk 163 00:06:17,460 --> 00:06:20,580 that can go away and I can still operate with no problems, 164 00:06:20,580 --> 00:06:22,920 because I can rely on the other disk. 165 00:06:22,920 --> 00:06:25,800 There is one big downside to using a RAID 1 though, 166 00:06:25,800 --> 00:06:28,650 and that's the loss of space on one of those disks, 167 00:06:28,650 --> 00:06:31,290 because if I have to take two one terabyte hard disks, 168 00:06:31,290 --> 00:06:34,350 I no longer get two terabytes of storage space to use, 169 00:06:34,350 --> 00:06:37,290 instead I'm only getting one terabyte of space. 170 00:06:37,290 --> 00:06:38,123 Why? 171 00:06:38,123 --> 00:06:39,270 Because I have a mirrored copy 172 00:06:39,270 --> 00:06:41,040 of every single piece of data 173 00:06:41,040 --> 00:06:43,440 and that means I have to make two full copies 174 00:06:43,440 --> 00:06:44,970 and it takes up twice the amount of space 175 00:06:44,970 --> 00:06:46,470 to store those files. 176 00:06:46,470 --> 00:06:48,720 So if you have half of the total space 177 00:06:48,720 --> 00:06:50,580 being used simply for redundancy, 178 00:06:50,580 --> 00:06:53,280 this means you're paying a lot more for storage costs, 179 00:06:53,280 --> 00:06:55,770 because you're going to be using something like a RAID 1 here, 180 00:06:55,770 --> 00:06:57,030 instead of a RAID 0 181 00:06:57,030 --> 00:06:59,430 that allows you to use all of the disk space. 182 00:06:59,430 --> 00:07:02,130 So if you want to get the best of both worlds, 183 00:07:02,130 --> 00:07:03,300 you really might want to consider 184 00:07:03,300 --> 00:07:05,580 using something like a RAID 5. 185 00:07:05,580 --> 00:07:08,670 A RAID 5 is going to give you the redundancy through parity. 186 00:07:08,670 --> 00:07:10,590 So when you hear RAID 5, 187 00:07:10,590 --> 00:07:12,450 I want you to remember the keywords. 188 00:07:12,450 --> 00:07:15,090 These are redundancy and parity. 189 00:07:15,090 --> 00:07:17,100 Now, a RAID 5 is a bit different 190 00:07:17,100 --> 00:07:19,440 in how it provides that redundancy though. 191 00:07:19,440 --> 00:07:22,200 Instead of using a full mirror like a RAID 1 did 192 00:07:22,200 --> 00:07:25,140 and taking up an entire disk worth of space, 193 00:07:25,140 --> 00:07:28,650 a RAID 5 is going to use striping and a computed parity 194 00:07:28,650 --> 00:07:30,780 across three or more disks. 195 00:07:30,780 --> 00:07:33,090 Now, it's important to remember and to take note 196 00:07:33,090 --> 00:07:35,970 that a RAID 5 can use three or more disks, 197 00:07:35,970 --> 00:07:38,700 but you must have at least three disks 198 00:07:38,700 --> 00:07:41,160 in order to configure a RAID 5. 199 00:07:41,160 --> 00:07:44,220 Now, a RAID 5 with three disks is going to work like this. 200 00:07:44,220 --> 00:07:46,710 First, we have three disks, disk zero, 201 00:07:46,710 --> 00:07:48,510 disk one, and disk two. 202 00:07:48,510 --> 00:07:51,120 If I want to store four different files on those drives, 203 00:07:51,120 --> 00:07:54,660 I can do this by putting pieces of file A, B, C and D 204 00:07:54,660 --> 00:07:56,370 on all of those drives. 205 00:07:56,370 --> 00:08:00,450 Notice that B, C and D have this thing called a Bp, 206 00:08:00,450 --> 00:08:02,550 a Cp and a Dp. 207 00:08:02,550 --> 00:08:04,740 This is the computed parity. 