►
From YouTube: TGI Kubernetes 115: Deploying an app
Description
Come hang out with Bryan Liles as he does a bit of hands on hacking of Kubernetes and related topics. Some of this will be Bryan talking about the things he knows. Some of this will be Bryan exploring something new with the audience. Come join the fun, ask questions, comment, and participate in the live chat!
A
All
right
hello,
good
afternoon
evening
and
morning,
if
you're
in
like
a
morning,
if
you're
my
bed,
sorry
about
the
little
bit
of
echo
there
but
good
board
good
afternoon
good
evening
and
good
morning,
if
you're
in
Hawaii,
so
I'm,
Bryan,
Liles
and
I'm
doing
TGI
cakes
today.
So
you
know,
TGI
kubernetes
been
doing
this
for
a
long
time
now.
This
is
the
hundred
and
15th
episode,
and
this
episode
will
be
a
little
bit
different.
I
wanted
to
I
wanted
to
actually
do
something
like
something
that
we
haven't
seen
before,
maybe
like.
A
B
B
A
Geared
bear
who
was
here
early,
hello
and
Bren,
hi
and
Maddie,
and
Maz
hello
and
Shiva
Kumar
and
Rory
and
carlesha
nice
to
see
you
and
Tim
and
Shekhar
and
party
band
and
Rodolfo
good
to
see
you
Erik
Brann
Dave
joy,
oh
my
gosh
there's!
So
many
people
in
this
chat
all
right,
I
can't
say
hello
to
all
you
all
right
now,
but
know
this
I'm
happy.
You
all
are
here.
So,
as
with
tradition,
let's
do
the
news!
B
A
Alright,
so
let's
actually
go
through
and
have
a
look
at
the
Week
in
Review,
so
at
the
top
of
our
list
is
actually
some
good
news
and
I'm
gonna
give
a
big
shout
out
to
Lachlan
Evanson
for
joining
the
kubernetes
steering
committee.
This
is
actually
you
know.
This
is
a
good
point
for
the
kubernetes
land,
where
we
have
people
who
have
have
actually
grown
up
around
kubernetes
and
now
there
they
are
actually
leading
the
charge
to
steer
us
where
we're
going
to
be
in
the
future.
A
So,
let's
see
what
other
kind
of
needs
we
have
so
first
link
here
is
what
happens
when
kubernetes
does
what
so,
what
this
is
is
just
the
explanation
of
really
what
kubernetes
is
doing
and
and
how
it's
actually
making
it
happen.
A
pretty
good
read
here.
This
is
definitely
already
in
my
book
parks
and
I've
gone
through
it
a
little
bit.
A
The
next
release
is
that
helm
is
at
three
point
two
point
zero.
This
is
a
I
think
there
were
security
fix
in
here.
Yes,
a
number
of
security
fixes
in
here,
but
actually
there's
something
else
going
on
with
helm.
Helm
is
pretty
close.
Oh
I'm
sorry
I'm
going
to
share
my
screen.
There
you
go
now.
Y'all
can
see
my
screen.
A
A
A
Helm
is
is
getting
close
to
graduating
in
the
CNC
F
and
joining
a
very
small
group
of
projects.
This
is
actually
a
big
deal.
This
is
huge
when
it
comes
to
the
stability
of
the
community
and
the
project
itself.
You
know
how
am
was
one
of
the
early
big
kubernetes
spin-off
project,
so
you
could
say,
and
now
that
they're
actually
going
to
make
release
status
or
graduated
status
on
CN
CF.
A
So
next
up
it
is
a
boom
to
the
2004
has
been
released.
This
is
a
big
deal,
because
every
two
years,
Ubuntu
actually
releases
an
LTS
and
an
LTS
means
that
its
long-term
support
so
a
lot
of
people.
We
started
way
back
on
Thabo
for
and
1404
and
1604,
and
then,
while
people
are
using
1804.
So
now
you
know
with
2004.
A
There's
gonna
be
two
years
of
stability
for
our
actually
two
years
plus
of
stability
and
a
boom
tool,
and
next
item
here
is
the
CSP
is
open
for
cube
con
in
Boston
and
go
to
this
page,
my
Prem,
my
successors,
constants
and
and
Stephen
are
going
to
put
together
a
great
show,
assuming
we
can
all
get
out
of
the
house.
Hopefully
we
can
all
make
it
out
of
the
house
by
November
and
make
it
up
to
Boston
the.
A
Is
a
great
video
by
one
of
my
co-workers,
Carly
Carly
Co
campus-
and
this
is
a
video
about
extending
the
kubernetes
api.
I
actually
did
watch
most
of
this
video
already,
but
I
would
like
to
give
a
shout
out
shout
out
to
carlesha,
because
not
only
is
she
creating
these
videos
she's
also
doing
things
for
her
people
crisis
is
not
a
Native
American.
She,
our
native
person
from
United
States.
