1 00:00:00,000 --> 00:00:00,900 W. Curtis Preston: Like my friend, Jeff. 2 00:00:00,929 --> 00:00:02,280 See you do cigars. 3 00:00:02,280 --> 00:00:03,090 You get COVID. 4 00:00:03,269 --> 00:00:03,719 That's it. 5 00:00:03,990 --> 00:00:05,550 That's that's the lesson that 6 00:00:05,550 --> 00:00:06,449 Jeff should learn. 7 00:00:07,275 --> 00:00:09,165 Prasanna Malaiyandi: Curtis and I are not medical doctors. 8 00:00:09,165 --> 00:00:10,875 We are not providing any medical advice. 9 00:00:11,055 --> 00:00:14,295 Please go talk to your doctor or do your research. 10 00:00:14,325 --> 00:00:14,715 Thank you. 11 00:00:14,879 --> 00:00:17,069 W. Curtis Preston: please don't get your medical advice on 12 00:00:17,369 --> 00:00:19,079 Backup Central's Restore it All 13 00:00:38,780 --> 00:00:41,690 hi and welcome to the Backup Central's Restore it All podcast. 14 00:00:41,690 --> 00:00:42,170 I'm your host. 15 00:00:42,170 --> 00:00:42,380 W. 16 00:00:42,380 --> 00:00:44,510 Curtis Preston AKA Mr. 17 00:00:44,510 --> 00:00:48,920 Backup and I have with me, my COVID isolation grief 18 00:00:48,950 --> 00:00:52,330 consultant, Prasanna Malaiyandi 19 00:00:54,305 --> 00:00:55,375 Prasanna Malaiyandi: oh, no. 20 00:00:56,840 --> 00:00:57,830 no. 21 00:00:58,070 --> 00:01:01,700 Oh, Curtis say it ain't so, 22 00:01:02,045 --> 00:01:08,555 W. Curtis Preston: So I don't have COVID, but I just spent six 23 00:01:08,555 --> 00:01:13,745 hours with Jeff who has COVID 24 00:01:14,430 --> 00:01:14,700 Prasanna Malaiyandi: oh, 25 00:01:14,765 --> 00:01:19,655 W. Curtis Preston: and I just got told, so I now have to isolate. 26 00:01:21,265 --> 00:01:25,615 For the next few days, you know, and here's the thing I had a 27 00:01:25,615 --> 00:01:28,135 trip planned, uh, and purchased. 28 00:01:28,135 --> 00:01:32,755 And so, you know, Druva's uh new corporate office opens up this month. 29 00:01:32,755 --> 00:01:34,225 So there's this big shindig. 30 00:01:34,765 --> 00:01:42,745 Marketing department has their first, uh, like social gathering since COVID. 31 00:01:43,165 --> 00:01:44,965 And I will be at neither of those things. 32 00:01:44,965 --> 00:01:48,595 I will be at home hoping that I don't get COVID. 33 00:01:48,680 --> 00:01:51,140 Prasanna Malaiyandi: When is the event supposed to be? 34 00:01:51,725 --> 00:01:53,345 W. Curtis Preston: uh, Wednesday. 35 00:01:53,940 --> 00:01:54,870 Prasanna Malaiyandi: Oh yeah. 36 00:01:54,930 --> 00:01:55,350 Okay. 37 00:01:55,440 --> 00:01:56,340 Yeah, you're not. 38 00:01:56,760 --> 00:01:58,710 Oh, I'm so sorry. 39 00:01:58,865 --> 00:01:59,105 W. Curtis Preston: yeah. 40 00:01:59,280 --> 00:01:59,760 Prasanna Malaiyandi: sucks. 41 00:02:00,875 --> 00:02:04,145 W. Curtis Preston: See, I knew that you would, you would console me in my grief. 42 00:02:05,260 --> 00:02:07,420 Prasanna Malaiyandi: I hope Jeff is feeling better or feeling OK. 43 00:02:07,475 --> 00:02:09,785 W. Curtis Preston: I called him, you know, he, he texted me. 44 00:02:09,785 --> 00:02:10,775 He's like, yeah. 45 00:02:10,775 --> 00:02:14,165 So I have COVID and I call him. 46 00:02:14,165 --> 00:02:18,995 I was like, dude, like of all the weeks, you know, and you know, and you know, 47 00:02:18,995 --> 00:02:20,585 not your fault, but of all the weeks. 48 00:02:20,585 --> 00:02:24,515 And he goes, well, if it makes you feel any better, you did the same thing to me. 49 00:02:24,545 --> 00:02:26,135 I'm like, oh yeah. 50 00:02:26,195 --> 00:02:26,675 Right. 51 00:02:26,825 --> 00:02:31,985 Cause when he, he came over for Christmas dinner and then, uh, I tested positive. 52 00:02:34,345 --> 00:02:38,575 Prasanna Malaiyandi: So since we're talking about, uh, isolate, not isolation, 53 00:02:38,575 --> 00:02:40,885 but sort of needing grief from sympathy. 54 00:02:41,185 --> 00:02:44,275 So I think I called you over the weekend and told you how I had food 55 00:02:44,275 --> 00:02:52,705 poisoning, which was not a fun thing at all, but I'm happy to say that 56 00:02:52,705 --> 00:02:53,965 I feel much, much, much better. 57 00:02:53,995 --> 00:02:57,295 It was like 12 hours in my system felt awful. 58 00:02:57,335 --> 00:02:58,805 But back to normal. 59 00:03:01,235 --> 00:03:03,875 And I, think that might be the second time I've ever had food poisoning, 60 00:03:05,050 --> 00:03:05,410 W. Curtis Preston: Yeah. 61 00:03:05,410 --> 00:03:06,430 I've had it more than that. 62 00:03:06,670 --> 00:03:09,610 Um, but, and you may recall I had it last year. 63 00:03:09,615 --> 00:03:09,795 Prasanna Malaiyandi: Yeah. 64 00:03:09,795 --> 00:03:11,175 Which is why I called you for advice. 65 00:03:11,175 --> 00:03:14,265 I'm like, Curtis, what do I do? 66 00:03:15,960 --> 00:03:16,650 W. Curtis Preston: yeah. 67 00:03:16,860 --> 00:03:17,070 Yeah. 68 00:03:17,070 --> 00:03:20,310 It's more like, I'm like, go get a big greasy hamburger. 69 00:03:20,955 --> 00:03:21,195 Prasanna Malaiyandi: Yeah. 70 00:03:22,050 --> 00:03:24,900 W. Curtis Preston: I just tried to spill the tea on my desk. 71 00:03:25,920 --> 00:03:28,170 Well, technically I did spill the tea on my desk. 72 00:03:28,170 --> 00:03:29,340 I just didn't spill all of it. 73 00:03:31,230 --> 00:03:31,710 Um, 74 00:03:33,375 --> 00:03:34,035 Prasanna Malaiyandi: So have 75 00:03:34,035 --> 00:03:34,365 W. Curtis Preston: me while 76 00:03:35,205 --> 00:03:37,635 Prasanna Malaiyandi: gotten tested yet or you're waiting 77 00:03:37,655 --> 00:03:39,125 W. Curtis Preston: no, I haven't tested yet. 78 00:03:39,185 --> 00:03:41,375 It's kind of a waste to test this early. 79 00:03:41,525 --> 00:03:43,325 You know, I've got tests. 80 00:03:43,565 --> 00:03:45,545 I'll, I'll wait a couple of days. 81 00:03:45,635 --> 00:03:48,215 Um, and then test and I'm probably fine. 82 00:03:48,215 --> 00:03:49,265 I probably don't have it. 83 00:03:49,505 --> 00:03:54,515 And if I do have it, um, I will have a mild case, but, uh, like the 84 00:03:54,515 --> 00:03:56,345 one I had back in December, but. 85 00:03:56,550 --> 00:03:56,790 Prasanna Malaiyandi: Yeah. 86 00:03:57,485 --> 00:04:00,665 W. Curtis Preston: I don't need to go to a large unmasked gathering 87 00:04:00,665 --> 00:04:06,275 of people right now, which is I don't want to be typhoid Mary. 88 00:04:06,880 --> 00:04:07,570 Prasanna Malaiyandi: Yes. 89 00:04:08,500 --> 00:04:11,140 I'm glad you're taking the initiative though. 90 00:04:11,140 --> 00:04:14,440 Not to go because I'm sure there are a lot of people in your position 91 00:04:14,440 --> 00:04:15,520 or you'd be like, yeah, it's fine. 92 00:04:15,580 --> 00:04:16,240 I'll just go. 93 00:04:17,530 --> 00:04:18,010 W. Curtis Preston: Yeah. 94 00:04:18,370 --> 00:04:18,820 Yeah. 95 00:04:20,110 --> 00:04:26,290 I'm uh, well, Jeff's wife, um, actually got on a plane before he found out. 96 00:04:26,530 --> 00:04:30,940 So, you know, obviously she spent lots of time with him and then she 97 00:04:30,940 --> 00:04:34,150 got on a plane this morning to go visit her relatives in Florida. 98 00:04:34,720 --> 00:04:38,020 And, you know, he notified her like in the air. 99 00:04:38,890 --> 00:04:42,610 So he's like, well, I guess you're going to be masking up 100 00:04:42,610 --> 00:04:43,850 when you're around your family. 101 00:04:45,275 --> 00:04:48,755 So, you know, it stinks cause you know, and the thing is, 102 00:04:48,845 --> 00:04:50,615 again, I'll be fine if I get it. 103 00:04:50,885 --> 00:04:54,995 Um, but I mean, due to both the vaccine and also, 104 00:04:56,280 --> 00:04:57,960 Prasanna Malaiyandi: Having had it back in December 105 00:04:58,130 --> 00:05:01,235 W. Curtis Preston: Yeah, I've had it and I should be fine, but I don't 106 00:05:01,235 --> 00:05:03,995 want to be the person who gives it to somebody else where it might have, 107 00:05:04,055 --> 00:05:05,495 you know, much more negative effects. 108 00:05:05,495 --> 00:05:08,855 So I'm going to take the precautionary stuff and it stinks. 109 00:05:08,885 --> 00:05:13,925 I literally found out like, as my daughter was stopping by the house 110 00:05:13,925 --> 00:05:15,995 and I'm like, yeah, by the way, I'm not going to come say hi to you. 111 00:05:15,995 --> 00:05:18,305 Cause, cause I got, you know, 112 00:05:19,145 --> 00:05:21,905 Prasanna Malaiyandi: Oh, well, be responsible. 113 00:05:21,905 --> 00:05:22,475 Be good. 114 00:05:23,170 --> 00:05:24,970 W. Curtis Preston: Be be responsible. 115 00:05:25,060 --> 00:05:25,300 Nope. 116 00:05:25,360 --> 00:05:25,990 No, that's not. 117 00:05:27,040 --> 00:05:28,930 No, that doesn't make a good cheer. 118 00:05:29,635 --> 00:05:30,085 Prasanna Malaiyandi: No, 119 00:05:31,125 --> 00:05:31,515 W. Curtis Preston: No. 120 00:05:32,445 --> 00:05:36,885 Well, uh, I threw out the Druva name, saw throw that out again, 121 00:05:36,885 --> 00:05:40,665 that a, this is a, that Prasanna and I worked for different companies. 122 00:05:40,665 --> 00:05:41,475 He works for Zoom. 123 00:05:41,475 --> 00:05:44,985 I work for Druva that's opening up a new. 124 00:05:46,230 --> 00:05:51,570 You know, corporate office that I won't be visiting, but this is 125 00:05:51,810 --> 00:05:53,370 not a podcast of either company. 126 00:05:53,640 --> 00:05:57,690 The opinions that you hear are ours, and you're going to get some opinions. 127 00:05:57,720 --> 00:06:00,090 This episode, I'll tell you, you will get some opinions, 128 00:06:00,240 --> 00:06:01,740 some very strong ones from me. 129 00:06:02,220 --> 00:06:08,290 And then, uh, also, uh, please rate this podcast at ratethispodcast.com/restore. 