1
00:00:00,480 --> 00:00:03,480
foreign

2
00:00:09,200 --> 00:00:13,620
's career has spanned two centuries if

3
00:00:12,240 --> 00:00:15,540
not millennia

4
00:00:13,620 --> 00:00:17,820
he has worked a variety of roles within

5
00:00:15,540 --> 00:00:20,039
the I.T industry and one day he may even

6
00:00:17,820 --> 00:00:22,560
pick one and stick with it

7
00:00:20,039 --> 00:00:24,600
today however is not that day

8
00:00:22,560 --> 00:00:26,580
Dave is going to tell us about how one

9
00:00:24,600 --> 00:00:29,519
person's ansible is another person's pie

10
00:00:26,580 --> 00:00:31,560
infra take it away Dave

11
00:00:29,519 --> 00:00:33,719
Daryl sorry

12
00:00:31,560 --> 00:00:35,940
[Applause]

13
00:00:33,719 --> 00:00:37,620
um I'd like to pay my respects to the

14
00:00:35,940 --> 00:00:39,600
traditional owners of the land on which

15
00:00:37,620 --> 00:00:41,579
we meet today the Ghana people of the

16
00:00:39,600 --> 00:00:43,980
Adelaide region and pay my respects to

17
00:00:41,579 --> 00:00:45,540
Elders past present and emerging I'd

18
00:00:43,980 --> 00:00:47,820
also like to thank our interstate and

19
00:00:45,540 --> 00:00:50,399
overseas visitors for visiting my home

20
00:00:47,820 --> 00:00:51,360
state it's lovely to have you here and

21
00:00:50,399 --> 00:00:52,940
you should definitely come back when

22
00:00:51,360 --> 00:00:55,199
it's a lot warmer

23
00:00:52,940 --> 00:00:58,920
preferably when we're not on fire though

24
00:00:55,199 --> 00:01:00,300
as an ex CFS member I'm qualified to say

25
00:00:58,920 --> 00:01:02,100
that

26
00:01:00,300 --> 00:01:04,920
I would also like to acknowledge my

27
00:01:02,100 --> 00:01:05,640
employers if they're watching

28
00:01:04,920 --> 00:01:07,619
um

29
00:01:05,640 --> 00:01:09,900
technically I'm self-employed but the

30
00:01:07,619 --> 00:01:11,520
people pay me to do money people pay me

31
00:01:09,900 --> 00:01:13,619
money so to do things to their computers

32
00:01:11,520 --> 00:01:15,479
and My Views probably don't represent

33
00:01:13,619 --> 00:01:18,600
theirs

34
00:01:15,479 --> 00:01:20,640
um and I'd also uh would like to ask for

35
00:01:18,600 --> 00:01:22,320
questions at the end of the talk

36
00:01:20,640 --> 00:01:24,720
um I'm a little bit nervous up here this

37
00:01:22,320 --> 00:01:27,960
first time for me so

38
00:01:24,720 --> 00:01:30,840
quite easily derailed in my thoughts

39
00:01:27,960 --> 00:01:32,400
I'd also like to add to that the GitHub

40
00:01:30,840 --> 00:01:34,020
repo isn't ready yet

41
00:01:32,400 --> 00:01:36,180
um the talk will be uploaded probably in

42
00:01:34,020 --> 00:01:38,540
the next couple of days so uh don't hit

43
00:01:36,180 --> 00:01:38,540
it yet

44
00:01:38,820 --> 00:01:44,100
so I'm going to make some assumptions

45
00:01:41,939 --> 00:01:46,200
um that you probably all got a passing

46
00:01:44,100 --> 00:01:48,360
understanding of python and you probably

47
00:01:46,200 --> 00:01:50,640
all got a passing understanding of

48
00:01:48,360 --> 00:01:51,659
interval which was why you're here if

49
00:01:50,640 --> 00:01:54,659
you're thinking this is an ansible

50
00:01:51,659 --> 00:01:58,439
beginner's talk not quite

51
00:01:54,659 --> 00:01:59,759
um and I'm not really seeking to convert

52
00:01:58,439 --> 00:02:02,240
you I'm just providing an alternate

53
00:01:59,759 --> 00:02:02,240
point of view

54
00:02:04,079 --> 00:02:07,439
so previous talk titles that I've

55
00:02:06,240 --> 00:02:08,160
wrestled with

56
00:02:07,439 --> 00:02:10,039
um

57
00:02:08,160 --> 00:02:14,120
is

58
00:02:10,039 --> 00:02:16,860
uh yes what you see here these are the

59
00:02:14,120 --> 00:02:18,420
the more family friendly ones

60
00:02:16,860 --> 00:02:20,640
um

61
00:02:18,420 --> 00:02:22,560
so and to be clear these are talk titles

62
00:02:20,640 --> 00:02:24,180
that were rejected by myself not by the

63
00:02:22,560 --> 00:02:25,920
program chair the program chair here has

64
00:02:24,180 --> 00:02:28,080
been lovely

65
00:02:25,920 --> 00:02:30,239
um I've been struggling with the theme

66
00:02:28,080 --> 00:02:31,800
of this talk and it's something I

67
00:02:30,239 --> 00:02:35,640
wrestle with constantly

68
00:02:31,800 --> 00:02:37,920
um because I'm very uh OCD about diving

69
00:02:35,640 --> 00:02:39,379
into problems and problems that I

70
00:02:37,920 --> 00:02:42,599
perceive

71
00:02:39,379 --> 00:02:45,180
so this is kind of what that talks about

72
00:02:42,599 --> 00:02:46,920
but I still haven't got basically the

73
00:02:45,180 --> 00:02:48,720
essence or the Zen of this talk out yet

74
00:02:46,920 --> 00:02:50,340
so

75
00:02:48,720 --> 00:02:52,640
um

76
00:02:50,340 --> 00:02:56,580
still a work in progress

77
00:02:52,640 --> 00:02:57,900
so it's also sort of documenting my

78
00:02:56,580 --> 00:03:00,000
struggles with ansible over the years

79
00:02:57,900 --> 00:03:02,819
and so

80
00:03:00,000 --> 00:03:05,760
which will lead me I guess to confession

81
00:03:02,819 --> 00:03:07,500
time which is what was taught currently

82
00:03:05,760 --> 00:03:09,959
kind of is

83
00:03:07,500 --> 00:03:11,459
um I struggle with ansible like I've

84
00:03:09,959 --> 00:03:14,819
been using it for probably five years

85
00:03:11,459 --> 00:03:17,400
now and not what it does or

86
00:03:14,819 --> 00:03:18,680
what it is but more how it does what it

87
00:03:17,400 --> 00:03:21,959
does

88
00:03:18,680 --> 00:03:24,780
I find ansible

89
00:03:21,959 --> 00:03:26,280
tends to be a tool there are some tools

90
00:03:24,780 --> 00:03:29,459
that sort of fit into this model where

91
00:03:26,280 --> 00:03:30,900
you're trying to solve a problem and you

92
00:03:29,459 --> 00:03:32,220
have to think about that problem in

93
00:03:30,900 --> 00:03:35,580
terms of the tool that you're using

94
00:03:32,220 --> 00:03:37,319
against it and and I find that quite

95
00:03:35,580 --> 00:03:38,640
often that was my case with ansible is

96
00:03:37,319 --> 00:03:40,980
like I'd have to

97
00:03:38,640 --> 00:03:45,080
structure it in terms of the way ansible

98
00:03:40,980 --> 00:03:45,080
thinks too to do things

99
00:03:45,180 --> 00:03:52,260
so and it sort of it it caused a fair

100
00:03:48,060 --> 00:03:55,560
bit of cognitive workload on myself

101
00:03:52,260 --> 00:03:57,180
doing that just having to deal with then

102
00:03:55,560 --> 00:03:58,560
I won't call them shortcomings they're

103
00:03:57,180 --> 00:04:01,140
just it's the way ansible's been

104
00:03:58,560 --> 00:04:02,340
developed and grown over the years

105
00:04:01,140 --> 00:04:04,080