208 00:08:04,740 --> 00:08:08,130 So when I take file B and I put half on disk zero 209 00:08:08,130 --> 00:08:10,642 and half on disk one, I then can do a calculation 210 00:08:10,642 --> 00:08:14,340 and put the results of that calculation onto disk two. 211 00:08:14,340 --> 00:08:16,080 Now, if I lose one of those disks, 212 00:08:16,080 --> 00:08:18,420 I can use the part that I have and the parity 213 00:08:18,420 --> 00:08:20,370 in order to calculate that file, 214 00:08:20,370 --> 00:08:22,650 or if I have the two parts and no parity, 215 00:08:22,650 --> 00:08:24,600 I can then simply recalculate the parity 216 00:08:24,600 --> 00:08:26,760 and recreate that file as well. 217 00:08:26,760 --> 00:08:28,710 Now, I know that sounds a little bit complicated, 218 00:08:28,710 --> 00:08:30,600 but I want you to think about it this way. 219 00:08:30,600 --> 00:08:33,299 Let's say I gave you two numbers and I give you an answer. 220 00:08:33,299 --> 00:08:36,390 For example, two plus three equals five. 221 00:08:36,390 --> 00:08:38,789 Now, if I take away any of the three numbers 222 00:08:38,789 --> 00:08:40,350 and I give you the other two, 223 00:08:40,350 --> 00:08:43,049 you could still figure out what the other one was, right? 224 00:08:43,049 --> 00:08:45,810 Of course you could, because if I gave you two and three, 225 00:08:45,810 --> 00:08:48,030 you could add them together and get five. 226 00:08:48,030 --> 00:08:50,880 If I gave you two plus something equals five, 227 00:08:50,880 --> 00:08:53,730 you could say, well, five minus two is three. 228 00:08:53,730 --> 00:08:55,650 So the answer must be three. 229 00:08:55,650 --> 00:08:58,350 Or if I gave you something plus three is five. 230 00:08:58,350 --> 00:09:00,630 You could say that five minus three is two. 231 00:09:00,630 --> 00:09:01,637 And therefore, you calculated 232 00:09:01,637 --> 00:09:03,960 that the missing number was two. 233 00:09:03,960 --> 00:09:06,420 That's what we're doing inside of a RAID 5. 234 00:09:06,420 --> 00:09:07,620 We're going to be able to recompute 235 00:09:07,620 --> 00:09:09,270 the missing parts of the file 236 00:09:09,270 --> 00:09:11,760 by using the known parts and the stored parity 237 00:09:11,760 --> 00:09:13,500 and this is how redundancy through parity 238 00:09:13,500 --> 00:09:14,850 is going to be achieved. 239 00:09:14,850 --> 00:09:16,170 The great thing about this, 240 00:09:16,170 --> 00:09:17,610 is that it takes up less disk space 241 00:09:17,610 --> 00:09:19,110 to store a computed parity 242 00:09:19,110 --> 00:09:21,540 than it does to store the entire mirrored file. 243 00:09:21,540 --> 00:09:24,330 So if I'm using three drives in a RAID 5, 244 00:09:24,330 --> 00:09:27,750 I'm only using one third of the space for that parity. 245 00:09:27,750 --> 00:09:29,220 If I use four drives, 246 00:09:29,220 --> 00:09:32,040 I'm losing only one fourth of the space for that parity. 247 00:09:32,040 --> 00:09:34,530 And if I'm using five drives, I'm only going to lose one fifth 248 00:09:34,530 --> 00:09:36,840 of the total space for that parity. 249 00:09:36,840 --> 00:09:39,210 This is a lot less than a full mirrored file. 250 00:09:39,210 --> 00:09:41,640 So it's a lot more efficient to create a RAID 5 251 00:09:41,640 --> 00:09:43,560 in terms of space, than using a RAID 1 252 00:09:43,560 --> 00:09:45,390 using a mirrored array. 253 00:09:45,390 --> 00:09:48,570 For this reason, RAID 5 is the most popular type of RAID 254 00:09:48,570 --> 00:09:50,910 in use today and you're going to find it heavily used 255 00:09:50,910 --> 00:09:53,040 in most server environments. 