130 00:06:08,670 --> 00:06:11,370 If you're listening to us on iTunes, just scroll down to the bottom, 131 00:06:11,460 --> 00:06:13,230 click some stars, make us a comment. 132 00:06:13,775 --> 00:06:17,255 Do all the things, you know, it just helps us to, to get the word out. 133 00:06:17,705 --> 00:06:21,665 And, um, and then also if you're interested in the kind of things we're 134 00:06:21,665 --> 00:06:26,435 interested, we want to have you on just reach out to me at what's that. 135 00:06:27,125 --> 00:06:28,115 Prasanna Malaiyandi: Come on the podcast. 136 00:06:28,540 --> 00:06:29,500 W. Curtis Preston: Come on the podcast. 137 00:06:29,500 --> 00:06:33,790 We love new people, you know, um, even if they say mean things like the one that, 138 00:06:34,190 --> 00:06:40,030 that we had last week, the one that got published today, the backup is evil. 139 00:06:40,300 --> 00:06:44,830 That was an interesting, uh, the first few minutes were rough. 140 00:06:44,860 --> 00:06:49,780 Uh, cause cause he was really just slamming down on backup, 141 00:06:49,810 --> 00:06:52,600 but, uh, that was a good episode. 142 00:06:52,600 --> 00:06:54,370 The backup is evil episode is a good one. 143 00:06:55,055 --> 00:06:56,615 Prasanna Malaiyandi: you don't have to always agree with us. 144 00:06:56,615 --> 00:06:57,815 It's totally fine. 145 00:06:58,590 --> 00:07:00,900 W. Curtis Preston: and you know, it would be nice to you, even if you think 146 00:07:00,990 --> 00:07:04,890 that, uh, for example, I'm going to make some really emphatic emphatical 147 00:07:05,040 --> 00:07:08,100 emphatic, emphatic, emphatic statements. 148 00:07:08,340 --> 00:07:11,340 This, this podcast, if you think that one or more of them are wrong, 149 00:07:11,640 --> 00:07:13,140 come on, we'll talk to you, right. 150 00:07:13,290 --> 00:07:13,890 We'll be nice. 151 00:07:13,890 --> 00:07:14,430 We promise. 152 00:07:14,790 --> 00:07:15,480 It's virtual. 153 00:07:15,480 --> 00:07:18,810 It's not like we can, uh, will Smith you, um, 154 00:07:19,160 --> 00:07:22,010 Prasanna Malaiyandi: Oh, uh, might be too soon. 155 00:07:22,010 --> 00:07:23,420 Curtis might be too soon. 156 00:07:24,290 --> 00:07:27,200 W. Curtis Preston: Well, yeah, I, you know, I went the thing with Jeff. 157 00:07:27,200 --> 00:07:32,720 We went to the academy museum again and saw a showing of the French connection 158 00:07:32,780 --> 00:07:38,090 introduced by the director of the French connection, which was amazing. 159 00:07:38,090 --> 00:07:39,170 He's 87 years old. 160 00:07:39,170 --> 00:07:39,920 It was an amazing thing. 161 00:07:39,920 --> 00:07:43,700 I don't have time to go into it, but I will say that he opened the thing 162 00:07:43,820 --> 00:07:45,800 by saying it's so great to be here. 163 00:07:46,410 --> 00:07:51,390 At the, uh, academy of motion pictures, a place known for, uh, will Smith. 164 00:07:55,685 --> 00:07:56,465 Prasanna Malaiyandi: Uh, 165 00:07:57,060 --> 00:07:57,930 W. Curtis Preston: He's like too soon? 166 00:07:58,020 --> 00:07:58,470 Too soon? 167 00:08:01,020 --> 00:08:05,580 The sense of humor and, and cognitive abilities of this 87 year old guy. 168 00:08:06,875 --> 00:08:08,075 They were great. 169 00:08:08,375 --> 00:08:09,095 It was just awesome. 170 00:08:10,295 --> 00:08:15,965 Anyway, so I, you know, the, I was inspired by the backup is evil episode, 171 00:08:15,965 --> 00:08:21,695 and I wanted to have a little bit another episode that delves into some 172 00:08:21,695 --> 00:08:27,425 of the things that we covered there, but, but just, um, because he was, 173 00:08:27,455 --> 00:08:30,365 he was very much heading towards. 174 00:08:32,005 --> 00:08:38,305 Obviously obvi, uh, the NetApp way of doing things, which is block 175 00:08:38,305 --> 00:08:42,175 level replication, incremental, incremental block level forever, 176 00:08:42,355 --> 00:08:44,575 and replication not backup. 177 00:08:44,805 --> 00:08:50,425 And let me just take a, take a moment to define what's different between 178 00:08:50,425 --> 00:08:55,915 backup and replication and for me and I don't, and I don't mean, I don't mean 179 00:08:55,915 --> 00:08:57,895 like snapshots and replication, what. 180 00:08:58,085 --> 00:08:59,295 Prasanna Malaiyandi: know, what's going on in my head. 181 00:09:00,135 --> 00:09:02,505 Remember you sent me, uh, the Reddit. 182 00:09:03,305 --> 00:09:09,845 A meme about with, uh, Chris rock and will Smith, where it was 183 00:09:09,875 --> 00:09:13,115 like, snapshots is not backup. 184 00:09:13,145 --> 00:09:13,925 And there was, 185 00:09:14,975 --> 00:09:15,255 W. Curtis Preston: Yeah. 186 00:09:15,255 --> 00:09:15,595 That one. 187 00:09:15,595 --> 00:09:15,865 Yeah, 188 00:09:16,835 --> 00:09:18,605 Prasanna Malaiyandi: that's what that's what was going on in my head. 189 00:09:18,965 --> 00:09:19,595 W. Curtis Preston: that's funny. 190 00:09:19,775 --> 00:09:22,115 It's not so much snapshots versus backup. 191 00:09:22,145 --> 00:09:24,365 It's replication versus backup. 192 00:09:24,395 --> 00:09:27,365 One big difference between what he was recommending. 193 00:09:27,725 --> 00:09:32,225 So he, he made this point of saying, Um, you know, the whole world based 194 00:09:32,255 --> 00:09:35,315 basically he, if he had his druthers, the entire world would move all 195 00:09:35,315 --> 00:09:36,875 their storage onto NetApp storage. 196 00:09:37,145 --> 00:09:40,085 And then you get basically backup and replication, storage, and primary and 197 00:09:40,085 --> 00:09:41,435 secondary and all that all in one thing. 198 00:09:41,705 --> 00:09:44,765 And he, but he acknowledged that that's only roughly 15% of the world. 199 00:09:45,755 --> 00:09:48,515 So I'm saying, all right, let's talk to the rest of the world. 200 00:09:49,295 --> 00:09:53,875 You're not going to move everything to, to, to net app or you're not, you, 201 00:09:53,875 --> 00:09:57,815 you're not even going to use if you've got a NetApp like architecture and you 202 00:09:57,815 --> 00:10:02,465 could be using snapshots and replication, but you don't, you haven't been able 203 00:10:02,465 --> 00:10:04,505 to convince your boss to do that. 204 00:10:04,535 --> 00:10:06,275 You're going to use backup instead. 205 00:10:06,275 --> 00:10:11,225 So I just want to, before we discuss backup, I just want to point my hands 206 00:10:11,225 --> 00:10:15,695 are really going to really going today before we discuss back up. 207 00:10:15,905 --> 00:10:16,775 I just wanted to say. 208 00:10:17,930 --> 00:10:23,930 What's the difference between a backup and what that does because I would, I would I, 209 00:10:24,500 --> 00:10:29,660 if you replicate snapshots to a secondary array, I would call that a backup. 210 00:10:29,930 --> 00:10:34,910 Some would not, or at least it's a, it's an element of backup, but what I'm 211 00:10:34,910 --> 00:10:41,570 differentiating here is when you make a copy into some other format, right? 212 00:10:41,600 --> 00:10:43,460 It's not all within. 213 00:10:44,615 --> 00:10:50,465 You know, uh, and typically when we say backup, although I do not, I am not 214 00:10:50,465 --> 00:10:56,525 this, um, whatever the opposite of all encompassing, I'm not this pedantic. 215 00:10:56,525 --> 00:10:57,155 If you will, 216 00:10:57,510 --> 00:10:57,720 Prasanna Malaiyandi: Yeah. 217 00:10:58,205 --> 00:11:00,185 W. Curtis Preston: I'm not just talking about tape backup. 218 00:11:00,215 --> 00:11:03,155 I'm not just talking about traditional. 219 00:11:04,205 --> 00:11:11,715 Um, like let's say Veritas NetBackup or, or Dell networker, that style. 220 00:11:12,095 --> 00:11:14,345 CommVault, it's not just that. 221 00:11:14,675 --> 00:11:20,015 I would also consider, you know, our competitors like Rubrik, Cohesity, 222 00:11:20,015 --> 00:11:23,075 Veeam, Clumio, these are all backup. 223 00:11:23,915 --> 00:11:31,565 Um, so, but what I thought we would do is talk about things that we 224 00:11:31,565 --> 00:11:36,275 should absolutely stop doing, right. 225 00:11:36,335 --> 00:11:39,575 That if you're still doing these things, you should really 226 00:11:39,575 --> 00:11:41,175 seriously consider a change. 227 00:11:42,005 --> 00:11:44,255 That was quite the build up quite a, quite a rant. 228 00:11:44,345 --> 00:11:45,665 And I haven't even said anything yet. 229 00:11:46,625 --> 00:11:47,165 What do you think? 230 00:11:47,815 --> 00:11:49,075 Prasanna Malaiyandi: no, I think so. 231 00:11:49,075 --> 00:11:51,805 I just wanted two points of clarification from you. 232 00:11:52,325 --> 00:11:52,895 W. Curtis Preston: Yeah. 233 00:11:53,405 --> 00:11:59,255 Prasanna Malaiyandi: Um, the first is, we're not saying that replication 234 00:11:59,285 --> 00:12:05,315 is evil or snapshots is a place in your environment, depending on your 235 00:12:05,315 --> 00:12:09,995 needs and objectives to have those mechanisms, but it does not truly 236 00:12:09,995 --> 00:12:12,935 replace backup and the sorts of things 237 00:12:12,935 --> 00:12:13,805 that, protects against. 238 00:12:14,315 --> 00:12:15,875 W. Curtis Preston: well, that's a different discussion, 239 00:12:16,345 --> 00:12:18,175 Prasanna Malaiyandi: Yeah, but I just want to be clear though, that 240 00:12:18,455 --> 00:12:18,725 W. Curtis Preston: right? 241 00:12:18,755 --> 00:12:20,615 That's not, that's not the point I'm trying to make. 242 00:12:20,645 --> 00:12:24,155 I'm just trying to say you're using a backup product. 243 00:12:24,575 --> 00:12:27,965 And I want to talk about some of the same concepts that he talked about. 