um

106
00:04:02,340 --> 00:04:05,459
and it sort of it takes away the

107
00:04:04,080 --> 00:04:08,360
workload from actually thinking about

108
00:04:05,459 --> 00:04:08,360
the problem itself

109
00:04:08,480 --> 00:04:12,900
some of the

110
00:04:10,860 --> 00:04:14,099
uh the cognitive workloads that I

111
00:04:12,900 --> 00:04:15,000
mentioned

112
00:04:14,099 --> 00:04:17,880
um

113
00:04:15,000 --> 00:04:20,459
keeping track of global state that sort

114
00:04:17,880 --> 00:04:24,120
of makes me feel like my old GW basic

115
00:04:20,459 --> 00:04:25,620
days and to be fair I came to GW basic

116
00:04:24,120 --> 00:04:29,040
after I'd worked in other Basics that

117
00:04:25,620 --> 00:04:30,900
had things like local variables and name

118
00:04:29,040 --> 00:04:32,460
subroutines so it was going back and

119
00:04:30,900 --> 00:04:33,960
trying to have to think of global

120
00:04:32,460 --> 00:04:37,220
variables again

121
00:04:33,960 --> 00:04:39,540
uh and winter and winter not Ginger

122
00:04:37,220 --> 00:04:43,199
Ginger coding still kicks my butt on a

123
00:04:39,540 --> 00:04:43,199
regular basis uh

124
00:04:43,979 --> 00:04:49,380
the obtuse error messages and this is a

125
00:04:46,800 --> 00:04:51,600
true saver message is more a case of um

126
00:04:49,380 --> 00:04:53,820
if you've forgotten to download a

127
00:04:51,600 --> 00:04:56,120
collection that you're using you'll come

128
00:04:53,820 --> 00:04:59,940
up with a rather bizarre error

129
00:04:56,120 --> 00:05:01,460
it's sort of it's a the equivalent of a

130
00:04:59,940 --> 00:05:06,120
code smell but in the design space

131
00:05:01,460 --> 00:05:09,180
because it's a parsing issue when that

132
00:05:06,120 --> 00:05:10,740
happens and it sort of it it again when

133
00:05:09,180 --> 00:05:12,419
it occurs I have to think about it and

134
00:05:10,740 --> 00:05:14,040
think why are you doing that is that

135
00:05:12,419 --> 00:05:16,680
something I've coded wrong and it's like

136
00:05:14,040 --> 00:05:19,460
oh no I've just forgotten to incorporate

137
00:05:16,680 --> 00:05:19,460
the correct collection

138
00:05:19,740 --> 00:05:25,620
um data structures I tend to a lot of my

139
00:05:23,460 --> 00:05:27,360
playbooks tend to be well complicated

140
00:05:25,620 --> 00:05:29,460
playbooks tend to be build data

141
00:05:27,360 --> 00:05:31,500
structures around how I want things to

142
00:05:29,460 --> 00:05:33,419
be deployed

143
00:05:31,500 --> 00:05:37,680
um if when they start getting moderately

144
00:05:33,419 --> 00:05:39,419
complex and coupled with uh templating

145
00:05:37,680 --> 00:05:41,900
it gets

146
00:05:39,419 --> 00:05:45,479
pretty messy pretty quickly

147
00:05:41,900 --> 00:05:46,860
include task versus import tasks I

148
00:05:45,479 --> 00:05:48,720
struggle with that all the time I can

149
00:05:46,860 --> 00:05:50,039
never remember when to do which I've

150
00:05:48,720 --> 00:05:51,840
always got to look up the documentation

151
00:05:50,039 --> 00:05:55,199
for that

152
00:05:51,840 --> 00:05:57,860
nested looping as probably also my

153
00:05:55,199 --> 00:06:01,979
favorite is there is many ways to do it

154
00:05:57,860 --> 00:06:04,440
in uh ansible

155
00:06:01,979 --> 00:06:07,580
not necessarily

156
00:06:04,440 --> 00:06:07,580
every way is equal

157
00:06:08,220 --> 00:06:11,820
there are just some dead ends that

158
00:06:10,380 --> 00:06:13,080
you'll go to well they're not that ends

159
00:06:11,820 --> 00:06:14,340
until you've actually gone down them and

160
00:06:13,080 --> 00:06:18,020
then you'll realize no I should have

161
00:06:14,340 --> 00:06:18,020
handled this differently so

162
00:06:18,120 --> 00:06:21,419
so

163
00:06:19,320 --> 00:06:23,880
I I tend to think long and hard about

164
00:06:21,419 --> 00:06:26,400
these things I spend a lot of time lying

165
00:06:23,880 --> 00:06:27,479
awake at night staring at my ceiling

166
00:06:26,400 --> 00:06:28,740
um

167
00:06:27,479 --> 00:06:33,120
I shouldn't do that I should have

168
00:06:28,740 --> 00:06:34,979
hobbies but I'd only waste them uh so

169
00:06:33,120 --> 00:06:36,479
while I was sitting there and again I

170
00:06:34,979 --> 00:06:38,580
tend to like research these things and

171
00:06:36,479 --> 00:06:39,840
trying to dig into the nature of what it

172
00:06:38,580 --> 00:06:41,819
is

173
00:06:39,840 --> 00:06:44,400
um somewhere between giving the previous

174
00:06:41,819 --> 00:06:46,740
version of this talk and this one which

175
00:06:44,400 --> 00:06:49,800
was probably about a year ago now in the

176
00:06:46,740 --> 00:06:52,800
python local Adelaide users group I came

177
00:06:49,800 --> 00:06:54,419
across this quote

178
00:06:52,800 --> 00:06:56,759
um

179
00:06:54,419 --> 00:07:00,419
which sort of

180
00:06:56,759 --> 00:07:04,259
sealed the deal for me it was like yeah

181
00:07:00,419 --> 00:07:06,479
it was it was quite an opener

182
00:07:04,259 --> 00:07:07,259
and it sort of it did actually appeal to

183
00:07:06,479 --> 00:07:08,460
me

184
00:07:07,259 --> 00:07:10,620
um it was it was pretty much a light

185
00:07:08,460 --> 00:07:12,380
bulb moment when I came across this

186
00:07:10,620 --> 00:07:15,660
ironically the

187
00:07:12,380 --> 00:07:18,419
coincidentally the uh question talking

188
00:07:15,660 --> 00:07:19,620
about this was talking about ansible uh

189
00:07:18,419 --> 00:07:21,060
and I wasn't actually researching

190
00:07:19,620 --> 00:07:23,220
ansible stuff in the time I was just

191
00:07:21,060 --> 00:07:25,139
having to be looking for something else

192
00:07:23,220 --> 00:07:26,880
but when I came across this it was just

193
00:07:25,139 --> 00:07:29,400
like oh okay he's

194
00:07:26,880 --> 00:07:30,840
studying he's suffering from similar

195
00:07:29,400 --> 00:07:32,280
issues with ansible and I was finding

196
00:07:30,840 --> 00:07:34,680
there's actually a small group of us

197
00:07:32,280 --> 00:07:36,960
like that that are struggling with some

198
00:07:34,680 --> 00:07:38,520
of these issues

199
00:07:36,960 --> 00:07:40,199
this is sounding like a self-help group

200
00:07:38,520 --> 00:07:41,580
sorry about that

201
00:07:40,199 --> 00:07:46,199
um

202
00:07:41,580 --> 00:07:48,180
and coincidentally uh he while I was

203
00:07:46,199 --> 00:07:49,860
sort of looking for a new product to

204
00:07:48,180 --> 00:07:52,740
pick up I'd already picked it up by the

205
00:07:49,860 --> 00:07:55,020
stage and he was also talking about the

206
00:07:52,740 --> 00:07:57,500
same product

207
00:07:55,020 --> 00:07:57,500