256 00:09:53,040 --> 00:09:55,320 Next, let's talk about RAID 6. 257 00:09:55,320 --> 00:09:57,360 What exactly is a RAID 6? 258 00:09:57,360 --> 00:10:00,330 Well, six is better than five by one 259 00:10:00,330 --> 00:10:01,950 and any time you hear RAID 6, 260 00:10:01,950 --> 00:10:04,260 I want you to think of five plus one. 261 00:10:04,260 --> 00:10:06,330 So everything that was true about a RAID 5 262 00:10:06,330 --> 00:10:08,520 is still true about a RAID 6, 263 00:10:08,520 --> 00:10:11,880 except we are going to have double parity in a RAID 6, 264 00:10:11,880 --> 00:10:13,530 instead of single parity. 265 00:10:13,530 --> 00:10:15,060 Because of this double parity, 266 00:10:15,060 --> 00:10:17,010 we also need to add an extra drive 267 00:10:17,010 --> 00:10:19,020 and have a minimum of at least four drives 268 00:10:19,020 --> 00:10:20,700 to create a RAID 6. 269 00:10:20,700 --> 00:10:24,660 Now, why would you want to use a RAID 6 instead of a RAID 5? 270 00:10:24,660 --> 00:10:27,750 Well, in a RAID 5, I can only lose one drive 271 00:10:27,750 --> 00:10:29,610 and keep operating that array. 272 00:10:29,610 --> 00:10:33,300 In RAID 6, I can actually lose two disks and keep operating. 273 00:10:33,300 --> 00:10:36,360 So if I have maybe five or 10 disk drives, 274 00:10:36,360 --> 00:10:37,860 using two of them for parity 275 00:10:37,860 --> 00:10:39,900 gives me a much better redundancy value 276 00:10:39,900 --> 00:10:41,400 than if I just use RAID 5 277 00:10:41,400 --> 00:10:44,070 with a single redundancy and single parity. 278 00:10:44,070 --> 00:10:47,790 This is the main difference between a RAID 5 and a RAID 6. 279 00:10:47,790 --> 00:10:49,710 Now, the last type of RAID we're going to discuss, 280 00:10:49,710 --> 00:10:51,660 is known as a RAID 10. 281 00:10:51,660 --> 00:10:54,540 A RAID 10 is really a RAID of RAIDs. 282 00:10:54,540 --> 00:10:57,570 What I mean by this, is that I actually have two RAID 1s 283 00:10:57,570 --> 00:11:00,990 and they're being placed inside a RAID 0 configuration. 284 00:11:00,990 --> 00:11:02,670 If we go back to a RAID 0, 285 00:11:02,670 --> 00:11:05,490 we remember this is all about speed and striping 286 00:11:05,490 --> 00:11:08,310 and so as the file comes in, I'm going to stripe it 287 00:11:08,310 --> 00:11:11,310 and put half to the left side and half to the right side. 288 00:11:11,310 --> 00:11:13,620 But because each of that left and right side 289 00:11:13,620 --> 00:11:15,300 is actually a mirrored RAID, 290 00:11:15,300 --> 00:11:17,700 I have all of the odd stuff on the left side 291 00:11:17,700 --> 00:11:19,920 and all of the even stuff on the right side, 292 00:11:19,920 --> 00:11:22,080 and now I have two full copies. 293 00:11:22,080 --> 00:11:25,140 Because of this setup, you need a minimum of four disks 294 00:11:25,140 --> 00:11:27,240 to be able to create a RAID 10. 295 00:11:27,240 --> 00:11:30,360 This means I'm going to lose 50% of my total disk space 296 00:11:30,360 --> 00:11:33,120 to create this striped array of mirrored to arrays, 297 00:11:33,120 --> 00:11:34,223 but I do get the benefit 298 00:11:34,223 --> 00:11:38,430 of having two fully redundant mirrors inside my RAID 10. 299 00:11:38,430 --> 00:11:40,470 RAID 10 is also pretty fast, 300 00:11:40,470 --> 00:11:43,080 because you're using that RAID 0 striping technique 301 00:11:43,080 --> 00:11:44,970 across two mirrored arrays. 