244 00:12:28,460 --> 00:12:32,900 But in the backup context, cause he was living in this world where you have 245 00:12:32,930 --> 00:12:37,880 assued backup and in favor of net, you know, I'm not saying I don't even want 246 00:12:37,880 --> 00:12:41,690 to have the discussion on, I mean, maybe we will on a different episode on 247 00:12:42,080 --> 00:12:45,230 what's the difference between, I just wanted to differentiate between the two. 248 00:12:45,230 --> 00:12:46,550 I'm not saying one is bad or better. 249 00:12:46,550 --> 00:12:51,050 I'm just saying your you're listening to this episode because 250 00:12:51,530 --> 00:12:53,390 you're using a backup product. 251 00:12:54,500 --> 00:12:57,500 And I want to talk about things that you shouldn't be doing anymore. 252 00:12:58,000 --> 00:12:58,320 Prasanna Malaiyandi: Okay. 253 00:12:58,600 --> 00:12:59,395 I'm good with that. 254 00:12:59,845 --> 00:13:01,855 W. Curtis Preston: So you really only asked one question. 255 00:13:02,725 --> 00:13:03,625 You had like three. 256 00:13:04,185 --> 00:13:04,885 Prasanna Malaiyandi: well, I had two. 257 00:13:04,905 --> 00:13:07,935 My second question and maybe we might get to it when we're talking. 258 00:13:08,355 --> 00:13:14,145 You also mentioned that you made the point about it should be in a different format, 259 00:13:16,875 --> 00:13:17,775 W. Curtis Preston: Yes. 260 00:13:17,985 --> 00:13:18,585 Yes. 261 00:13:19,475 --> 00:13:19,985 Prasanna Malaiyandi: And 262 00:13:21,265 --> 00:13:24,385 W. Curtis Preston: In order to fall into this definition, that's all I'm saying. 263 00:13:24,385 --> 00:13:25,665 I'm not, I'm not making. 264 00:13:26,360 --> 00:13:27,800 Well, there is a judgment call. 265 00:13:27,860 --> 00:13:29,120 That is a different discussion. 266 00:13:29,480 --> 00:13:32,060 I'm just trying to just delineate it. 267 00:13:32,135 --> 00:13:32,465 Prasanna Malaiyandi: Yeah. 268 00:13:32,525 --> 00:13:38,105 The only reason I was asking is typically a lot of backup products 269 00:13:38,105 --> 00:13:44,705 that might be doing native format backups don't really modify the format. 270 00:13:45,695 --> 00:13:45,935 And 271 00:13:46,120 --> 00:13:48,820 W. Curtis Preston: don't modify it, but they store it in a different way. 272 00:13:49,390 --> 00:13:49,900 Prasanna Malaiyandi: Okay. 273 00:13:50,550 --> 00:13:50,760 Okay. 274 00:13:50,770 --> 00:13:53,450 So It's but 275 00:13:53,805 --> 00:13:57,135 W. Curtis Preston: not just a replicated, like, again, it's easier to say what, 276 00:13:57,135 --> 00:14:04,485 what, what isn't so NetApp SnapMirror, you know, snapshots and SnapMirror that 277 00:14:04,485 --> 00:14:07,875 doesn't fit this definition because it's the exact same thing on both sides, 278 00:14:08,480 --> 00:14:11,240 Prasanna Malaiyandi: but if someone was scp from one Linux box to 279 00:14:11,240 --> 00:14:14,300 another Linux box and copying the data that would be considered 280 00:14:14,455 --> 00:14:15,685 W. Curtis Preston: That would be something different. 281 00:14:18,140 --> 00:14:21,920 Generally, it will change formats, but not always. 282 00:14:22,430 --> 00:14:22,820 Right. 283 00:14:22,880 --> 00:14:26,270 I know for example, Veeam tends to store data in a way. 284 00:14:26,540 --> 00:14:29,120 And actually I think Rubrik and Cohesity do as well. 285 00:14:29,660 --> 00:14:31,820 They store data in a way that it can be accessed. 286 00:14:31,820 --> 00:14:38,540 But, but it's just about, again, it's about changing the manner 287 00:14:38,540 --> 00:14:39,740 in which the data is stored. 288 00:14:40,295 --> 00:14:40,715 Prasanna Malaiyandi: Okay. 289 00:14:40,955 --> 00:14:41,795 I just looking 290 00:14:41,900 --> 00:14:42,890 W. Curtis Preston: don't have a better way. 291 00:14:43,205 --> 00:14:43,865 Prasanna Malaiyandi: No, that's okay. 292 00:14:44,420 --> 00:14:44,750 W. Curtis Preston: Yeah. 293 00:14:44,810 --> 00:14:45,230 Okay. 294 00:14:46,460 --> 00:14:47,210 So, 295 00:14:47,255 --> 00:14:48,155 Prasanna Malaiyandi: should we be stopped doing 296 00:14:49,300 --> 00:14:52,690 W. Curtis Preston: well, the first thing I'm going to talk about is tape, 297 00:14:53,620 --> 00:14:55,210 and I'm sorry to my tape friends. 298 00:14:55,430 --> 00:14:56,055 Prasanna Malaiyandi: But you're Mr. 299 00:14:56,055 --> 00:14:57,015 Back up? 300 00:14:58,350 --> 00:15:00,360 W. Curtis Preston: Yeah, but see backup doesn't mean tape. 301 00:15:00,870 --> 00:15:02,640 It has for many, many years. 302 00:15:02,880 --> 00:15:09,030 And, and, and, and it's sure meant that to, to, to Ricky on that episode, 303 00:15:09,150 --> 00:15:11,220 but backup doesn't mean tape to me. 304 00:15:11,220 --> 00:15:13,050 And it hasn't for a long time. 305 00:15:13,290 --> 00:15:16,530 I'm not saying do away with tape, but here's what, here's 306 00:15:16,530 --> 00:15:18,240 what we have to do away with. 307 00:15:18,510 --> 00:15:22,020 And w, and again, everything I say has exceptions. 308 00:15:23,030 --> 00:15:23,480 Right. 309 00:15:23,960 --> 00:15:25,730 Everything I say has exceptions. 310 00:15:26,420 --> 00:15:30,260 There are reasons why you might want to continue doing things, but as 311 00:15:30,260 --> 00:15:37,970 a general rule, I don't think that most companies should be backing up 312 00:15:38,180 --> 00:15:41,690 from a server directly to a tape. 313 00:15:42,830 --> 00:15:43,250 Right. 314 00:15:43,550 --> 00:15:50,480 That design hasn't been a good design for 15 years. 315 00:15:51,315 --> 00:15:56,055 Maybe more it's because of that problem that we've talked about before, it's 316 00:15:56,055 --> 00:15:57,705 the tape speed mismatch problem. 317 00:15:57,705 --> 00:16:01,395 The tape wants to go too fast and you want to go, your backup wants to 318 00:16:01,395 --> 00:16:04,305 go really slow and the only way to. 319 00:16:05,885 --> 00:16:09,545 To address that speed mismatch is to do massive levels of 320 00:16:09,545 --> 00:16:13,145 multiplexing, where you're, where you're putting data together. 321 00:16:13,415 --> 00:16:19,615 Um, in order to, to generate a stream of data fast enough to 322 00:16:19,615 --> 00:16:20,865 keep that tape drive happy. 323 00:16:21,165 --> 00:16:21,445 Right. 324 00:16:22,175 --> 00:16:24,035 And we'll talk about multiplexing in a bit. 325 00:16:24,060 --> 00:16:26,400 Prasanna Malaiyandi: Yeah, it's funny as you're saying 326 00:16:26,400 --> 00:16:27,460 that we should stop doing tape. 327 00:16:27,460 --> 00:16:30,690 I was just going back and thinking how many episodes on this podcast 328 00:16:30,690 --> 00:16:32,280 have we actually talked about tape? 329 00:16:32,610 --> 00:16:34,320 And I agree with you though, right? 330 00:16:34,320 --> 00:16:36,270 That there is a use case per tape. 331 00:16:36,900 --> 00:16:39,540 And there is improvements in the technology, but just like you 332 00:16:39,540 --> 00:16:44,640 said, using tape in, in general using tape as your first copy 333 00:16:45,040 --> 00:16:49,180 W. Curtis Preston: As your initial copy is very problematic. 334 00:16:49,720 --> 00:16:53,410 We can have a different discussion as to whether or not it makes a good 335 00:16:53,410 --> 00:16:57,130 secondary copy or a doomsday copy. 336 00:16:57,190 --> 00:16:57,580 Right. 337 00:16:57,610 --> 00:17:00,910 So, Brian, right, Brian came on and Brian made a good. 338 00:17:01,875 --> 00:17:06,975 Case Brian Greenberg and his, his coworker whose name I forgot, but 339 00:17:07,035 --> 00:17:10,415 I've known what's that Cameron. 340 00:17:10,445 --> 00:17:10,895 Thank you. 341 00:17:11,645 --> 00:17:18,275 They, they, they made a good case for tape for, for a copy on tape. 342 00:17:18,305 --> 00:17:19,805 I'm not saying tape is bad. 343 00:17:20,015 --> 00:17:22,655 I'm just saying tape is really, really bad at receiving incremental 344 00:17:22,655 --> 00:17:26,435 backups unless you've spooled them all to some fast disk first. 345 00:17:26,435 --> 00:17:29,165 And then you're just copying all those backups over. 346 00:17:29,675 --> 00:17:30,185 That's what I mean. 347 00:17:31,175 --> 00:17:31,535 Prasanna Malaiyandi: can I. 348 00:17:32,345 --> 00:17:36,435 And I know this will probably fall under that in general, but 349 00:17:36,435 --> 00:17:38,085 here might be an exception case. 350 00:17:38,355 --> 00:17:41,205 Um, would it change? 351 00:17:41,205 --> 00:17:44,805 So if, depending on what you're backing up, for instance, if you're backing 352 00:17:44,805 --> 00:17:49,995 up object store which not many people probably do, but maybe in that instance, 353 00:17:50,085 --> 00:17:51,995 would it make sense to go to tape? 354 00:17:53,665 --> 00:17:56,000 W. Curtis Preston: All of my recommendation is based on. 355 00:17:57,090 --> 00:17:59,250 How lousy incremental backups are to tape. 356 00:17:59,280 --> 00:18:02,340 If you have a scenario where you don't have this problem, I don't care. 357 00:18:02,340 --> 00:18:04,710 I don't have a problem with the reliability of tape. 358 00:18:05,070 --> 00:18:07,140 I have a problem with it's in compatibility with the 359 00:18:07,140 --> 00:18:08,310 way incremental backups run. 360 00:18:09,600 --> 00:18:09,840 Okay. 361 00:18:10,680 --> 00:18:11,100 Prasanna Malaiyandi: Gotcha. 362 00:18:11,520 --> 00:18:12,480 W. Curtis Preston: I I'm even okay. 363 00:18:12,870 --> 00:18:15,240 Maybe for using them for full backups. 364 00:18:15,720 --> 00:18:22,200 If, if you've got a scenario where you can make that happen, many environments, 365 00:18:22,230 --> 00:18:25,800 they just can't even do full backups or just, you know, they've got a 10 366 00:18:25,800 --> 00:18:30,285 exabyte storage system, how are you going to do a full backup on that? 367 00:18:30,345 --> 00:18:30,615 Right? 