hello

208
00:07:58,639 --> 00:08:04,699
certainly so you have your my permission

209
00:08:02,160 --> 00:08:04,699
to touch me

210
00:08:09,720 --> 00:08:12,680
okay

211
00:08:14,060 --> 00:08:19,259
it's me it's my uh well now I can hear

212
00:08:17,699 --> 00:08:22,160
myself that's worse

213
00:08:19,259 --> 00:08:22,160
please don't do that

214
00:08:24,660 --> 00:08:30,139
I'm sorry that's okay

215
00:08:27,000 --> 00:08:30,139
it could be my metal implant

216
00:08:32,839 --> 00:08:37,740
maybe I'll just try not to move so much

217
00:08:35,640 --> 00:08:40,580
was that better

218
00:08:37,740 --> 00:08:40,580
I would just

219
00:08:40,860 --> 00:08:44,539
this is going to be awkward

220
00:08:45,180 --> 00:08:50,279
um so

221
00:08:47,100 --> 00:08:51,420
I was reading um lox blog post

222
00:08:50,279 --> 00:08:54,000
um he was talking about problems with

223
00:08:51,420 --> 00:08:55,440
ansible but he was also saying oh I've

224
00:08:54,000 --> 00:08:58,080
also found the solution which I'd

225
00:08:55,440 --> 00:08:59,940
already started dabbling in and by

226
00:08:58,080 --> 00:09:01,019
dabbling I mean actually wholeheartedly

227
00:08:59,940 --> 00:09:06,480
committing

228
00:09:01,019 --> 00:09:08,220
and the thing in question was Pi infra

229
00:09:06,480 --> 00:09:10,560
when I came across Pine for I sort of

230
00:09:08,220 --> 00:09:12,360
went oh because it's rare that I do that

231
00:09:10,560 --> 00:09:14,040
for a project these days

232
00:09:12,360 --> 00:09:17,339
um but it was hitting all

233
00:09:14,040 --> 00:09:18,660
it was solving a lot of no solving's

234
00:09:17,339 --> 00:09:20,399
probably not the correct word but it was

235
00:09:18,660 --> 00:09:22,320
covering a lot of the pain points that I

236
00:09:20,399 --> 00:09:23,459
was having with ansible

237
00:09:22,320 --> 00:09:27,420
um

238
00:09:23,459 --> 00:09:29,100
so uh basically it's very similar it

239
00:09:27,420 --> 00:09:31,620
runs in the sort of the same problem

240
00:09:29,100 --> 00:09:34,100
space as ansible it's uh

241
00:09:31,620 --> 00:09:37,440
an agentless

242
00:09:34,100 --> 00:09:38,519
orchestration tool a impotent for the

243
00:09:37,440 --> 00:09:39,180
most part

244
00:09:38,519 --> 00:09:40,680
um

245
00:09:39,180 --> 00:09:42,839
you can still sort of work around that

246
00:09:40,680 --> 00:09:45,080
but that's

247
00:09:42,839 --> 00:09:47,820
something you shouldn't do

248
00:09:45,080 --> 00:09:50,820
and it has inventories and playbooks

249
00:09:47,820 --> 00:09:51,420
well the equivalent of playbooks

250
00:09:50,820 --> 00:09:56,060
um

251
00:09:51,420 --> 00:09:56,060
but they're all written in Python

252
00:09:56,420 --> 00:10:01,080
it's

253
00:09:58,260 --> 00:10:03,420
so generally it's it's it's a similar

254
00:10:01,080 --> 00:10:07,760
model to python it runs basically on a

255
00:10:03,420 --> 00:10:10,140
control node sorry python ansible

256
00:10:07,760 --> 00:10:12,540
however it sort of differs from ansible

257
00:10:10,140 --> 00:10:15,120
ansible tends to upload well what tends

258
00:10:12,540 --> 00:10:17,640
to ansible compiles a module uploads it

259
00:10:15,120 --> 00:10:19,519
to the remote end and runs it expects

260
00:10:17,640 --> 00:10:21,899
you to have python on there

261
00:10:19,519 --> 00:10:23,640
Pi in for a dozen buying for just

262
00:10:21,899 --> 00:10:27,480
expects

263
00:10:23,640 --> 00:10:29,880
a working shell of some sort at SSH is

264
00:10:27,480 --> 00:10:32,640
across and not necessarily just SSH it

265
00:10:29,880 --> 00:10:34,760
has a whole bunch of connectors but SSH

266
00:10:32,640 --> 00:10:36,779
is one of them and is the default one

267
00:10:34,760 --> 00:10:39,540
executes a bunch of commands on there

268
00:10:36,779 --> 00:10:42,060
copies the results back or retrieves the

269
00:10:39,540 --> 00:10:44,519
results back and then processes that and

270
00:10:42,060 --> 00:10:47,600
then continues on from there it has

271
00:10:44,519 --> 00:10:50,339
facts for doing a lot of that

272
00:10:47,600 --> 00:10:52,260
so it will use the fax to determine the

273
00:10:50,339 --> 00:10:55,260
current state if you

274
00:10:52,260 --> 00:10:56,760
structure you deploy correctly it'll use

275
00:10:55,260 --> 00:10:57,720
fax to determine the current state of

276
00:10:56,760 --> 00:11:00,240
the system

277
00:10:57,720 --> 00:11:02,279
then apply a set of operations to bring

278
00:11:00,240 --> 00:11:03,779
it to the desired state that you want so

279
00:11:02,279 --> 00:11:06,540
this is where the eye dependency comes

280
00:11:03,779 --> 00:11:08,459
in high dependency

281
00:11:06,540 --> 00:11:11,459
sorry practice that word and I still

282
00:11:08,459 --> 00:11:12,899
have trouble speaking it out loud so

283
00:11:11,459 --> 00:11:14,220
this is where the idepatency comes in

284
00:11:12,899 --> 00:11:15,959
where you would sit there if you

285
00:11:14,220 --> 00:11:17,579
re-execute a Playbook and there's no

286
00:11:15,959 --> 00:11:20,459
changes to be made it should not make

287
00:11:17,579 --> 00:11:24,060
any changes to that

288
00:11:20,459 --> 00:11:25,800
to the desired state of the system uh I

289
00:11:24,060 --> 00:11:27,959
was saying before it has connectors what

290
00:11:25,800 --> 00:11:29,579
it what it calls connectors so SSH is

291
00:11:27,959 --> 00:11:31,200
usually the bog standard one you use

292
00:11:29,579 --> 00:11:33,600
unless you're talking to your local

293
00:11:31,200 --> 00:11:36,060
machine and it's got a specific one

294
00:11:33,600 --> 00:11:37,620
called local for that basically a

295
00:11:36,060 --> 00:11:40,380
connector is any means that it can use

296
00:11:37,620 --> 00:11:43,019
to talk to the remote endpoint other

297
00:11:40,380 --> 00:11:46,160
connectors that it includes a connectors

298
00:11:43,019 --> 00:11:46,160
for Docker terraform

299
00:11:46,220 --> 00:11:51,060
and there's a few others as well which

300
00:11:48,240 --> 00:11:53,279
completely escaped my mind the SSH

301
00:11:51,060 --> 00:11:54,320
sessions remain open for the duration of

302
00:11:53,279 --> 00:11:57,600
the run

303
00:11:54,320 --> 00:11:59,880
if you run ansible and your playbook has

304
00:11:57,600 --> 00:12:02,220
multiple plays in it it tends to do an

305
00:11:59,880 --> 00:12:05,459
SSH connection for each one if you've

306
00:12:02,220 --> 00:12:07,980
got uh like some environments I've

307
00:12:05,459 --> 00:12:10,440
worked in previously

308
00:12:07,980 --> 00:12:12,300
um sorry waving to a colleague there X

309
00:12:10,440 --> 00:12:14,100
Colleen

310