302 00:11:44,970 --> 00:11:48,270 So with RAID 10, you get the benefits of RAID 1 303 00:11:48,270 --> 00:11:50,520 and the benefits of RAID 0, 304 00:11:50,520 --> 00:11:53,580 but you're having to use four disks to do this. 305 00:11:53,580 --> 00:11:56,340 All right, now that we've reviewed the types of RAIDs, 306 00:11:56,340 --> 00:11:57,210 you might be wondering, 307 00:11:57,210 --> 00:11:58,994 how do I get information about my RAID 308 00:11:58,994 --> 00:12:01,440 from my Linux system and its Kernel? 309 00:12:01,440 --> 00:12:06,210 Well, to do that, you're going to use /proc/mdstat. 310 00:12:06,210 --> 00:12:08,370 This is a file and it's a text-based file 311 00:12:08,370 --> 00:12:12,060 that contains a snapshot of the Kernel's RAID or MD state. 312 00:12:12,060 --> 00:12:13,770 To read this file on your system, 313 00:12:13,770 --> 00:12:17,340 simply type cat /proc/mdstat 314 00:12:17,340 --> 00:12:20,130 in the command line interface and then hit ENTER. 315 00:12:20,130 --> 00:12:21,840 This will display the information to your screen 316 00:12:21,840 --> 00:12:25,020 about any known RAID configurations on your system. 317 00:12:25,020 --> 00:12:27,120 Now, for example, on one of my systems, 318 00:12:27,120 --> 00:12:30,240 I get the following returned when I enter this command. 319 00:12:30,240 --> 00:12:32,400 Under personalities, the file's going to tell me 320 00:12:32,400 --> 00:12:34,920 which RAIDs could be supported by my Kernel. 321 00:12:34,920 --> 00:12:37,710 In this case, it says I can support a RAID 6, 322 00:12:37,710 --> 00:12:39,930 a RAID 5 or a RAID 4. 323 00:12:39,930 --> 00:12:42,000 Next, we see that md0, 324 00:12:42,000 --> 00:12:44,430 which is the name of my array, is active. 325 00:12:44,430 --> 00:12:45,837 It's configured as a RAID 5 326 00:12:45,837 --> 00:12:49,290 and it's using three disk drives to make up that RAID 5. 327 00:12:49,290 --> 00:12:53,220 Sda1, sdd1 and sdb1. 328 00:12:53,220 --> 00:12:55,620 The numbers after each of these tells me which disk 329 00:12:55,620 --> 00:12:57,660 is physically located in the array 330 00:12:57,660 --> 00:13:01,050 and this is actually disk zero, which is sda1, 331 00:13:01,050 --> 00:13:05,820 disk one which is sdd1, and disk two which is sdd1. 332 00:13:05,820 --> 00:13:08,310 Next, we see the number of blocks supported, 333 00:13:08,310 --> 00:13:10,830 the size of the chunks and the status of the algorithm 334 00:13:10,830 --> 00:13:13,980 and then we can see the status of the disks in that RAID. 335 00:13:13,980 --> 00:13:15,390 And this is shown as UUU_. 336 00:13:17,070 --> 00:13:20,160 This means up, up, up and down. 337 00:13:20,160 --> 00:13:21,990 Notice, there's that underscore 338 00:13:21,990 --> 00:13:23,700 and when there's an underscore shown, 339 00:13:23,700 --> 00:13:25,920 this means there's a disk that is down. 340 00:13:25,920 --> 00:13:28,170 In this case, we can see the fourth disk, 341 00:13:28,170 --> 00:13:30,480 which is probably going to be sdc1. 342 00:13:30,480 --> 00:13:33,210 It's going to be down, because there's an underscore shown. 343 00:13:33,210 --> 00:13:35,760 This explains why the Kernel supports RAID 6, 344 00:13:35,760 --> 00:13:39,300 which requires four disks, but we're only using a RAID 5, 345 00:13:39,300 --> 00:13:41,370 which uses three disks, because we have three disks 346 00:13:41,370 --> 00:13:43,170 that are up and configured. 