368 00:18:30,615 --> 00:18:31,665 You you're just not. 369 00:18:32,595 --> 00:18:36,885 And so, but if you can, if you can do a full backup to tape, that's 370 00:18:36,915 --> 00:18:40,305 possibly a way to use tape in your environment, but generally speaking, 371 00:18:40,485 --> 00:18:41,835 but then it complicates things. 372 00:18:41,835 --> 00:18:44,055 We're going to do full backups to tape, but incrementals to desk. 373 00:18:44,055 --> 00:18:44,415 And we're, you 374 00:18:44,415 --> 00:18:44,625 know, 375 00:18:45,270 --> 00:18:45,630 Prasanna Malaiyandi: Managing. 376 00:18:45,675 --> 00:18:47,865 W. Curtis Preston: I'd much rather you just simplify things 377 00:18:47,865 --> 00:18:49,665 and use, use disk for everything. 378 00:18:50,205 --> 00:18:53,715 Um, uh, although we're gonna, we're going to come back to disk as well. 379 00:18:53,865 --> 00:18:54,165 Prasanna Malaiyandi: Yeah. 380 00:18:55,305 --> 00:18:55,575 So. 381 00:18:55,575 --> 00:18:58,455 since you touched on tape and continuing along that train 382 00:18:58,455 --> 00:18:59,685 of thought, you brought up 383 00:18:59,685 --> 00:19:00,315 multiplexing. 384 00:19:01,010 --> 00:19:03,050 W. Curtis Preston: Yeah, multiplexing. 385 00:19:03,495 --> 00:19:03,855 Prasanna Malaiyandi: Yeah. 386 00:19:04,095 --> 00:19:07,815 And I know in the previous tape conversations, right. 387 00:19:07,905 --> 00:19:13,755 It's kind of the only way to actually keep tape happy, but like, you've like, we've 388 00:19:13,755 --> 00:19:18,840 also talked about in previous podcasts, It's great for doing backups, but when 389 00:19:18,840 --> 00:19:23,520 you have to do restores, that's where things just go really, really, really 390 00:19:23,520 --> 00:19:23,820 bad. 391 00:19:24,580 --> 00:19:27,280 W. Curtis Preston: Yeah, because basically you have to read the whole 392 00:19:27,280 --> 00:19:31,600 tape and throw away 90% of it, or, or maybe even 95 or more percent of it. 393 00:19:31,930 --> 00:19:36,850 Now I know of at least one product, their name is escaping me. 394 00:19:36,850 --> 00:19:39,280 It might be, no, I don't want to say that. 395 00:19:39,340 --> 00:19:40,420 I don't want to say the name. 396 00:19:40,510 --> 00:19:43,150 See, you heard that notification share the notification. 397 00:19:43,210 --> 00:19:43,740 I have do not. 398 00:19:45,705 --> 00:19:49,155 Oh, you, you probably did because you didn't because I got a unidirectional mic, 399 00:19:49,185 --> 00:19:51,225 but slack just gave me a notification. 400 00:19:52,215 --> 00:19:55,365 So what I have do not disturb for everything on. 401 00:19:55,365 --> 00:19:57,795 So why is slacking beeping on my computer anyway? 402 00:19:59,265 --> 00:20:01,065 I'm sorry. 403 00:20:01,545 --> 00:20:03,285 I digress, but that's what happens. 404 00:20:03,285 --> 00:20:04,635 Cause I got a squirrel. 405 00:20:05,865 --> 00:20:06,915 What were we talking about? 406 00:20:07,200 --> 00:20:07,920 Prasanna Malaiyandi: Multiplexing 407 00:20:08,865 --> 00:20:09,135 W. Curtis Preston: Oh yeah. 408 00:20:09,240 --> 00:20:09,480 Prasanna Malaiyandi: there's 409 00:20:09,480 --> 00:20:10,110 a tape vendor. 410 00:20:10,485 --> 00:20:12,735 W. Curtis Preston: Multiplexing before we talk about multiplexer. 411 00:20:12,745 --> 00:20:14,715 Let me just round out about tape again. 412 00:20:14,805 --> 00:20:19,815 I think tape there's an argument to be made for using it as a secondary 413 00:20:19,815 --> 00:20:23,895 copy as a way for creating a backup to send off site mostly in smaller 414 00:20:23,895 --> 00:20:28,245 environments though, because when they get big using tape as a DR mechanism, 415 00:20:28,245 --> 00:20:29,595 I think is really problematic, 416 00:20:29,950 --> 00:20:30,130 Prasanna Malaiyandi: Yeah. 417 00:20:30,705 --> 00:20:35,265 W. Curtis Preston: um, and also using it for long-term storage, there's a 418 00:20:35,265 --> 00:20:36,735 really, really good argument there. 419 00:20:36,800 --> 00:20:37,070 Prasanna Malaiyandi: Yep. 420 00:20:37,220 --> 00:20:38,690 And archive is not backup. 421 00:20:39,725 --> 00:20:40,265 W. Curtis Preston: Correct. 422 00:20:42,195 --> 00:20:44,555 Is it, did you have another comment you want to make? 423 00:20:44,560 --> 00:20:46,330 Prasanna Malaiyandi: No, I was just going to talk about, yeah. 424 00:20:46,360 --> 00:20:48,760 Tape makes perfect sense in an archiving sense as 425 00:20:48,760 --> 00:20:49,240 well. 426 00:20:49,365 --> 00:20:49,665 W. Curtis Preston: Yeah. 427 00:20:50,625 --> 00:20:50,925 Active 428 00:20:51,040 --> 00:20:51,430 Prasanna Malaiyandi: different than 429 00:20:51,430 --> 00:20:51,820 backup. 430 00:20:52,000 --> 00:20:52,210 Yep. 431 00:20:52,665 --> 00:20:53,355 W. Curtis Preston: Absolutely. 432 00:20:53,715 --> 00:20:58,155 Um, and so, yeah, so let's go back to multiplexing .Multiplexing. 433 00:20:58,735 --> 00:21:00,195 It's a bit like, have you ever heard that? 434 00:21:00,195 --> 00:21:04,995 Um, I dunno if this is one of those like fake quotes, but uh, I think. 435 00:21:06,380 --> 00:21:11,210 Th my memory is that it's Winston Churchill that allegedly said democracy 436 00:21:11,210 --> 00:21:12,500 is the worst form of government. 437 00:21:12,500 --> 00:21:13,760 It's just better than the other ones. 438 00:21:14,360 --> 00:21:16,190 Um, something like that. 439 00:21:16,640 --> 00:21:20,180 And multiplexing is the worst backup technology ever. 440 00:21:21,090 --> 00:21:24,630 It just was really helpful in solving a particular problem. 441 00:21:24,930 --> 00:21:25,140 Right? 442 00:21:25,140 --> 00:21:28,470 It was, it was a, it was a necessary evil because tape drives were getting 443 00:21:28,470 --> 00:21:31,980 faster and faster, faster, and we didn't, we didn't yet have multiplexing. 444 00:21:32,400 --> 00:21:32,790 I'm sorry. 445 00:21:32,790 --> 00:21:35,250 We didn't yet have deduplication, 446 00:21:35,370 --> 00:21:38,320 which, which is what allowed us to use disk. 447 00:21:39,180 --> 00:21:44,630 So stop using tape and multiplexing and use disk and deduplication. 448 00:21:44,820 --> 00:21:47,160 And you know, that solves that problem. 449 00:21:48,170 --> 00:21:49,665 Prasanna Malaiyandi: It solves on problem 450 00:21:50,195 --> 00:21:51,085 W. Curtis Preston: solves one problem. 451 00:21:51,265 --> 00:21:51,625 Yeah. 452 00:21:51,895 --> 00:21:52,075 Yeah. 453 00:21:52,075 --> 00:21:53,035 We're gonna, yeah. 454 00:21:53,035 --> 00:21:54,265 disk creates another problem. 455 00:21:54,265 --> 00:21:55,135 We'll get to that in a minute. 456 00:21:55,705 --> 00:21:56,365 Prasanna Malaiyandi: But here's 457 00:21:56,425 --> 00:21:57,145 W. Curtis Preston: The next 458 00:21:57,145 --> 00:21:57,655 thing, 459 00:21:58,075 --> 00:21:58,945 Prasanna Malaiyandi: before we jump. 460 00:22:00,355 --> 00:22:08,005 So even with the multiplexing and tape backup, are there any products other 461 00:22:08,005 --> 00:22:11,515 than tape, are there other displays products that use multiplexing? 462 00:22:12,530 --> 00:22:14,300 W. Curtis Preston: No, because there's absolutely no reason 463 00:22:14,840 --> 00:22:16,250 to use multiplexing on desk. 464 00:22:17,120 --> 00:22:20,600 And by the way, if, if I'm wrong, uh, then that's the dumbest product 465 00:22:20,930 --> 00:22:22,490 in the history of dumb products. 466 00:22:23,540 --> 00:22:23,960 Right? 467 00:22:23,990 --> 00:22:28,430 The, oh, what I was saying before I got interrupted by the slack thing is 468 00:22:28,430 --> 00:22:32,090 there is at least one backup product that does multiplexing differently. 469 00:22:32,690 --> 00:22:34,870 It sort of spools. 470 00:22:35,385 --> 00:22:38,775 I don't know if it does it in Ram or if it does it in flash or disk or 471 00:22:38,775 --> 00:22:43,005 whatever, but it does multiplexing in such a way that they put big 472 00:22:43,005 --> 00:22:45,825 segments on the tape contiguously. 473 00:22:46,245 --> 00:22:51,615 And so they're able to read, seek, read, seek rather than what most backup 474 00:22:51,645 --> 00:22:54,525 products do, which is read everything and just throw away everything. 475 00:22:54,525 --> 00:22:55,245 It doesn't need. 476 00:22:55,945 --> 00:22:59,875 There's that, but just tape and multiplexing, that's just gotta stop. 477 00:22:59,905 --> 00:23:02,875 Like if you're doing tape and multiplexing, you really have to stop. 478 00:23:02,905 --> 00:23:04,705 It is a horrible, horrible way. 479 00:23:05,275 --> 00:23:07,675 And if you're like, well, how am I going to make my tape drives happy? 480 00:23:07,735 --> 00:23:08,845 Stop using them. 481 00:23:09,145 --> 00:23:11,005 That's how right. 482 00:23:11,005 --> 00:23:12,505 Use them for secondary copies. 483 00:23:12,505 --> 00:23:15,505 Use them to create a copy to hand, to hand to iron mountain. 484 00:23:17,725 --> 00:23:18,115 All right. 485 00:23:19,255 --> 00:23:20,755 So the next 486 00:23:20,755 --> 00:23:21,145 one. 487 00:23:21,240 --> 00:23:23,190 Prasanna Malaiyandi: Have you caught in that all out of your system? 488 00:23:23,900 --> 00:23:27,380 W. Curtis Preston: Yeah, the tape thing, the next is, and this is 489 00:23:27,380 --> 00:23:32,090 what we talked a lot about with Ricky and that is full backups. 