00:12:12,300 --> 00:12:16,680
um

311
00:12:14,100 --> 00:12:18,180
uh for SSH connections if you're

312
00:12:16,680 --> 00:12:20,100
spawning off a bunch of them a

313
00:12:18,180 --> 00:12:22,200
connection create quite a sort of uh

314
00:12:20,100 --> 00:12:23,880
quite a workload behind

315
00:12:22,200 --> 00:12:26,040
um sorry that's me making noises again

316
00:12:23,880 --> 00:12:28,800
sorry

317
00:12:26,040 --> 00:12:31,560
I don't think I can keep still

318
00:12:28,800 --> 00:12:33,260
um if you spawn a bunch a large bunch of

319
00:12:31,560 --> 00:12:35,820
SSH connections you can actually create

320
00:12:33,260 --> 00:12:38,760
uh quite the workload behind on your

321
00:12:35,820 --> 00:12:40,500
authentication back-end system so uh

322
00:12:38,760 --> 00:12:43,260
with this it spawns the SSH connection

323
00:12:40,500 --> 00:12:46,560
but it remains for the duration of the

324
00:12:43,260 --> 00:12:48,180
entire run it doesn't uh sit there and

325
00:12:46,560 --> 00:12:51,060
reconnect

326
00:12:48,180 --> 00:12:52,500
it also opens ssh in a slightly

327
00:12:51,060 --> 00:12:55,860
different mode to what ansible does

328
00:12:52,500 --> 00:12:58,740
ansible is quite noisy if you do uh for

329
00:12:55,860 --> 00:13:00,360
example it logs in using a regular user

330
00:12:58,740 --> 00:13:02,220
login account so if you do a who on the

331
00:13:00,360 --> 00:13:04,200
system afterwards you can see you've

332
00:13:02,220 --> 00:13:07,440
just done a manageable run because your

333
00:13:04,200 --> 00:13:10,440
ansible user is logged in quite often

334
00:13:07,440 --> 00:13:12,660
this uses ssh in a

335
00:13:10,440 --> 00:13:14,639
uh in more of the remote connect mode

336
00:13:12,660 --> 00:13:16,740
like when you're typing SSH hostname

337
00:13:14,639 --> 00:13:18,420
command so it doesn't actually appear in

338
00:13:16,740 --> 00:13:19,800
your login but it still appears in your

339
00:13:18,420 --> 00:13:23,839
ortholog so if you need to go back and

340
00:13:19,800 --> 00:13:23,839
do any sort of research for that

341
00:13:24,000 --> 00:13:27,240
um so it's saying it uses the remote

342
00:13:25,440 --> 00:13:29,880
Shell at the far end it's not

343
00:13:27,240 --> 00:13:33,180
particularly wedded to the remote shell

344
00:13:29,880 --> 00:13:34,740
that you use it's got a variety of um

345
00:13:33,180 --> 00:13:36,899
different sort of plugins where you can

346
00:13:34,740 --> 00:13:39,240
sit there and and depend on your back

347
00:13:36,899 --> 00:13:41,480
end shell that you connect to you could

348
00:13:39,240 --> 00:13:44,180
be connecting to some proprietary device

349
00:13:41,480 --> 00:13:47,940
like a switch for example

350
00:13:44,180 --> 00:13:50,040
and uh and it's particular set of

351
00:13:47,940 --> 00:13:52,920
commands to use for that

352
00:13:50,040 --> 00:13:55,019
the deploy process runs as two passes so

353
00:13:52,920 --> 00:13:56,240
the the initial run

354
00:13:55,019 --> 00:13:58,519
um sits there

355
00:13:56,240 --> 00:14:00,480
executes all the code

356
00:13:58,519 --> 00:14:02,100
gathers all the facts that it requires

357
00:14:00,480 --> 00:14:03,959
to do that and it builds up a set of

358
00:14:02,100 --> 00:14:06,240
operations and then the second part just

359
00:14:03,959 --> 00:14:08,160
blats those operations on top of the the

360
00:14:06,240 --> 00:14:10,639
system to bring it into the state that's

361
00:14:08,160 --> 00:14:10,639
required

362
00:14:10,700 --> 00:14:17,839
so I present you a concrete example

363
00:14:14,100 --> 00:14:20,220
um this is a

364
00:14:17,839 --> 00:14:22,620
lab maintenance script I call it a home

365
00:14:20,220 --> 00:14:26,760
lab maintenance script that I used so I

366
00:14:22,620 --> 00:14:29,399
run a small set of vps's and whatnot

367
00:14:26,760 --> 00:14:32,339
um I quite often

368
00:14:29,399 --> 00:14:33,839
SSH into them as we all probably do

369
00:14:32,339 --> 00:14:35,760
um but

370
00:14:33,839 --> 00:14:38,820
what I don't want to see when I SSH into

371
00:14:35,760 --> 00:14:40,380
them is that

372
00:14:38,820 --> 00:14:41,820
or usually I don't want to see the

373
00:14:40,380 --> 00:14:43,920
initial connect message that the initial

374
00:14:41,820 --> 00:14:46,079
connect message is usually you've never

375
00:14:43,920 --> 00:14:48,600
logged into this host before do you want

376
00:14:46,079 --> 00:14:50,399
to and it's like I think it's the host

377
00:14:48,600 --> 00:14:53,579
I'm connecting to I always have that

378
00:14:50,399 --> 00:14:57,480
moment of Doubt I've

379
00:14:53,579 --> 00:14:59,100
dagnabbit now I can hear it sorry guys

380
00:14:57,480 --> 00:15:01,920
um I always have that moment of Doubt

381
00:14:59,100 --> 00:15:04,260
logging into it and I know of a previous

382
00:15:01,920 --> 00:15:06,360
Linux conference where someone did do

383
00:15:04,260 --> 00:15:09,300
that and ignored the warning and were

384
00:15:06,360 --> 00:15:10,639
subsequently slightly pwned

385
00:15:09,300 --> 00:15:15,839
so

386
00:15:10,639 --> 00:15:18,779
uh for SSH key hygiene so

387
00:15:15,839 --> 00:15:21,480
I tend to now be quite security

388
00:15:18,779 --> 00:15:24,360
conscious I tend to make sure that I

389
00:15:21,480 --> 00:15:27,360
populate my global

390
00:15:24,360 --> 00:15:28,800
known hosts SSH keys with all the keys

391
00:15:27,360 --> 00:15:30,959
of the machines I'm connecting to so

392
00:15:28,800 --> 00:15:33,000
that when I do connect to them if I ever

393
00:15:30,959 --> 00:15:34,980
get a scary message like before I know

394
00:15:33,000 --> 00:15:36,779
something is really wrong and I never

395
00:15:34,980 --> 00:15:39,740
get the you're connecting to this

396
00:15:36,779 --> 00:15:39,740
machine for the first time

397
00:15:41,240 --> 00:15:44,660
sorry just

398
00:15:44,880 --> 00:15:51,180
what's up so host names this is uh good

399
00:15:48,959 --> 00:15:53,220
Lord my slide is incredibly small sorry

400
00:15:51,180 --> 00:15:55,339
I've got to appear around at this one um

401
00:15:53,220 --> 00:15:58,920
so this is what we're manipulating

402
00:15:55,339 --> 00:16:00,199
uh the top one is the name of is the

403
00:15:58,920 --> 00:16:03,839
the public

404
00:16:00,199 --> 00:16:06,720
uh known hosts also the public host key

405
00:16:03,839 --> 00:16:08,160
that the server will be known by and

406
00:16:06,720 --> 00:16:11,040
what we've got to do is get the contents

407
00:16:08,160 --> 00:16:13,079
of that into the SSH known host file the

408
00:16:11,040 --> 00:16:13,740
global one

409
00:16:13,079 --> 00:16:16,500