347 00:13:43,170 --> 00:13:45,990 Next, we're going to discuss one of the major applications 348 00:13:45,990 --> 00:13:47,250 of the device mapper, 349 00:13:47,250 --> 00:13:51,120 which is called the logical volume manager, or LVM. 350 00:13:51,120 --> 00:13:54,750 The LVM will map whole physical devices and partitions 351 00:13:54,750 --> 00:13:58,530 into one or more virtual containers, known as volume groups. 352 00:13:58,530 --> 00:14:02,370 Within these volume groups, are one or more logical volumes. 353 00:14:02,370 --> 00:14:05,463 Ultimately, these logical volumes become the storage devices 354 00:14:05,463 --> 00:14:08,580 that the system, the user and the applications 355 00:14:08,580 --> 00:14:10,980 are going to interact with and work with. 356 00:14:10,980 --> 00:14:14,310 With the logical volume manager, you can dynamically create, 357 00:14:14,310 --> 00:14:16,110 delete and resize volumes 358 00:14:16,110 --> 00:14:17,970 without having to reboot the system. 359 00:14:17,970 --> 00:14:20,130 You can also map multiple logical volumes 360 00:14:20,130 --> 00:14:22,290 across multiple physical devices 361 00:14:22,290 --> 00:14:26,040 and you can create virtual snapshots of each logical volume 362 00:14:26,040 --> 00:14:28,470 so that you can quickly and easily revert a volume 363 00:14:28,470 --> 00:14:30,330 to a specific state. 364 00:14:30,330 --> 00:14:34,470 The /dev/mapper directory contains all the logical volumes 365 00:14:34,470 --> 00:14:36,510 on a given system that are being managed 366 00:14:36,510 --> 00:14:38,700 by the logical volume manager. 367 00:14:38,700 --> 00:14:39,765 Devices in this directory 368 00:14:39,765 --> 00:14:43,830 are typically going to be called /dev/mapper, 369 00:14:43,830 --> 00:14:47,700 the volume group name, -the logical volume name. 370 00:14:47,700 --> 00:14:49,140 the logical volume manager 371 00:14:49,140 --> 00:14:50,477 divides its volume management tools 372 00:14:50,477 --> 00:14:54,030 into three categories based on three different components, 373 00:14:54,030 --> 00:14:57,120 including physical volume tools, volume group tools 374 00:14:57,120 --> 00:14:59,160 and logical volume tools. 375 00:14:59,160 --> 00:15:02,190 Some logical volume managers' physical volume tools 376 00:15:02,190 --> 00:15:04,380 will include things like pvscan, 377 00:15:04,380 --> 00:15:06,960 which is used to scan for all the physical devices 378 00:15:06,960 --> 00:15:10,080 that are being mounted and used as physical volumes. 379 00:15:10,080 --> 00:15:12,960 You also have pvcreate, which is used to initialize 380 00:15:12,960 --> 00:15:16,260 a driver partition to use as a physical volume. 381 00:15:16,260 --> 00:15:18,810 We also have pvdisplay, which is used to list 382 00:15:18,810 --> 00:15:21,090 the attributes of physical volumes. 383 00:15:21,090 --> 00:15:23,880 We have pvchange, which is used to change attributes 384 00:15:23,880 --> 00:15:25,530 of a physical volume. 385 00:15:25,530 --> 00:15:27,966 We have pvs, which is used to display information 386 00:15:27,966 --> 00:15:30,000 about physical volumes. 387 00:15:30,000 --> 00:15:32,430 We have pvck, which is used to check 388 00:15:32,430 --> 00:15:34,650 the metadata of physical volumes. 389 00:15:34,650 --> 00:15:36,450 And we have pvremove, 390 00:15:36,450 --> 00:15:39,120 which is used to remove physical volumes. 391 00:15:39,120 --> 00:15:41,400 Now, we also have some volume group tools. 