490 00:23:32,090 --> 00:23:36,020 And by that, I mean, repeated full backups, occasional, regular, full 491 00:23:36,020 --> 00:23:39,890 backups once a week, once a month, once a quarter, once a year, I don't 492 00:23:39,890 --> 00:23:41,000 care how often you're doing them. 493 00:23:41,150 --> 00:23:42,020 They're dumb. 494 00:23:42,470 --> 00:23:45,920 And that's where Ricky and I, a hundred percent agree the idea 495 00:23:46,610 --> 00:23:50,900 of moving everything from a to B. 496 00:23:51,790 --> 00:23:55,990 'cause that's why we've always done it is the dumbest idea. 497 00:23:56,830 --> 00:24:01,630 Again, it was a great idea back in the day, because we were using tape 498 00:24:02,230 --> 00:24:07,000 and if you're using tape, you do need to do the occasional, full backup or 499 00:24:07,000 --> 00:24:09,070 else the restore can take forever. 500 00:24:10,090 --> 00:24:13,030 And you know, there are backup products that have this problem. 501 00:24:13,060 --> 00:24:16,720 I won't name them specifically, although Ricky didn't have a problem. 502 00:24:19,570 --> 00:24:23,080 Prasanna Malaiyandi: So I agree with you because it doesn't make sense, 503 00:24:23,080 --> 00:24:26,740 especially when you also have deduplicated disk where you're going to throw away 504 00:24:26,740 --> 00:24:29,050 90% of it, the full backup anyway. 505 00:24:29,380 --> 00:24:29,650 Right. 506 00:24:29,650 --> 00:24:33,580 Why even occupy the additional network, right. 507 00:24:33,580 --> 00:24:35,950 That you need in order to be able to move all the data, plus all 508 00:24:35,950 --> 00:24:37,300 of the compute and the storage 509 00:24:37,895 --> 00:24:40,195 W. Curtis Preston: well, the compute and everything on, on the, on the, 510 00:24:40,355 --> 00:24:41,645 on the side that you're doing it. 511 00:24:41,645 --> 00:24:41,975 Yeah. 512 00:24:42,060 --> 00:24:42,240 Prasanna Malaiyandi: Yep. 513 00:24:42,475 --> 00:24:42,745 W. Curtis Preston: Yeah. 514 00:24:43,260 --> 00:24:47,430 Prasanna Malaiyandi: And, but I do have a question 515 00:24:47,520 --> 00:24:48,030 whether it's. 516 00:24:48,055 --> 00:24:50,155 W. Curtis Preston: You know that whenever you say a bunch of stuff, and 517 00:24:50,155 --> 00:24:53,545 then you say, but you just dismissed everything you said before the but, right? 518 00:24:54,210 --> 00:24:57,120 Prasanna Malaiyandi: So I totally agree with, we should never do full 519 00:24:57,120 --> 00:24:57,750 backups, 520 00:24:58,395 --> 00:24:58,785 W. Curtis Preston: Okay. 521 00:25:00,195 --> 00:25:00,675 So 522 00:25:02,295 --> 00:25:08,175 you have a concern, you know, you pause, you just pause, you agree. 523 00:25:08,235 --> 00:25:10,185 And then you say, I have a concern. 524 00:25:10,875 --> 00:25:13,595 So you, that way you don't negate everything that you said before. 525 00:25:15,100 --> 00:25:15,615 Prasanna Malaiyandi: I have a 526 00:25:15,685 --> 00:25:17,065 W. Curtis Preston: I'm really good at arguing. 527 00:25:17,155 --> 00:25:20,995 I'm just saying, this is, this is a skill that I picked up when you 528 00:25:20,995 --> 00:25:22,435 say, blah, blah, blah, blah, blah. 529 00:25:22,495 --> 00:25:25,855 But, or however you just dismissed everything that you just said. 530 00:25:25,855 --> 00:25:31,045 So now I've learned to say, I agree with what you said about tape, not 531 00:25:31,135 --> 00:25:36,345 or about not doing full backups, the concern I have C you on that 532 00:25:36,650 --> 00:25:37,760 Prasanna Malaiyandi: Yeah, much better. 533 00:25:37,810 --> 00:25:38,070 Yeah. 534 00:25:38,245 --> 00:25:39,415 W. Curtis Preston: I'm teaching you how to argue. 535 00:25:41,580 --> 00:25:44,970 Prasanna Malaiyandi: The concern I have is for. 536 00:25:46,650 --> 00:25:48,870 Applications and certain workloads. 537 00:25:48,900 --> 00:25:52,470 I think you can't get away from never doing full backups. 538 00:25:53,115 --> 00:25:56,115 W. Curtis Preston: That is a valid statement. 539 00:25:56,595 --> 00:26:08,005 In fact, the product that Ricky mentioned, they only do forever 540 00:26:08,005 --> 00:26:09,385 incrementals on file systems. 541 00:26:09,385 --> 00:26:10,345 They don't do. 542 00:26:11,290 --> 00:26:12,940 Forever incremental on databases. 543 00:26:13,360 --> 00:26:19,810 If your database basically doesn't know how to not do a full backup, you 544 00:26:19,810 --> 00:26:21,580 have to do the occasional full backup. 545 00:26:22,660 --> 00:26:24,550 There are ways around. 546 00:26:24,550 --> 00:26:30,640 So like for example, the most popular of those databases would be Oracle and 547 00:26:30,940 --> 00:26:35,260 there is a way to do forever incremental with Oracle and hardly anybody uses it. 548 00:26:35,905 --> 00:26:36,205 Prasanna Malaiyandi: Yep. 549 00:26:36,385 --> 00:26:39,415 W. Curtis Preston: The idea of the image copy and, 550 00:26:39,895 --> 00:26:41,155 Prasanna Malaiyandi: Oracle incremental merge. 551 00:26:41,785 --> 00:26:43,455 W. Curtis Preston: Right, Oracle incremental margin. 552 00:26:43,455 --> 00:26:45,505 We should cover that on podcasts, but that on the list, 553 00:26:45,565 --> 00:26:46,855 we'll cover that on a podcast. 554 00:26:47,215 --> 00:26:49,225 I think it's a great way to do you. 555 00:26:49,345 --> 00:26:51,385 You end up with a copy of Oracle. 556 00:26:51,385 --> 00:26:57,385 You end up with multiple copies of Oracle on disc, that multiple versions, if you 557 00:26:57,385 --> 00:27:02,335 will, of Oracle on disk and each of them only takes up the incremental new blocks. 558 00:27:03,520 --> 00:27:07,540 So it's a really good way to do incremental forever. 559 00:27:10,750 --> 00:27:12,850 You don't like incremental forever with what don't you 560 00:27:12,955 --> 00:27:17,395 Prasanna Malaiyandi: I don't, there are some downsides to using Oracle incremental 561 00:27:17,395 --> 00:27:21,235 merge, especially when it comes to integrating with other applications. 562 00:27:21,745 --> 00:27:25,435 Um, because when it comes to integrating with your native backup 563 00:27:25,435 --> 00:27:32,095 applications, uh, because it's not fully supporting of the streaming. 564 00:27:33,505 --> 00:27:36,745 Transferred the SBT library that Oracle has, right. 565 00:27:36,745 --> 00:27:38,695 It's completely managed separately and all the rest. 566 00:27:38,695 --> 00:27:41,335 So it makes it a little bit more complex to manage. 567 00:27:41,575 --> 00:27:45,625 I agree with you that there are a lot of great use cases with Oracle incremental 568 00:27:45,625 --> 00:27:50,245 merge and doing backups in that format and being able to just, get the incrementals. 569 00:27:50,785 --> 00:27:53,575 Um, and then there is also a second downside of Oracle incremental 570 00:27:53,575 --> 00:27:57,055 merge, which we can talk about in a separate podcast, which is you 571 00:27:57,055 --> 00:27:59,605 actually have to read the data back in order to apply the incremental 572 00:28:00,895 --> 00:28:01,255 W. Curtis Preston: right. 573 00:28:02,485 --> 00:28:02,875 Prasanna Malaiyandi: But we can 574 00:28:02,875 --> 00:28:03,475 say that for another 575 00:28:03,895 --> 00:28:05,815 W. Curtis Preston: there are, yeah, there are exceptions, dare 576 00:28:05,855 --> 00:28:06,805 exceptions, just like everything. 577 00:28:06,805 --> 00:28:10,045 There are exceptions to the rule and obviously if your application forces you 578 00:28:10,045 --> 00:28:11,605 to do full, so then you got to do fulls. 579 00:28:11,935 --> 00:28:15,415 What I'm talking about is using a backup product that it 580 00:28:15,415 --> 00:28:16,615 doesn't know how to not do it. 581 00:28:16,995 --> 00:28:17,985 Prasanna Malaiyandi: Like VMs, right? 582 00:28:18,015 --> 00:28:18,435 VMs. 583 00:28:18,435 --> 00:28:20,505 You should never be doing full backup daily. 584 00:28:20,535 --> 00:28:20,775 Right? 585 00:28:20,775 --> 00:28:21,585 That's ridiculous. 586 00:28:22,560 --> 00:28:22,860 W. Curtis Preston: Right. 587 00:28:23,700 --> 00:28:29,130 And if the, and there are some products that say, well, we don't force a full, 588 00:28:29,130 --> 00:28:34,290 we just do a sink, a synthetic full, honestly, that's not much better. 589 00:28:35,185 --> 00:28:39,055 It's it's a little better, but it's the same thing. 590 00:28:39,055 --> 00:28:43,795 Your backup product is not yet in the 21st century, your backup product 591 00:28:43,915 --> 00:28:46,915 still thinks it needs a full backup. 592 00:28:47,215 --> 00:28:49,855 And so you're going to create a full backup and by the way, synthetic 593 00:28:49,855 --> 00:28:52,435 fulls have limitations themselves. 594 00:28:52,435 --> 00:28:52,735 Right? 595 00:28:54,265 --> 00:28:57,175 I don't want to go down that route, but maybe we add that to the list. 596 00:28:57,205 --> 00:28:59,695 You know, we should talk about synthetic foals versus real fulls. 597 00:29:01,030 --> 00:29:04,480 But, yeah, I'm just saying in general, if you're still using a backup product 598 00:29:04,480 --> 00:29:07,840 that forces you to do an occasional full, however, often that occasional 599 00:29:07,840 --> 00:29:10,870 and whether or not it's synthetic, I'm just saying that should just stop full 600 00:29:10,870 --> 00:29:15,040 backups unless when you're forced to do so, should be a thing of the past. 601 00:29:16,270 --> 00:29:16,810 It's all I'm saying. 602 00:29:17,680 --> 00:29:17,890 Prasanna Malaiyandi: Oh, 603 00:29:18,760 --> 00:29:20,410 W. Curtis Preston: And re huh. 604 00:29:21,070 --> 00:29:21,880 Prasanna Malaiyandi: I agree with you. 605 00:29:22,295 --> 00:29:24,095 W. Curtis Preston: Okay right after that. 606 00:29:24,155 --> 00:29:26,795 And again, unless you don't have any other choice 607 00:29:27,650 --> 00:29:27,830 Prasanna Malaiyandi: Yeah. 608 00:29:28,145 --> 00:29:31,295 W. Curtis Preston: is full file, incremental backups. 