um

410
00:16:13,740 --> 00:16:19,620
slight wrinkle these days is that

411
00:16:16,500 --> 00:16:22,860
um we hash the host names because

412
00:16:19,620 --> 00:16:24,120
security I'm 100 sure why but it's

413
00:16:22,860 --> 00:16:25,680
probably

414
00:16:24,120 --> 00:16:28,260
if they do it it must be a good reason

415
00:16:25,680 --> 00:16:30,300
so we don't tend to keep the host names

416
00:16:28,260 --> 00:16:32,279
in the clear these days anymore

417
00:16:30,300 --> 00:16:35,399
uh the other

418
00:16:32,279 --> 00:16:38,759
key point there is that there's also the

419
00:16:35,399 --> 00:16:41,519
uh so the in the SSH known host file

420
00:16:38,759 --> 00:16:43,440
there's the uh the hashed hostname

421
00:16:41,519 --> 00:16:45,720
there's the key type and the key value

422
00:16:43,440 --> 00:16:48,380
and usually there's a bunch of key types

423
00:16:45,720 --> 00:16:48,380
per host

424
00:16:51,360 --> 00:16:55,259
so this is the ansible Playbook that I

425
00:16:53,220 --> 00:16:57,320
originally had for it

426
00:16:55,259 --> 00:16:57,320
um

427
00:16:57,480 --> 00:17:03,899
and to fit it I've stripped out no I've

428
00:17:02,519 --> 00:17:05,459
stripped out a lot of the comments and

429
00:17:03,899 --> 00:17:06,959
now when I go back to look at it I quite

430
00:17:05,459 --> 00:17:10,140
often find it quite difficult to read

431
00:17:06,959 --> 00:17:12,240
without the comments which is the part

432
00:17:10,140 --> 00:17:14,339
of the whole beef of this talk

433
00:17:12,240 --> 00:17:16,679
um so

434
00:17:14,339 --> 00:17:19,319
this Playbook basically it logs into the

435
00:17:16,679 --> 00:17:21,959
room well it uses uh

436
00:17:19,319 --> 00:17:24,360
it fetches from the remote host

437
00:17:21,959 --> 00:17:26,579
uh the list of files that we need to get

438
00:17:24,360 --> 00:17:27,500
actually prior to that sorry

439
00:17:26,579 --> 00:17:31,620
um

440
00:17:27,500 --> 00:17:33,000
we know the remote host by one of three

441
00:17:31,620 --> 00:17:35,160
ways we know by its fully qualified

442
00:17:33,000 --> 00:17:37,140
domain name it's fully qualified domain

443
00:17:35,160 --> 00:17:38,760
name with a DOT on the end of it and

444
00:17:37,140 --> 00:17:40,020
there's good DNS reasons for that which

445
00:17:38,760 --> 00:17:42,299
I won't go into here

446
00:17:40,020 --> 00:17:44,280
and then it's also known by its IP

447
00:17:42,299 --> 00:17:47,580
address which is the long convoluted

448
00:17:44,280 --> 00:17:49,740
host vars underneath it

449
00:17:47,580 --> 00:17:52,020
we then connect to the server and we get

450
00:17:49,740 --> 00:17:53,520
a list of the public key files that we

451
00:17:52,020 --> 00:17:55,200
want to bring back across

452
00:17:53,520 --> 00:17:58,679
and

453
00:17:55,200 --> 00:18:00,900
with a new slip to fetch them slip

454
00:17:58,679 --> 00:18:03,720
brings them back as a base64 encoded

455
00:18:00,900 --> 00:18:05,880
file because it's not uh it's only ASCII

456
00:18:03,720 --> 00:18:07,440
clean the connection so

457
00:18:05,880 --> 00:18:11,340
um the

458
00:18:07,440 --> 00:18:13,440
we retrieve the the key back actually

459
00:18:11,340 --> 00:18:18,059
we've achieve all parts of the public

460
00:18:13,440 --> 00:18:19,140
file The Next Step then splits it up

461
00:18:18,059 --> 00:18:21,840
got

462
00:18:19,140 --> 00:18:24,059
it's boring me even

463
00:18:21,840 --> 00:18:26,640
it's not boring it's a lot of detail and

464
00:18:24,059 --> 00:18:28,500
what I find and this is

465
00:18:26,640 --> 00:18:30,960
partly what I'm trying to convey is what

466
00:18:28,500 --> 00:18:32,220
I find is that with the comment strip I

467
00:18:30,960 --> 00:18:33,780
get confused as to what it's doing

468
00:18:32,220 --> 00:18:36,900
easily and I have to stare at it for

469
00:18:33,780 --> 00:18:39,059
quite a while and again it's the problem

470
00:18:36,900 --> 00:18:40,740
I'm trying to solve not the problem I'm

471
00:18:39,059 --> 00:18:43,580
trying to solve in the tool that I'm

472
00:18:40,740 --> 00:18:43,580
trying to solve it with

473
00:18:43,620 --> 00:18:48,539
um there's all sorts of chicanery in

474
00:18:45,840 --> 00:18:51,299
here there's the uh the use of default

475
00:18:48,539 --> 00:18:52,980
for the SSH host key because SSH host

476
00:18:51,299 --> 00:18:55,200
key when you do the original pass

477
00:18:52,980 --> 00:18:57,360
through it is blank or it's actually

478
00:18:55,200 --> 00:18:59,039
doesn't exist so if you don't have the

479
00:18:57,360 --> 00:19:00,480
default ansible then throws an error and

480
00:18:59,039 --> 00:19:01,620
saying this key has never been used

481
00:19:00,480 --> 00:19:03,660
before and it's like well I'm just

482
00:19:01,620 --> 00:19:05,400
looping through it typically if I was

483
00:19:03,660 --> 00:19:06,900
doing this in Python I'd just set the

484
00:19:05,400 --> 00:19:08,700
variable beforehand so I know it's

485
00:19:06,900 --> 00:19:10,940
default State and then proceed through

486
00:19:08,700 --> 00:19:10,940
it

487
00:19:14,100 --> 00:19:18,360
uh

488
00:19:16,140 --> 00:19:20,340
this goes on

489
00:19:18,360 --> 00:19:22,440
this then gets to the the

490
00:19:20,340 --> 00:19:25,320
the uh

491
00:19:22,440 --> 00:19:26,820
the uh the end of the wedge of the

492
00:19:25,320 --> 00:19:29,039
matter I guess

493
00:19:26,820 --> 00:19:30,600
um this then actually updates the known

494
00:19:29,039 --> 00:19:32,520
host file

495
00:19:30,600 --> 00:19:35,520
we do this on the local machine we have

496
00:19:32,520 --> 00:19:36,780
delegate 2 on localhost which says for

497
00:19:35,520 --> 00:19:38,280
all the data you've just collected from

498
00:19:36,780 --> 00:19:40,559
all the remote machines we're now just

499
00:19:38,280 --> 00:19:43,200
applying it to in this case it's usually

500
00:19:40,559 --> 00:19:45,840
the laptop that I'm running it from

501
00:19:43,200 --> 00:19:48,900
um we're doing a lot of looping we're

502
00:19:45,840 --> 00:19:50,700
doing looping with products and again if

503
00:19:48,900 --> 00:19:53,220
I don't go back and review the ansible

504
00:19:50,700 --> 00:19:54,419
documentation for I think what is it I'm

505
00:19:53,220 --> 00:19:57,539
doing here

506
00:19:54,419 --> 00:19:59,880
it just slips out we also have in Loop

507
00:19:57,539 --> 00:20:02,220
control where we can sit there and as

508
00:19:59,880 --> 00:20:03,720
it's as edible is looping through it it

509
00:20:02,220 --> 00:20:05,460
sits there and reports only on the

510
00:20:03,720 --> 00:20:07,200
subset of the information otherwise it