A
She
grew
up
in
in
Brazil
and
speaks
Portuguese
and
a
while
ago
and
I
speak
tiny
little
bit
of
Portuguese
I
actually
sat
there
and
watched
most
of
the
video
where
they
actually
did
a
kubernetes
call
in
Portuguese.
It's
great
to
see
people
doing
these
things
and
other
languages,
because
even
though
English
is
prevalent,
it's
not
the
only
thing
we
speak
out
here.
A
The
next
item
on
the
list
is
the
kubernetes
podcast
celebrates
is
hundreds
with
Parris
Pittman
and
I.
Just
wanna
give
a
good
shout
out
to
Parris.
Pittman
I
grew
up
down
the
street
from
Paris
when
she
lived
in
Baltimore
many
years
ago,
no
pears
for
a
very
long
time.
So
it's
great
to
see
her
come
from
where
she
was
to
all
these
great
things
that
she's
done
now.
Community
Builder
is
definitely
a
is
definitely
a
title
that
she
wears
well.
A
So
the
next
item
on
the
list
is
the
contributor
communications
team
for
kubernetes,
and
this
is
a
new
effort.
So
really,
when
you
get
to
a
project,
the
size
of
kubernetes
thinking
about
building
better
communications
process
is
actually
pretty
important
at
this
point.
So
now
we
have
great
set
of
volunteers
actually
putting
this
together.
So
please
take
a
look
at
this
link
to
see
what
they
have
going
on.
A
The
next
item
item
here
is
a
cluster
API.
V1
alpha-3
has
been
featured
on
the
kubernetes
blog
as
of
two
days
ago,
and
for
those
of
you
don't
know
what
cluster
API
is
think
about
this
Chelsie
Hightower,
Joe,
Beda
and
I
like
to
say
this
all
the
time
as
well
is
that
kubernetes
is
a
platform
for
building
platforms,
so
with
cluster
API
you
can
actually
create
clusters
by
typing.
Nip
you
lating
resources
within
kubernetes.
This
is
actually
a
great
idea.
A
Whenever
you
think
about
hey,
we
have
lots
of
virtual
machine
compute
and
we
don't
need
huge
clusters.
We
can
actually
have
smaller
clusters.
So
this
is
a
it's
good
to
see
they're
getting
to
this
level
of
stability,
and
the
neat
thing
is
is
that
you
can
build.
There's,
there's
lots
of
implementations
of
this
where
I
mean
I
worked
at
VMware,
and
we
have
it
here
at
VMware,
but
Asher
and
AWS,
and
and
quite
a
few
more
so
it's
good
to
see
that
this
team
has
been
able
to
move
this
from
just
an
idea.
A
So
it's
good
to
see
that
this
project
has
grown
to
the
point
where,
where
we
feel
that
the
community
that
we
really
should
be
donating
this
to
the
community
and
the
community
should
definitely
have
all
the
inputs
so
with
that,
let's
get
down
to
it.
So
this
is
an
interesting
place,
so,
like
I
said
before,
I
wanted
to
do
something
a
little
bit
different
for
this
episode
of
kubernetes
I.
A
Think
Pat
clackin
August
of
last
year,
I
talked
about
some
software
contour
that
I
are
not
contour
octant,
that
I
wrote
and
I
didn't
like
that,
and
and
really
I
went
heads
down
and
did
a
whole
bunch
of
other
things,
but
I
kind
of
realized.
It's
time
for
me
to
come
talk
about
something
else,
and
what
I
wanted
to
talk
about
was
there
are
actually
let
me
actually
show
you
this
on.
A
So
I
will
pull
this
over
here
and
I
will
go
back
a
page,
so
we're
shooting
the
115
episode
of
t
GI
kubernetes
right
now.
There
are
a
hundred
plus
episodes
of
just
different
types
of
content
and
and
what
I
wanted
to
see,
because,
yes,
I
am
an
infrastructure
person
but
I'm
also
an
application
person.
I
wanted
to
actually
just
say
why
don't
we
just
take
some
of
these
ideas
that
were
in
in
these
previous
videos
and
why.
B
A
We
just
work
on
doing
something
basic,
which
we
will
show
is
actually
not
as
basic
as
we
think
and,
let's
think
about
deploying
an
application
of
kubernetes
and
before
everyone
says.
Well,
no
Brian,
that's
easy
yeah!
It
is,
if
you
understand,
all
the
parts,
but
there
is
actually
a
very
important
thing
about
understand
all
the
parts.
So
what
is
the
first
thing
we're
going
to
do
before
we
deploy
an
application
to
production
or
anything
into
kubernetes?
B
A
One
and
I'll
show
you
how
it
works,
because
this
is
the
actual
demo
that
we're
going
to
use
throughout
the
entire
session.
So
I
wrote
this
little
thing.
It's
actually
not
much,
and
it's
just
this
little
go
app.