609 00:29:32,015 --> 00:29:33,215 What do I mean by that? 610 00:29:34,235 --> 00:29:41,015 There is, there are so many products that, well, there are products that. 611 00:29:43,075 --> 00:29:46,435 Can do block level incremental backup on those files. 612 00:29:47,275 --> 00:29:52,135 The idea of backing up an entire file, just because one byte of that 613 00:29:52,135 --> 00:29:57,025 file has changed is another dumb idea whose time should go away. 614 00:29:57,985 --> 00:30:02,305 And if your backup product doesn't know how to do block level 615 00:30:02,305 --> 00:30:07,225 incremental backups or source side deduplication, so that you're, you 616 00:30:07,225 --> 00:30:08,515 know, you're doing it that way. 617 00:30:08,935 --> 00:30:10,855 And by the way, replication is a way to do it. 618 00:30:12,395 --> 00:30:16,985 So if your product doesn't know how to do that, and you're still backing up the, 619 00:30:17,055 --> 00:30:20,555 it's an order of magnitude difference in the size of an incremental backup. 620 00:30:20,945 --> 00:30:25,465 You're looking at 10 to 15% versus 1% of the, of the size. 621 00:30:25,485 --> 00:30:29,345 Like just, I mean, that's a total made up number, but it's based on my experience. 622 00:30:29,525 --> 00:30:29,855 Okay. 623 00:30:30,815 --> 00:30:34,085 Um, what do you think about that one full file incrementals? 624 00:30:34,795 --> 00:30:37,825 Prasanna Malaiyandi: And I think just to put more context for users. 625 00:30:38,950 --> 00:30:42,640 Don't think of the case where you have a word document and a byte changes. 626 00:30:42,940 --> 00:30:43,180 Right? 627 00:30:43,180 --> 00:30:48,130 Think about also large files that might be in your system, that you are constantly 628 00:30:48,130 --> 00:30:50,050 backing up day after, day after day. 629 00:30:50,590 --> 00:30:50,980 Right. 630 00:30:51,040 --> 00:30:54,520 And if those, if you make a small change to it, that causes the 631 00:30:54,520 --> 00:30:56,410 entire file to be backed up again, 632 00:30:56,950 --> 00:30:57,250 right? 633 00:30:57,410 --> 00:30:58,760 W. Curtis Preston: PST files coming to mind. 634 00:30:59,240 --> 00:30:59,540 Prasanna Malaiyandi: Yep. 635 00:31:00,245 --> 00:31:03,305 W. Curtis Preston: And similar files like that large files that 636 00:31:03,725 --> 00:31:05,435 get a little bit of change each day. 637 00:31:05,885 --> 00:31:08,795 And if you have a full file incremental, backup you back up the entire thing. 638 00:31:09,610 --> 00:31:09,880 Prasanna Malaiyandi: Yep. 639 00:31:10,720 --> 00:31:11,950 And I agree with you. 640 00:31:12,010 --> 00:31:15,190 There's no reason that you should be copying that entire file and backing 641 00:31:15,190 --> 00:31:17,350 up that entire file every single time. 642 00:31:18,570 --> 00:31:18,870 W. Curtis Preston: Yeah. 643 00:31:20,340 --> 00:31:20,580 All right. 644 00:31:20,580 --> 00:31:24,510 Let's the next, one's going to harken back to the earlier discussion 645 00:31:24,510 --> 00:31:27,810 because I was like, I want you to stop storing your backups on tape. 646 00:31:28,950 --> 00:31:29,280 Now. 647 00:31:29,280 --> 00:31:33,240 I want you to stop storing your backups on disk by that. 648 00:31:33,240 --> 00:31:33,900 I mean, 649 00:31:33,900 --> 00:31:36,060 Prasanna Malaiyandi: Turn is Curtis. 650 00:31:37,980 --> 00:31:43,475 W. Curtis Preston: I mean, Directly accessible disc, an SMB 651 00:31:43,475 --> 00:31:50,105 Mount, or an NFS Mount, a windows server with some DAS, right. 652 00:31:50,105 --> 00:31:51,545 That's direct attached storage. 653 00:31:51,665 --> 00:31:54,365 So just regular old JBOD or a disk array or whatever. 654 00:31:55,465 --> 00:32:01,505 An NFS mounted data domain or quantum or whatever, whatever box you're using. 655 00:32:02,645 --> 00:32:07,295 If it's, if the backups are directly accessible in the operating system 656 00:32:07,415 --> 00:32:12,335 of the backup server, this is what I, what we have to stop. 657 00:32:13,595 --> 00:32:15,365 And, and why is that? 658 00:32:16,390 --> 00:32:19,060 Prasanna Malaiyandi: well, because it now becomes easily accessible for 659 00:32:19,060 --> 00:32:23,290 ransomware and other people to go blow away, encrypt your backups, and now 660 00:32:23,290 --> 00:32:24,520 you can't restore your environment. 661 00:32:25,395 --> 00:32:25,605 W. Curtis Preston: Yeah. 662 00:32:27,885 --> 00:32:30,495 Prasanna Malaiyandi: Just one more point on that. 663 00:32:30,765 --> 00:32:32,985 It's uh, shoot. 664 00:32:32,985 --> 00:32:33,915 I lost my train of thought. 665 00:32:33,945 --> 00:32:35,505 Oh, you said backup server. 666 00:32:35,505 --> 00:32:40,425 I think you could also be from any client as well accessing that server. 667 00:32:41,135 --> 00:32:44,765 W. Curtis Preston: Well, any, any server that has direct 668 00:32:45,085 --> 00:32:45,415 Prasanna Malaiyandi: Yeah. 669 00:32:46,025 --> 00:32:49,025 W. Curtis Preston: read-write access to the backups, you are right? 670 00:32:49,025 --> 00:32:55,145 That there are some backup configurations where the backup client writes directly 671 00:32:55,145 --> 00:32:58,445 to the storage, not via like a protocol. 672 00:32:58,565 --> 00:33:01,355 Well, unless we're talking about NFS or SMB is the protocol. 673 00:33:02,195 --> 00:33:04,895 If there, if they can, if they have read, write access directly 674 00:33:04,895 --> 00:33:07,445 to the backups and yes, that's exactly what I'm talking about. 675 00:33:07,910 --> 00:33:10,160 Prasanna Malaiyandi: Especially with all these network protocols, 676 00:33:10,490 --> 00:33:13,100 most backup vendors, right. 677 00:33:13,670 --> 00:33:17,300 Have a proprietary protocol or a proprietary connection for actually 678 00:33:17,300 --> 00:33:18,800 writing the data to the storage. 679 00:33:19,160 --> 00:33:22,130 So you really shouldn't be using open protocols. 680 00:33:23,080 --> 00:33:29,265 Or I want to, I wonder if we could even extend it and say, even if you are using 681 00:33:29,265 --> 00:33:33,615 something that is open, you probably want a really, really, really good way 682 00:33:33,615 --> 00:33:36,765 to secure it and limit the attack scope. 683 00:33:38,675 --> 00:33:39,575 It's not ideal. 684 00:33:40,520 --> 00:33:41,870 W. Curtis Preston: that's a great point. 685 00:33:42,230 --> 00:33:44,210 And because. 686 00:33:45,285 --> 00:33:48,825 If the backup server, I'm sorry. 687 00:33:48,915 --> 00:33:54,975 If the backup software administrative accounts are compromised, it doesn't 688 00:33:54,975 --> 00:33:59,535 matter how you wrote the backups on that disc, whether you used a protocol or not, 689 00:34:00,105 --> 00:34:02,445 you can go in and delete those backups. 690 00:34:02,730 --> 00:34:03,030 Prasanna Malaiyandi: Yep. 691 00:34:04,285 --> 00:34:09,805 W. Curtis Preston: However, the exception to that would be if you've 692 00:34:09,805 --> 00:34:13,165 written it to storage, that is truly. 693 00:34:14,130 --> 00:34:14,790 Prasanna Malaiyandi: Immutable 694 00:34:15,240 --> 00:34:15,780 W. Curtis Preston: Immutable. 695 00:34:16,800 --> 00:34:25,260 So if you, uh, if you're unable to delete the backups, even if you have 696 00:34:26,550 --> 00:34:29,670 administrative access, then that would, that would solve this problem. 697 00:34:29,670 --> 00:34:29,910 Right. 698 00:34:30,090 --> 00:34:32,640 But this is, this is the thing you have to look at now, look at the 699 00:34:32,640 --> 00:34:35,970 attack surface that you have, look at how you're writing backups, find 700 00:34:35,970 --> 00:34:39,510 out if there is a more secure way to store your disk-based backups. 701 00:34:40,050 --> 00:34:42,420 In most cases, the answer is almost always yes. 702 00:34:42,955 --> 00:34:43,285 Prasanna Malaiyandi: Yeah. 703 00:34:43,555 --> 00:34:46,945 And so this goes back to one of those exception cases, right? 704 00:34:46,945 --> 00:34:50,725 So you have to do your due diligence to ensure that everything is protected 705 00:34:51,025 --> 00:34:53,725 and secured in the right way before sort of saying, yeah, I'm just going 706 00:34:53,725 --> 00:34:59,705 to write all my backups via NFS or SMB. 707 00:35:01,175 --> 00:35:02,870 W. Curtis Preston: Um, and, and. 708 00:35:04,475 --> 00:35:08,945 I don't have a problem with like using disk as sort of a caching mechanism. 709 00:35:09,485 --> 00:35:09,815 Right. 710 00:35:09,815 --> 00:35:10,925 Especially if you're using tape. 711 00:35:11,765 --> 00:35:15,135 So it's not that writing backups to directly accessible disk 712 00:35:15,155 --> 00:35:17,105 is necessarily an evil thing. 713 00:35:17,105 --> 00:35:20,225 It's just that if that's the only way you're writing backups, it's 714 00:35:20,225 --> 00:35:21,755 a really, really insecure way. 715 00:35:22,265 --> 00:35:24,875 If it's one of your many copies, I don't have a problem. 716 00:35:25,770 --> 00:35:30,240 But, uh, if, uh, unless you have another option, if you have another 717 00:35:30,240 --> 00:35:33,300 option, a more secure way than use that way, that's all I'm saying. 718 00:35:34,190 --> 00:35:36,440 Prasanna Malaiyandi: And there are also other benefits for some of these 719 00:35:36,440 --> 00:35:41,540 protocols, because it allows things like source-side deduplication and helping 720 00:35:41,540 --> 00:35:43,840 you efficient or save bandwidth, right. 