511
00:20:05,460 --> 00:20:09,419
just tends to blurt out

512
00:20:07,200 --> 00:20:11,100
great deal of noise onto the screen and

513
00:20:09,419 --> 00:20:14,120
if something breaks you're not actually

514
00:20:11,100 --> 00:20:14,120
sure what's broken

515
00:20:19,740 --> 00:20:24,660
so I come to

516
00:20:23,100 --> 00:20:25,980
there's

517
00:20:24,660 --> 00:20:28,440
there's some ways you can work around

518
00:20:25,980 --> 00:20:31,500
this like you can

519
00:20:28,440 --> 00:20:34,440
deploy you can modularize stuff to tasks

520
00:20:31,500 --> 00:20:36,000
but again once you do so you've got to

521
00:20:34,440 --> 00:20:37,740
sit there and you've got to be aware of

522
00:20:36,000 --> 00:20:40,740
what variables are going into and out of

523
00:20:37,740 --> 00:20:43,740
tasks you've got to track that yourself

524
00:20:40,740 --> 00:20:46,740
you know we have methods in Python where

525
00:20:43,740 --> 00:20:48,120
you can pass in local variables

526
00:20:46,740 --> 00:20:49,620
these are things I don't have to worry

527
00:20:48,120 --> 00:20:51,240
about these are just things that occur

528
00:20:49,620 --> 00:20:53,460
naturally to me when I'm coding it's

529
00:20:51,240 --> 00:20:55,500
like but when I'm coding an ansible I've

530
00:20:53,460 --> 00:20:57,720
got to like no I can't do it that way I

531
00:20:55,500 --> 00:21:00,059
have to track all of this state myself

532
00:20:57,720 --> 00:21:02,160
that's the computer's job I don't want

533
00:21:00,059 --> 00:21:04,740
to be doing the computer's job

534
00:21:02,160 --> 00:21:06,720
I've got other things to do stare at my

535
00:21:04,740 --> 00:21:09,120
ceiling

536
00:21:06,720 --> 00:21:11,700
so

537
00:21:09,120 --> 00:21:13,140
again as I'm obsessing over this

538
00:21:11,700 --> 00:21:14,820
um

539
00:21:13,140 --> 00:21:15,660
I love this quote

540
00:21:14,820 --> 00:21:17,700
um

541
00:21:15,660 --> 00:21:19,320
it's from an old web comment which no

542
00:21:17,700 --> 00:21:21,980
longer exists unfortunately but I've

543
00:21:19,320 --> 00:21:24,480
still talked to Phil Barnes occasionally

544
00:21:21,980 --> 00:21:26,400
Owens the program manager and Desmond's

545
00:21:24,480 --> 00:21:27,539
the programmer

546
00:21:26,400 --> 00:21:29,220
and this is one of the things I

547
00:21:27,539 --> 00:21:32,400
struggled with is like this is part of

548
00:21:29,220 --> 00:21:35,700
my struggle with this whole concept is

549
00:21:32,400 --> 00:21:38,400
I'm still writing code you know and part

550
00:21:35,700 --> 00:21:40,919
of ansible's reason was to be able to be

551
00:21:38,400 --> 00:21:41,760
usable by the every man sorry every

552
00:21:40,919 --> 00:21:44,340
person

553
00:21:41,760 --> 00:21:46,200
sorry that's gender inappropriate

554
00:21:44,340 --> 00:21:49,140
um usable by everyone

555
00:21:46,200 --> 00:21:51,539
it should be relatively simple to put

556
00:21:49,140 --> 00:21:53,340
together and for simple cases it is

557
00:21:51,539 --> 00:21:54,900
ansible was quite simple but when you

558
00:21:53,340 --> 00:21:57,840
get to start doing complex things with

559
00:21:54,900 --> 00:21:59,760
it it just for me and it might not be

560
00:21:57,840 --> 00:22:01,200
everyone here but for me personally it

561
00:21:59,760 --> 00:22:05,000
goes off the routes fairly quickly when

562
00:22:01,200 --> 00:22:07,620
I have to burn brain CPU Cycles on

563
00:22:05,000 --> 00:22:09,059
details that I'd rather be burning

564
00:22:07,620 --> 00:22:11,400
elsewhere

565
00:22:09,059 --> 00:22:14,820
so I was concerned

566
00:22:11,400 --> 00:22:16,799
when moving to python when moving to Pi

567
00:22:14,820 --> 00:22:18,179
infra but not just paying for it I was

568
00:22:16,799 --> 00:22:20,640
looking at other programs in the same

569
00:22:18,179 --> 00:22:22,740
space I was concerned am I just swapping

570
00:22:20,640 --> 00:22:23,760
one program from another one problem for

571
00:22:22,740 --> 00:22:26,580
another

572
00:22:23,760 --> 00:22:29,520
but I don't feel that's quite the case

573
00:22:26,580 --> 00:22:31,679
um again it's more the power of

574
00:22:29,520 --> 00:22:35,659
abstraction that you get from Python and

575
00:22:31,679 --> 00:22:35,659
being able to express problems in Python

576
00:22:35,820 --> 00:22:40,440
so I've converted that script I'm sorry

577
00:22:38,820 --> 00:22:41,220
that Playbook

578
00:22:40,440 --> 00:22:42,740
um

579
00:22:41,220 --> 00:22:45,780
Pi infra

580
00:22:42,740 --> 00:22:48,539
and as I said point for basically the

581
00:22:45,780 --> 00:22:52,080
inventory and the deploy scripts what

582
00:22:48,539 --> 00:22:53,400
its versions of playbooks are

583
00:22:52,080 --> 00:22:56,100
um

584
00:22:53,400 --> 00:22:58,799
for all intents and purposes python so

585
00:22:56,100 --> 00:23:01,620
first up is the inventory

586
00:22:58,799 --> 00:23:03,900
um for so for my example

587
00:23:01,620 --> 00:23:06,059
um we just this is the inventory.pi that

588
00:23:03,900 --> 00:23:07,799
we pass into paying for it's just

589
00:23:06,059 --> 00:23:09,780
basically a list of servers normally

590
00:23:07,799 --> 00:23:10,860
it's just a straight list of servers I

591
00:23:09,780 --> 00:23:14,820
actually

592
00:23:10,860 --> 00:23:17,400
like again I like my running dialogue as

593
00:23:14,820 --> 00:23:18,600
things are running to be reasonably

594
00:23:17,400 --> 00:23:19,620
cleaned so you don't have to pick

595
00:23:18,600 --> 00:23:21,179
through and try and pick through the

596
00:23:19,620 --> 00:23:22,320
error messages so I like my server name

597
00:23:21,179 --> 00:23:27,299
short

598
00:23:22,320 --> 00:23:29,520
and my SSH known assh hostname's long so

599
00:23:27,299 --> 00:23:31,520
basically when playing for runs it knows

600
00:23:29,520 --> 00:23:33,740
that the host that it has to connect to

601
00:23:31,520 --> 00:23:36,120
known as server one is actually

602
00:23:33,740 --> 00:23:37,220
server1.example.com and so on you can

603
00:23:36,120 --> 00:23:41,100
pass other

604
00:23:37,220 --> 00:23:44,159
uh variables in sorry you can assign

605
00:23:41,100 --> 00:23:45,539
other variables into this these are just

606
00:23:44,159 --> 00:23:48,000
the ones that I needed for the purpose

607
00:23:45,539 --> 00:23:49,620
the

608
00:23:48,000 --> 00:23:52,799
upshot this though is that it's

609
00:23:49,620 --> 00:23:55,140
executable python code so you can use

610
00:23:52,799 --> 00:23:58,020
something like requests and

611
00:23:55,140 --> 00:23:59,640
grab your inventory from you know a

612
00:23:58,020 --> 00:24:01,740