392 00:15:41,400 --> 00:15:43,410 This includes vgscan, 393 00:15:43,410 --> 00:15:46,830 which is going to scan all physical devices for volume groups. 394 00:15:46,830 --> 00:15:50,010 We have vgcreate to create our volume groups. 395 00:15:50,010 --> 00:15:52,257 We have vgdisplay to list out attributes 396 00:15:52,257 --> 00:15:53,940 of our volume groups. 397 00:15:53,940 --> 00:15:56,520 We have vgchange to change the attributes 398 00:15:56,520 --> 00:15:57,960 of our volume groups. 399 00:15:57,960 --> 00:16:01,890 We have vgs to display information about our volume groups. 400 00:16:01,890 --> 00:16:06,150 We have vgck to check the metadata of those volume groups. 401 00:16:06,150 --> 00:16:09,150 We have vgrename to rename a volume group. 402 00:16:09,150 --> 00:16:11,880 We have vgreduce to remove physical volumes 403 00:16:11,880 --> 00:16:14,670 from a volume group to reduce its size. 404 00:16:14,670 --> 00:16:17,730 We have vgextend, which allows us to add physical volumes 405 00:16:17,730 --> 00:16:19,080 to the volume groups. 406 00:16:19,080 --> 00:16:21,030 We have vgmerge, where we can merge 407 00:16:21,030 --> 00:16:22,620 two volume groups together. 408 00:16:22,620 --> 00:16:24,900 We have vgsplit, which can be used to split 409 00:16:24,900 --> 00:16:26,700 a volume group into two groups 410 00:16:26,700 --> 00:16:28,020 and we have vgremove, 411 00:16:28,020 --> 00:16:30,420 where we can remove those volume groups. 412 00:16:30,420 --> 00:16:32,730 When we get down to our logical volume tools, 413 00:16:32,730 --> 00:16:34,410 we have things like lvscan, 414 00:16:34,410 --> 00:16:37,770 which scans all the physical devices for logical volumes. 415 00:16:37,770 --> 00:16:40,380 We have lvcreate to create logical volumes 416 00:16:40,380 --> 00:16:41,820 in a volume group. 417 00:16:41,820 --> 00:16:45,810 We have lvdisplay to list the attributes of logical volumes. 418 00:16:45,810 --> 00:16:48,027 Lvchange is going to be used to change attributes 419 00:16:48,027 --> 00:16:49,860 of the volumes. 420 00:16:49,860 --> 00:16:53,340 Lvs will display information about logical volumes 421 00:16:53,340 --> 00:16:56,730 and lvrename is used to rename the logical volumes. 422 00:16:56,730 --> 00:16:58,050 If you use lvreduce, 423 00:16:58,050 --> 00:17:00,780 that'll reduce the size of your logical volumes. 424 00:17:00,780 --> 00:17:04,319 Lvextend will extend the size of your logical volumes. 425 00:17:04,319 --> 00:17:07,290 Lvresize will resize the logical volumes 426 00:17:07,290 --> 00:17:11,130 and finally, lvremove will remove the logical volumes. 427 00:17:11,130 --> 00:17:12,359 Now, you may be wondering, 428 00:17:12,359 --> 00:17:14,396 do I have to memorize that long list of tools 429 00:17:14,396 --> 00:17:16,170 that Jason just read out? 430 00:17:16,170 --> 00:17:19,650 And the answer is no, because you always have the help files 431 00:17:19,650 --> 00:17:21,329 and the man files at your system 432 00:17:21,329 --> 00:17:23,220 when you're doing this in the real world. 433 00:17:23,220 --> 00:17:24,938 For the exam, they're not going to ask you 434 00:17:24,938 --> 00:17:27,089 each and every one of these commands, 435 00:17:27,089 --> 00:17:29,010 but it is something I wanted to introduce you to 436 00:17:29,010 --> 00:17:30,660 so you can get comfortable with the idea 437 00:17:30,660 --> 00:17:32,850 of how to work with logical volumes 438 00:17:32,850 --> 00:17:34,893 inside the logical volume manager.