721 00:35:43,910 --> 00:35:46,500 By not sending all the data and getting tossed on the storage side. 722 00:35:47,860 --> 00:35:50,530 W. Curtis Preston: And, and, uh, you know, I'll, I'll put this out. 723 00:35:51,645 --> 00:35:53,055 That yes. 724 00:35:53,085 --> 00:35:58,125 The way that Druva stores, the backups meets this definition, meaning that 725 00:35:58,605 --> 00:36:03,075 it never stores the backups in a, in a way that, you know, we do have a cloud 726 00:36:03,075 --> 00:36:05,655 cache that you can optionally use. 727 00:36:05,655 --> 00:36:08,565 It's not required, but you can, in that case, you may have a 728 00:36:08,565 --> 00:36:13,455 local copy of your data, um, that is also stored to the cloud, but 729 00:36:13,455 --> 00:36:18,325 the main working reference copy. 730 00:36:18,325 --> 00:36:19,135 If you will. 731 00:36:19,405 --> 00:36:24,235 The, the copy of record is stored in the cloud and it's stored in object 732 00:36:24,235 --> 00:36:28,465 storage and is stored in an account that you don't have any access to it's as 733 00:36:28,525 --> 00:36:34,135 it's as removed as it could be from you without actually putting it on a tape 734 00:36:34,375 --> 00:36:35,965 and then handing it to a man in a van. 735 00:36:37,075 --> 00:36:38,905 So we're not the only ones that do that. 736 00:36:38,905 --> 00:36:41,965 We're not, and that's not the only way to do that, but I'm just saying. 737 00:36:42,930 --> 00:36:46,950 That is a way to solve this problem is to just not use disk. 738 00:36:47,060 --> 00:36:50,550 Not use regular disk, uh, regular backup servers, uh, at all. 739 00:36:51,190 --> 00:36:53,680 Prasanna Malaiyandi: You know, you should get t-shirts printed with that, 740 00:36:54,210 --> 00:36:54,660 W. Curtis Preston: What 741 00:36:55,760 --> 00:36:56,920 Prasanna Malaiyandi: not the man in the van 742 00:36:57,350 --> 00:36:58,280 or way. 743 00:36:59,590 --> 00:37:00,550 W. Curtis Preston: not the man in the van. 744 00:37:00,550 --> 00:37:00,970 I like it. 745 00:37:01,720 --> 00:37:06,790 Um, so the last thing I wanted to talk about, and this is 746 00:37:06,790 --> 00:37:08,200 another age old practice. 747 00:37:09,700 --> 00:37:15,880 Is an, and it's an age old practice with even more experienced than, than some of 748 00:37:15,880 --> 00:37:17,050 the other ones that we've talked about. 749 00:37:17,050 --> 00:37:23,800 And that is this idea of waiting for the disaster or the attack 750 00:37:24,100 --> 00:37:26,020 to start your disaster recovery. 751 00:37:27,990 --> 00:37:34,380 And by that, I mean, when I think back to my early days, RDR our DR as I. 752 00:37:35,220 --> 00:37:40,200 As I make quotes, DR plan was a box and a tape and a set of 753 00:37:40,200 --> 00:37:47,940 instructions, and we never had to fire it in anger, as I say, right. 754 00:37:47,940 --> 00:37:52,590 We never had to actually use it because we got a flood or a 755 00:37:53,190 --> 00:37:58,020 hurricane or a terrorist attack, or obviously not a ransomware attack. 756 00:37:59,510 --> 00:38:03,740 But I know that if we had actually done that, we would have been down 757 00:38:03,740 --> 00:38:09,980 for ages because I know that it took us an entire weekend to restore the, 758 00:38:10,640 --> 00:38:14,720 the, just the one or two servers that we would do on a, on a DR test. 759 00:38:15,890 --> 00:38:22,670 And so what I'm saying is that it used to be that. 760 00:38:25,155 --> 00:38:30,315 Only the high end companies, only the financial trading firms or whatever, 761 00:38:30,345 --> 00:38:35,945 where they could put downtime measured in millions of dollars an hour. 762 00:38:36,195 --> 00:38:36,405 Right. 763 00:38:36,405 --> 00:38:38,235 We lose a million dollars an hour. 764 00:38:38,235 --> 00:38:38,805 If we're down. 765 00:38:39,015 --> 00:38:44,955 Those were the folks that had the super HA stuff and the hot, the hot standby site 766 00:38:44,955 --> 00:38:46,545 and all that stuff that we used to have. 767 00:38:48,025 --> 00:38:51,015 What I'm saying is that's no longer the case. 768 00:38:51,615 --> 00:38:58,395 It's no, there are too many options for you to be able to have a disaster recovery 769 00:38:58,395 --> 00:39:05,115 plan that is ransomware friendly so that you should be able to restore your data. 770 00:39:06,045 --> 00:39:09,795 Uh, and have it essentially restored for you ready to go in a disaster. 771 00:39:09,795 --> 00:39:10,545 That's what I'm saying. 772 00:39:10,905 --> 00:39:13,605 If you're just waiting for the restore to happen, you say, then we're 773 00:39:13,605 --> 00:39:15,195 going to do our disaster recovery. 774 00:39:16,335 --> 00:39:18,255 What I'm saying is don't do it that way. 775 00:39:18,255 --> 00:39:23,955 I'm saying use one of the methods that allows you to restore the 776 00:39:23,955 --> 00:39:25,695 data before it's ever needed. 777 00:39:25,695 --> 00:39:29,535 You, you have this hot standby copy ready to go, and this will 778 00:39:29,595 --> 00:39:30,945 definitely be in the cloud. 779 00:39:30,975 --> 00:39:33,255 That's the whole point of this is that. 780 00:39:34,540 --> 00:39:38,440 Because of the cloud, you don't have to pay for the compute infrastructure 781 00:39:38,440 --> 00:39:40,450 until you need to actually test it. 782 00:39:40,810 --> 00:39:43,540 All you have to pay for is the backup copy. 783 00:39:44,020 --> 00:39:45,910 That's essentially ready to go. 784 00:39:45,910 --> 00:39:48,730 Not, not sitting there in backup format or anything like that. 785 00:39:48,730 --> 00:39:51,070 It needs to be ready to go. 786 00:39:51,580 --> 00:39:54,310 And, um, that's what I'm talking about. 787 00:39:55,960 --> 00:39:59,860 Prasanna Malaiyandi: And I think because of the cloud and other 788 00:39:59,860 --> 00:40:03,580 technologies that have come out, it's now become a lot more affordable 789 00:40:04,000 --> 00:40:07,330 before, like you said, there weren't any options, unless you had millions 790 00:40:07,330 --> 00:40:08,800 and millions of dollars in your budget. 791 00:40:09,600 --> 00:40:12,240 Now you can get a DR copy. 792 00:40:12,240 --> 00:40:14,430 That's good enough to deal with a lot of the ransomware 793 00:40:14,430 --> 00:40:16,620 scenarios and other things, right. 794 00:40:16,620 --> 00:40:18,960 That is affordable for most organizations. 795 00:40:19,680 --> 00:40:21,870 So go take a look at things out there. 796 00:40:23,180 --> 00:40:23,870 W. Curtis Preston: Absolutely. 797 00:40:25,010 --> 00:40:27,110 And I'll, I'll throw out a few random things. 798 00:40:27,110 --> 00:40:28,100 We don't need to cover them. 799 00:40:28,130 --> 00:40:31,430 I mean, these are more like ma maybe, maybe, maybe. 800 00:40:32,670 --> 00:40:34,170 Well, I'm going to come back, run back to you. 801 00:40:34,200 --> 00:40:35,490 Maybe we'll do an episode. 802 00:40:36,060 --> 00:40:38,550 Let's do an episode things you better be doing. 803 00:40:39,210 --> 00:40:39,540 Prasanna Malaiyandi: Oh, 804 00:40:39,810 --> 00:40:42,330 W. Curtis Preston: These are things you should definitely no longer be doing. 805 00:40:43,140 --> 00:40:44,970 We'll talk about things that you better be doing. 806 00:40:45,030 --> 00:40:47,820 Uh, one of them being, um, MFA. 807 00:40:48,470 --> 00:40:50,840 Prasanna Malaiyandi: yeah, I have one for you should stop doing this. 808 00:40:51,330 --> 00:40:51,750 W. Curtis Preston: Okay. 809 00:40:52,720 --> 00:40:56,860 Prasanna Malaiyandi: Um, you should stop running your backup apps 810 00:40:56,860 --> 00:41:00,080 as root and your agents as root. 811 00:41:00,720 --> 00:41:02,585 W. Curtis Preston: So you mean administering them as root? 812 00:41:03,435 --> 00:41:05,835 Prasanna Malaiyandi: At both administering or even having agents 813 00:41:05,835 --> 00:41:08,085 being run as root in your environment, 814 00:41:09,275 --> 00:41:15,305 W. Curtis Preston: Well, my only concern of that is like in a Unix environment, 815 00:41:15,395 --> 00:41:21,605 it has to run as UID zero in order to get the, the power that it needs or the 816 00:41:21,635 --> 00:41:25,385 Prasanna Malaiyandi: but not always, not all ways as 817 00:41:26,150 --> 00:41:27,200 W. Curtis Preston: What would you do? 818 00:41:27,530 --> 00:41:34,850 Other than that, it runs a system or root, like in order to get the 819 00:41:34,850 --> 00:41:39,140 permissions, to be able to see everything in a Unix world in a windows world 820 00:41:39,140 --> 00:41:41,450 has to run is the equivalent of Mr. 821 00:41:41,540 --> 00:41:42,380 Administrator. 822 00:41:42,680 --> 00:41:46,280 And what I'm saying is you could make another user ID, but in the 823 00:41:46,280 --> 00:41:47,870 Unix world, it would still have UID. 824 00:41:48,575 --> 00:41:48,875 Prasanna Malaiyandi: yeah. 825 00:41:49,485 --> 00:41:53,085 W. Curtis Preston: And so, but, and so I'm, I agree with your recommendation. 826 00:41:53,295 --> 00:41:59,895 I just want to just slightly, uh, massage it to mean if you are logging in as 827 00:41:59,895 --> 00:42:04,365 root or something, equivalent to root or administrator, and you're running your 828 00:42:04,365 --> 00:42:06,435 backup, you're running your backup system. 829 00:42:06,445 --> 00:42:08,445 That way that is absolutely wrong. 830 00:42:09,375 --> 00:42:12,315 Prasanna Malaiyandi: So I would also say though, in the case where you 831 00:42:12,315 --> 00:42:15,885 don't need to be running your agent as root, that it should not be running 832 00:42:15,885 --> 00:42:18,795 as root because not all cases need 833 00:42:19,500 --> 00:42:19,980 W. Curtis Preston: Agreed. 834 00:42:20,040 --> 00:42:22,590 I don't know what those, I don't know what those cases are. 835 00:42:22,650 --> 00:42:24,000 Give me an example of a case. 836 00:42:24,385 --> 00:42:27,235 Prasanna Malaiyandi: for database backups, you don't need to be running as root. 837 00:42:27,355 --> 00:42:27,445 In 838 00:42:27,680 --> 00:42:30,230 W. Curtis Preston: I, I, I made that up in my head, as I was saying, 839 00:42:30,230 --> 00:42:32,120 I was like, oh, he's probably talking about database backup. 