convenient Json server that's located

613
00:23:59,640 --> 00:24:03,539
nearby but has all your stuff there if

614
00:24:01,740 --> 00:24:05,820
it doesn't come back in the format you

615
00:24:03,539 --> 00:24:09,000
know that's that's like this you can

616
00:24:05,820 --> 00:24:10,640
Mudge it within the script you don't

617
00:24:09,000 --> 00:24:14,220
have to write

618
00:24:10,640 --> 00:24:15,840
a uh a separate inventory source as you

619
00:24:14,220 --> 00:24:19,020
would do with ansible you can actually

620
00:24:15,840 --> 00:24:19,919
do it it's it's python you can do it as

621
00:24:19,020 --> 00:24:23,039
is

622
00:24:19,919 --> 00:24:24,900
uh so paying for when running this it's

623
00:24:23,039 --> 00:24:27,299
just concerned about prod servers it can

624
00:24:24,900 --> 00:24:30,000
it basically any uh variable that's

625
00:24:27,299 --> 00:24:32,100
sorry any yeah variable that's set in

626
00:24:30,000 --> 00:24:34,980
the module space it uses

627
00:24:32,100 --> 00:24:36,900
uh as the inventory so you can have prod

628
00:24:34,980 --> 00:24:38,220
servers you could have Dev servers and

629
00:24:36,900 --> 00:24:39,780
they'd be selectable from the command

630
00:24:38,220 --> 00:24:41,460
line as well or it'll just execute

631
00:24:39,780 --> 00:24:43,500
against the lot

632
00:24:41,460 --> 00:24:46,380
uh

633
00:24:43,500 --> 00:24:49,020
uh I'll come back to it local shortly uh

634
00:24:46,380 --> 00:24:50,640
but at local as I said before it's uh

635
00:24:49,020 --> 00:24:54,620
basically saying I want to connect to

636
00:24:50,640 --> 00:24:54,620
the local server not virus sh

637
00:24:55,039 --> 00:24:59,159
but I'm sorry I won't come back to it

638
00:24:57,480 --> 00:25:01,020
just I'll come back to it right now uh

639
00:24:59,159 --> 00:25:02,820
the reason we do that is that Pi infra

640
00:25:01,020 --> 00:25:05,100
doesn't have the concept of a delegate

641
00:25:02,820 --> 00:25:06,720
to so when previous Playbook had a

642
00:25:05,100 --> 00:25:08,520
delegate to to say I need you to run

643
00:25:06,720 --> 00:25:11,520
this on the Local Host

644
00:25:08,520 --> 00:25:13,500
time for it doesn't have that but it

645
00:25:11,520 --> 00:25:16,400
just lets you run your playbook against

646
00:25:13,500 --> 00:25:16,400
localhost anyway

647
00:25:17,340 --> 00:25:22,500
uh and we have group files as well

648
00:25:20,640 --> 00:25:24,480
same as ansible so almost feature

649
00:25:22,500 --> 00:25:27,000
complete for that and this is basically

650
00:25:24,480 --> 00:25:28,440
saying this is the SSH user which isn't

651
00:25:27,000 --> 00:25:31,440
my user in case any of you thinking

652
00:25:28,440 --> 00:25:33,179
about trying to hack a new my framework

653
00:25:31,440 --> 00:25:35,580
um and the key

654
00:25:33,179 --> 00:25:38,039
which one of you have either but um be

655
00:25:35,580 --> 00:25:40,820
sure to commit that to uh GitHub later

656
00:25:38,039 --> 00:25:42,960
uh

657
00:25:40,820 --> 00:25:44,640
interesting bug and it's probably worth

658
00:25:42,960 --> 00:25:47,400
keeping and this will be uploaded to the

659
00:25:44,640 --> 00:25:50,460
repository um interesting but paramico

660
00:25:47,400 --> 00:25:52,980
and modern versions of Debian and Ubuntu

661
00:25:50,460 --> 00:25:54,779
SSH servers

662
00:25:52,980 --> 00:25:57,120
can't negotiate

663
00:25:54,779 --> 00:25:59,400
can't play nicely there's a couple of

664
00:25:57,120 --> 00:26:01,919
algorithms there that it seems to it

665
00:25:59,400 --> 00:26:03,600
tries to use and then dies badly but

666
00:26:01,919 --> 00:26:05,880
it's a simple matter to disable those

667
00:26:03,600 --> 00:26:08,720
and say just exclude these algorithms in

668
00:26:05,880 --> 00:26:11,820
your initial setup and negotiation

669
00:26:08,720 --> 00:26:15,240
and then go nuts

670
00:26:11,820 --> 00:26:17,880
um the last line uh

671
00:26:15,240 --> 00:26:22,440
that's also a paramico thing uh paramico

672
00:26:17,880 --> 00:26:27,059
just uses your local your personal home

673
00:26:22,440 --> 00:26:30,980
directory SSH known hosts by default

674
00:26:27,059 --> 00:26:30,980
holy duly I am so sorry

675
00:26:31,679 --> 00:26:35,760
have you wiped that five minute thing in

676
00:26:33,059 --> 00:26:37,080
me yet you did sorry about that um speed

677
00:26:35,760 --> 00:26:38,700
run

678
00:26:37,080 --> 00:26:40,500
I apologize for this because this is

679
00:26:38,700 --> 00:26:43,200
what I swear I wouldn't do

680
00:26:40,500 --> 00:26:44,820
ah the deploy script proper

681
00:26:43,200 --> 00:26:47,340
ly to the deploy script

682
00:26:44,820 --> 00:26:49,740
uh the deploy script only needs to

683
00:26:47,340 --> 00:26:52,080
execute on the local host and this is as

684
00:26:49,740 --> 00:26:54,360
you can see pure python

685
00:26:52,080 --> 00:26:55,919
um I grabbed some bits from paying for

686
00:26:54,360 --> 00:26:57,419
it that I need I also grabbed some bits

687
00:26:55,919 --> 00:27:00,360
from help script that I need

688
00:26:57,419 --> 00:27:03,240
and then I sit there and I go right I

689
00:27:00,360 --> 00:27:05,220
need to slurp up all the keys from all

690
00:27:03,240 --> 00:27:07,320
the hosts that aren't me

691
00:27:05,220 --> 00:27:08,940
hence the I name does not equal that

692
00:27:07,320 --> 00:27:10,320
local and the ones I've successfully

693
00:27:08,940 --> 00:27:13,320
connected to so if there's any servers

694
00:27:10,320 --> 00:27:16,580
offline it skips over them as well it

695
00:27:13,320 --> 00:27:19,919
grabs the connection IP the SSH hostname

696
00:27:16,580 --> 00:27:22,400
and then executes that fact SSH known

697
00:27:19,919 --> 00:27:25,020
host keys

698
00:27:22,400 --> 00:27:26,460
it then sits there and creates all the

699
00:27:25,020 --> 00:27:27,720
host names based on that so that's the

700
00:27:26,460 --> 00:27:30,720
fully qualified domain name with them

701
00:27:27,720 --> 00:27:34,440
without the dot and the connection IP

702
00:27:30,720 --> 00:27:35,340
and then bunks that all into a list of

703
00:27:34,440 --> 00:27:37,860
lists

704
00:27:35,340 --> 00:27:41,159
and then it calls this is H known host

705
00:27:37,860 --> 00:27:44,159
subset which sits there and says passes

706
00:27:41,159 --> 00:27:45,179
that all off to the SSH host module

707
00:27:44,159 --> 00:27:47,220
um

708
00:27:45,179 --> 00:27:48,500
which is the operation that Pi infra

709
00:27:47,220 --> 00:27:51,120
runs

710
00:27:48,500 --> 00:27:53,820
so it does that tells you the name of

711
00:27:51,120 --> 00:27:56,760
the path the path to the message you

712
00:27:53,820 --> 00:27:59,520
want to update name is the descriptive

713
00:27:56,760 --> 00:28:01,039
field that comes up as it's running so

714
00:27:59,520 --> 00:28:03,059
you can

715
00:28:01,039 --> 00:28:04,860
so you know what it's doing in the play

716
00:28:03,059 --> 00:28:08,039
and of course we need pseudo access to

717
00:28:04,860 --> 00:28:10,340
do that so just say give me sudo or give

718
00:28:08,039 --> 00:28:10,340
me death

719
00:28:17,400 --> 00:28:22,500
I have cheated slightly but to be fair

720
00:28:20,580 --> 00:28:24,900
answer we'll start at it

721
00:28:22,500 --> 00:28:27,240
ansible has a lot of building facts and

722
00:28:24,900 --> 00:28:29,000
modules that we can leverage on

723
00:28:27,240 --> 00:28:32,279
um

724
00:28:29,000 --> 00:28:33,539
that we do not have so I've actually had

725
00:28:32,279 --> 00:28:37,279
to write

726
00:28:33,539 --> 00:28:37,279
two facts and an operation

727
00:28:37,740 --> 00:28:41,279
I'm going to really speak on through

728
00:28:39,419 --> 00:28:42,960
this

729
00:28:41,279 --> 00:28:44,520
this is a fact

730
00:28:42,960 --> 00:28:46,860
these are very simple

731
00:28:44,520 --> 00:28:48,539
these are so simple I've spent no time

732
00:28:46,860 --> 00:28:50,700
thinking about writing these as it

733
00:28:48,539 --> 00:28:52,380
should be again this is the problem that

734
00:28:50,700 --> 00:28:54,659
I was trying to solve I do not want to

735
00:28:52,380 --> 00:28:55,860
be thinking about how to solve these

736
00:28:54,659 --> 00:28:57,299
problems I want to be thinking about the

737
00:28:55,860 --> 00:29:00,539
problem that I'm trying to solve in

738
00:28:57,299 --> 00:29:03,500
ansible sorry the automation problem not

739
00:29:00,539 --> 00:29:03,500
in ansible no

740
00:29:03,840 --> 00:29:08,580
um so a fact is basically we derive it

741
00:29:07,080 --> 00:29:11,539
from what's called fact base which is

742
00:29:08,580 --> 00:29:13,980
just the you know the the superclass

743
00:29:11,539 --> 00:29:15,539
there's two methods that we need to

744
00:29:13,980 --> 00:29:18,240
implement command which in this case

745
00:29:15,539 --> 00:29:19,559
we're actually there is a helper script

746
00:29:18,240 --> 00:29:21,240
that applying for supplies that just

747
00:29:19,559 --> 00:29:22,860
lets us go hey we just want to cap the

748
00:29:21,240 --> 00:29:25,020
output of all those files

749
00:29:22,860 --> 00:29:26,520
and then there's a process which it runs

750
00:29:25,020 --> 00:29:29,820
which basically grabs the output of all

751
00:29:26,520 --> 00:29:32,039
those files which is executed across SSH

752
00:29:29,820 --> 00:29:33,240
brought brought back and then we just go

753
00:29:32,039 --> 00:29:36,000
hey

754
00:29:33,240 --> 00:29:37,679
do with it whatever and what I'm doing

755
00:29:36,000 --> 00:29:39,000
with it here is basically just returning

756
00:29:37,679 --> 00:29:40,980
a dictionary that says this is the key

757
00:29:39,000 --> 00:29:43,020
type this is the key value imagine says

758
00:29:40,980 --> 00:29:43,860
one minute

759
00:29:43,020 --> 00:29:45,539
um

760
00:29:43,860 --> 00:29:47,820
but this is the thing is that the fact

761
00:29:45,539 --> 00:29:49,140
is structured in a way that helps me

762
00:29:47,820 --> 00:29:51,299
solve the problem that I'm trying to

763
00:29:49,140 --> 00:29:54,320
solve

764
00:29:51,299 --> 00:29:54,320
and skip over the next fact

765
00:29:54,480 --> 00:29:59,279
should note right

766
00:29:57,539 --> 00:30:02,700
and then this is the operation the

767
00:29:59,279 --> 00:30:04,860
operation is basically it grabs all the

768
00:30:02,700 --> 00:30:07,520
data that you've collected and then sits

769
00:30:04,860 --> 00:30:07,520
there and

770
00:30:08,100 --> 00:30:11,100
where what I'm doing here is irrelevant

771
00:30:09,720 --> 00:30:12,000
skip that

772
00:30:11,100 --> 00:30:13,860
um

773
00:30:12,000 --> 00:30:15,840
but we need to basically say in the SSH

774
00:30:13,860 --> 00:30:19,320
known case files we need to delete Keys

775
00:30:15,840 --> 00:30:20,760
we need to append Keys we sit there and

776
00:30:19,320 --> 00:30:22,919
we work over

777
00:30:20,760 --> 00:30:24,179
we do basically technical Mumbo jumper

778
00:30:22,919 --> 00:30:26,340
don't need to care about it but what we

779
00:30:24,179 --> 00:30:27,779
do at the end is we go for the lines

780
00:30:26,340 --> 00:30:29,399
that I want to lead out of the key file

781
00:30:27,779 --> 00:30:31,380
this is the said command I want to

782
00:30:29,399 --> 00:30:32,700
execute for all the keys I need to

783
00:30:31,380 --> 00:30:34,980
append these are the keys I need to

784
00:30:32,700 --> 00:30:36,600
append and if I've made any of those

785
00:30:34,980 --> 00:30:39,840
changes

786
00:30:36,600 --> 00:30:41,580
it's almost up if I if I have made any

787
00:30:39,840 --> 00:30:43,860
of those changes then I need to use SSH

788
00:30:41,580 --> 00:30:46,020
Keygen to sit there and reformat the

789
00:30:43,860 --> 00:30:49,440
output to rehash the output

790
00:30:46,020 --> 00:30:50,760
and I think I'm done and I apologize

791
00:30:49,440 --> 00:30:53,539
um

792
00:30:50,760 --> 00:30:53,539
30 seconds

793
00:30:53,700 --> 00:30:56,760
right

794
00:30:54,659 --> 00:30:57,600
there will be no questions

795
00:30:56,760 --> 00:31:00,600
um

796
00:30:57,600 --> 00:31:02,220
if you do two months fee operations to

797
00:31:00,600 --> 00:31:04,140
Pine for when you run it you get to see

798
00:31:02,220 --> 00:31:06,899
the commands that it spits out over the

799
00:31:04,140 --> 00:31:08,640
top commands are the facts that it's the

800
00:31:06,899 --> 00:31:10,860
shell commands that it executes for the

801
00:31:08,640 --> 00:31:13,500
facts the bottom commands are the ones

802
00:31:10,860 --> 00:31:17,100
that the output of the operation that I

803
00:31:13,500 --> 00:31:19,080
ran and that shows you yep exactly what

804
00:31:17,100 --> 00:31:21,059
it is it's doing if you use three minus

805
00:31:19,080 --> 00:31:22,620
fees it actually sits there and shows

806
00:31:21,059 --> 00:31:24,360
you the command the response from those

807
00:31:22,620 --> 00:31:26,840
commands coming back this is great for

808
00:31:24,360 --> 00:31:26,840
debugging

809
00:31:27,179 --> 00:31:32,399
any Lessons Learned yes time up talks

810
00:31:29,340 --> 00:31:34,380
more properly uh no questions

811
00:31:32,399 --> 00:31:36,860
this time we're really done I apologize

812
00:31:34,380 --> 00:31:36,860
sorry

813
00:31:37,620 --> 00:31:41,240
honestly pretty good speed run