840 00:42:32,150 --> 00:42:32,390 Yes. 841 00:42:32,570 --> 00:42:35,530 If the agent doesn't have to run as root then correct, don't run it as root 842 00:42:35,745 --> 00:42:35,985 Prasanna Malaiyandi: Yup. 843 00:42:36,980 --> 00:42:40,100 W. Curtis Preston: Yeah, apply the concept of least privilege to your backups. 844 00:42:40,860 --> 00:42:41,130 Prasanna Malaiyandi: Yup. 845 00:42:41,480 --> 00:42:45,350 W. Curtis Preston: And we should put that, you know, MFA, least privilege. 846 00:42:45,350 --> 00:42:48,770 We'll talk about those in are things you should be doing episode. 847 00:42:48,940 --> 00:42:49,180 Prasanna Malaiyandi: Yeah. 848 00:42:51,880 --> 00:42:54,010 W. Curtis Preston: do you think we've yelled at people enough? 849 00:42:55,145 --> 00:42:56,165 Prasanna Malaiyandi: I think so 850 00:42:57,610 --> 00:42:59,140 W. Curtis Preston: This is a very angry episode. 851 00:42:59,140 --> 00:43:00,580 Stop doing this. 852 00:43:01,675 --> 00:43:04,945 Prasanna Malaiyandi: well, it would be interesting to see how many people 853 00:43:04,945 --> 00:43:08,185 are using tools that actually do this, 854 00:43:09,815 --> 00:43:10,265 or 855 00:43:10,270 --> 00:43:11,290 W. Curtis Preston: It would be interesting. 856 00:43:11,830 --> 00:43:22,330 Well, I would say still the bulk of the backup world is doing, using backup 857 00:43:22,330 --> 00:43:24,850 products that still do full backups. 858 00:43:25,510 --> 00:43:25,780 Prasanna Malaiyandi: Yep. 859 00:43:25,940 --> 00:43:28,570 W. Curtis Preston: That still force an occasional full backup, 860 00:43:28,600 --> 00:43:30,070 even if it's a synthetic one. 861 00:43:30,610 --> 00:43:30,940 Prasanna Malaiyandi: Yeah. 862 00:43:31,880 --> 00:43:32,780 W. Curtis Preston: Think about it, right. 863 00:43:32,810 --> 00:43:34,160 Net backup, networker. 864 00:43:34,160 --> 00:43:34,850 Commvault. 865 00:43:36,485 --> 00:43:42,245 Not TSM, uh, ArcServe, I mean, you know, pick, pick a backup product 866 00:43:42,245 --> 00:43:45,995 that's been around for 20 years or more, it's doing occasional full backups. 867 00:43:46,290 --> 00:43:46,560 Prasanna Malaiyandi: Yep. 868 00:43:47,075 --> 00:43:49,235 W. Curtis Preston: It's only backup products that have been designed in 869 00:43:49,235 --> 00:43:56,585 the last 10 years or so, which would include Veeam would include, actually 870 00:43:56,625 --> 00:44:01,685 seems a little older than that, but, but the rubric Cohesity Druva. 871 00:44:03,585 --> 00:44:08,395 Trying to think Clumio , these are products that, that they're 872 00:44:08,395 --> 00:44:11,935 like, basically they're the, they're the disk generation. 873 00:44:12,565 --> 00:44:14,005 They're, they're making an assumption. 874 00:44:14,005 --> 00:44:17,095 We're going to store backups on disk or something behaving like disc. 875 00:44:17,095 --> 00:44:19,645 And so therefore the idea of a full backup is just dumb. 876 00:44:19,715 --> 00:44:20,525 Prasanna Malaiyandi: bonkers. 877 00:44:20,875 --> 00:44:21,835 W. Curtis Preston: And so let's not do it. 878 00:44:21,865 --> 00:44:25,615 Let's let's design, let's design a backup product from the beginning 879 00:44:25,615 --> 00:44:26,915 to not need a full backup. 880 00:44:28,200 --> 00:44:29,270 A repeated full backup. 881 00:44:29,300 --> 00:44:30,140 Obviously we always 882 00:44:30,730 --> 00:44:31,400 always need 883 00:44:32,275 --> 00:44:32,605 Prasanna Malaiyandi: At least 884 00:44:32,810 --> 00:44:33,230 W. Curtis Preston: backup, 885 00:44:33,950 --> 00:44:42,750 just one, one, and done baby one and done just like, no, I was just going 886 00:44:42,750 --> 00:44:48,320 to try to make a go go North Carolina. 887 00:44:48,380 --> 00:44:49,040 You don't even know. 888 00:44:49,040 --> 00:44:50,340 Cause you don't follow NCAA. 889 00:44:50,420 --> 00:44:52,880 You follow, you don't really follow basketball. 890 00:44:52,880 --> 00:44:53,900 You follow football. 891 00:44:54,085 --> 00:44:54,445 Prasanna Malaiyandi: Nope. 892 00:44:54,565 --> 00:44:55,555 I just follow football. 893 00:44:56,125 --> 00:44:56,635 the NFL. 894 00:44:56,760 --> 00:45:01,530 W. Curtis Preston: Yeah, North Carolina won over Duke this past weekend. 895 00:45:01,835 --> 00:45:02,675 Prasanna Malaiyandi: It was an upset 896 00:45:03,420 --> 00:45:04,380 W. Curtis Preston: It was an upset. 897 00:45:04,530 --> 00:45:08,340 And you know, and the guy, the coach that, you know, the coach ended his 898 00:45:08,340 --> 00:45:12,390 career on a loss, which is a shame, but he's had an amazing career. 899 00:45:12,390 --> 00:45:13,080 So good for him. 900 00:45:13,980 --> 00:45:16,740 Well, Prasanna, thanks for letting me rant. 901 00:45:17,555 --> 00:45:18,125 Prasanna Malaiyandi: you feeling better. 902 00:45:19,800 --> 00:45:23,670 W. Curtis Preston: Well, I'm still going to be hanging out in this office 903 00:45:23,790 --> 00:45:24,420 for the next. 904 00:45:24,530 --> 00:45:26,540 Prasanna Malaiyandi: but are you at least feeling better? 905 00:45:27,345 --> 00:45:28,575 W. Curtis Preston: Yeah, I'm feeling better. 906 00:45:28,725 --> 00:45:33,135 Uh, you know, I don't know if I, I thought I made a, I don't know if you've ever 907 00:45:33,135 --> 00:45:36,855 been in a house where somebody had lice, 908 00:45:38,135 --> 00:45:38,895 you've been in a house. 909 00:45:39,110 --> 00:45:39,500 Prasanna Malaiyandi: Yep. 910 00:45:39,925 --> 00:45:42,855 W. Curtis Preston: Did you start like psychosomatically 911 00:45:42,885 --> 00:45:44,325 feeling lice in your hair. 912 00:45:46,045 --> 00:45:46,475 Yeah. 913 00:45:46,895 --> 00:45:55,535 So that's where I'm like, I mean, if, if I got infected, I I'm not 914 00:45:55,535 --> 00:45:58,955 going to have symptoms yet, but still my body is like, I think is that 915 00:45:58,955 --> 00:46:02,645 a fever said, you know, said so. 916 00:46:03,575 --> 00:46:05,835 Yeah, good times. 917 00:46:06,235 --> 00:46:06,875 Prasanna Malaiyandi: lots of tea. 918 00:46:07,865 --> 00:46:08,945 W. Curtis Preston: Drink, lots of tea. 919 00:46:09,305 --> 00:46:13,925 Um, not, not whiskey, although I don't know why I said whiskey. 920 00:46:13,925 --> 00:46:14,795 I don't drink whiskey. 921 00:46:14,930 --> 00:46:15,260 Prasanna Malaiyandi: This 922 00:46:16,235 --> 00:46:21,335 W. Curtis Preston: I'm not a whiskey bourbon or scotch person or cigars. 923 00:46:21,335 --> 00:46:22,235 Like my friend, Jeff. 924 00:46:22,265 --> 00:46:23,615 See you do cigars. 925 00:46:23,615 --> 00:46:24,425 You get COVID. 926 00:46:24,605 --> 00:46:25,055 That's it. 927 00:46:25,325 --> 00:46:26,885 That's that's the lesson that 928 00:46:26,885 --> 00:46:27,785 Jeff should learn. 929 00:46:28,610 --> 00:46:30,500 Prasanna Malaiyandi: Curtis and I are not medical doctors. 930 00:46:30,500 --> 00:46:32,210 We are not providing any medical advice. 931 00:46:32,390 --> 00:46:35,630 Please go talk to your doctor or do your research. 932 00:46:35,660 --> 00:46:36,050 Thank you. 933 00:46:36,215 --> 00:46:38,405 W. Curtis Preston: please don't get your medical advice on 934 00:46:38,705 --> 00:46:40,415 Backup Central's Restore it All 935 00:46:40,415 --> 00:46:42,725 we, that is not, it's 936 00:46:42,770 --> 00:46:43,070 Prasanna Malaiyandi: at least. 937 00:46:43,070 --> 00:46:44,030 from the two of us. 938 00:46:44,795 --> 00:46:45,155 W. Curtis Preston: at least. 939 00:46:45,155 --> 00:46:45,425 Yeah. 940 00:46:45,635 --> 00:46:46,115 Right. 941 00:46:46,565 --> 00:46:52,265 We got to have Lindsey back on and I'm just waiting for like, I don't know. 942 00:46:52,265 --> 00:46:55,145 I kinda just want to bring her back on and say, is it over? 943 00:46:55,955 --> 00:46:58,475 Can we, can we move on now? 944 00:46:59,045 --> 00:47:01,685 That's that's the episode I want to have Lindsay back on for. 945 00:47:02,900 --> 00:47:06,830 And if you guys don't know what I'm talking about, we did half a 946 00:47:06,830 --> 00:47:11,570 dozen episodes in the early days of the pandemic with, uh, Dr. 947 00:47:11,570 --> 00:47:12,350 Lindsey Schultz. 948 00:47:12,410 --> 00:47:18,130 And she, she actually is a doctor medical doctor with a, with a, a 949 00:47:18,200 --> 00:47:20,570 master's in public health as well. 950 00:47:20,840 --> 00:47:25,280 And she specializes in harms reduction. 951 00:47:25,280 --> 00:47:28,970 It's a kind of, kind of logic that you use when. 952 00:47:29,735 --> 00:47:32,555 Wanting to say, if we're going to make this public policy. 953 00:47:33,665 --> 00:47:38,675 And, um, anyway, so yeah, I want to have her back on, but I want 954 00:47:38,675 --> 00:47:41,165 to, I just, I want to have her on when it's like just good news. 955 00:47:41,375 --> 00:47:42,425 Yes, we're done. 956 00:47:43,025 --> 00:47:44,615 We can move on with our life. 957 00:47:44,855 --> 00:47:45,575 No more. 958 00:47:46,085 --> 00:47:48,065 COVID scares like the one I'm in right now. 959 00:47:49,205 --> 00:47:51,125 Damn it anyway. 960 00:47:51,575 --> 00:47:51,905 All right. 961 00:47:51,965 --> 00:47:56,855 I'm going to end this thing before, before I, before I start to despair. 962 00:47:57,185 --> 00:48:02,015 So thank you Prasanna, you know, like, like I said, for letting me rant 963 00:48:02,540 --> 00:48:04,790 Prasanna Malaiyandi: Anytime Curtis, and I hope you feel better. 964 00:48:06,025 --> 00:48:10,165 W. Curtis Preston: and, uh, thank you to the listeners and be sure to 965 00:48:10,165 --> 00:48:12,775 subscribe so that you can restore it all.