That's
going
to
be
the
application
that
I'm
going
to
use
for
the
the
rest
of
the
session.
So
let
me
actually
run
it
real
quick.
So
if
I
just
go
go
run
main
god,
nothing
shows
up
like
I
did
not
put
in
the
logging
stuff
in
here,
but
let
me
actually
show
you
what
it
does.
A
It's
a
Cal
State
application
and
why
is
it
a
Cal
say
a
patient?
Well,
because
I
was
thinking
well,
I
could
make
a
web
service,
but
then
you
have
to
do
curls
and-
and
things
like
that,
so
I
was
like
I'll
make
something
you
show
up
in
the
web
page
and
keep
it
fun.
So
everyone
likes
Cal
say
so.
This
is
Cal
say
as
a
web
application
and
it's
pretty
static.
But
the
neat
part
about
this
application
and
will
be
important
for
this-
is
that
I
can
actually
do
things
like
change
the
color.
A
So
we'll
change
the
background
color
to
black,
we'll
change
the
foreground,
color
to
white
and
we'll
change
the
message
to
CGI
K
and
we'll
change
the
type
to
now.
Let's
keep
it
like
that,
we'll
just
keep
it
as
cows,
so
app
is
running
again,
but
the
app
is
configurable,
the
colors
and
the
message
for
the
most
part
configurable.
So
now
we
have
a
life
white
cow
on
a
black
background,
so
we're
going
to
so
now.
We're
we've
already
made
the
decision
here.
A
Our
teams
are
using
kubernetes
we're
not
going
to
talk
about
why
kubernetes
right
this
second,
but
we're
gonna
talk
about
how
do
we
get
this
into
the
cluster
and
let's
go
through
some
of
these
steps
because
sometimes
I
feel
like
we,
we
jump
over
that
and
we
assume
that
everyone
has
this
experience
to
actually
do
this.
So
let
me
actually
go
back
clip
this
and
the
first
thing
we're
going
to
want
to
do
is
write
dr.
files,
and
actually
this
is
a
huge
pain
point.
A
This
is
why
I
wanted
to
bring
it
up,
how
many
people
can
actually
write,
docker
files
from
the
command
line
and
and
that's
and
that's
just
a
very
good
question
and
something
that
I've
been
using
docker
since
the
beginning
and
even
I
struggle
with
it.
Sometimes
so.
I'm
gonna
launch
yourself
in
my
favorite
editor,
just
because
I
want
to.
A
File
and
because-
and
we
won't
add
it
and
because
it's
let
me
go
here,
make
another
one
of
these,
because
I
don't
want
to
bore
you
all
with
my
typing
I
actually
created
one
I
created
all
I
created
this
beforehand,
because
this
could
definitely
be
a
lot
of
typing
and
that's
not
what
this
is
about.
This
is
an
app
about
Brian,
doing
typing
slowly,.
B
A
Not
much
going
on
here
so
what
we
have
in
this
docker
file.
This
is
a
two-stage
docker
file.
It
builds
in
one
stage
and
then
it
builds.
It
builds
a
smaller
scratch
image
in
another.
So
what
I
can
actually
do
is
like
just
go,
build
this
thing
and
we'll
call
it
Brian
Elbe,
let's
say
and
we'll
just
build:
no,
not
right
directory.
A
We
go
so
what's
going
on
here,
and
this
will
be
pretty
quick.
This
app,
that's
not
very
big,
is
that
we
are
going
ahead
and
we
are
building
this
image
so
just
to
show
you
that
we're
building
its
image,
I'll
go
docker,
run
and
I'll
just
map
this
thing
locally
and
make
it
go
away.
Just
to
let
you
know
so
inside
of
my
docker,
it
worked.
It
works
again
and
that's
just
docker,
but
this
is
not
important
to
us.
A
Next
is:
let's
go
up
to
another:
let's
actually
make
this
docker
file
just
a
little
bit
better
and
once
again,
I
pretty
typed
all
this,
because
I
didn't
want
to
type
it
all
in
and
if
you're
paying
attention,
there's
one
there's
literally
one
line
change
in
here:
I
change
this
image
from
scratch
to
this
image
from
Google
Cloud,
where
they
have
this
idea
of
destroy
lists.
So
let
me
show
you
this
and
although
GCP
destroy
lists
and
what
distro
list
is
it's
there
are
there.
A
They
are
base
images
that
can
be
used
by
anyone
to
actually
serve
all
the
things
and
the
reason
I
like
this
is
because
there's
things
because
sometimes
there's
actually
a
big
discrepancy
from
having
nothing
to
having
maybe
a
full
boo
to
or
full
Debian.
So,
instead
of
doing
that,
what
distro
list
does
is
they?
Actually,
someone
has
gone
through
and
paired
all
these
things
down.
So
let
me
actually
show
you
what
I
mean
by
this.
So
if
I
go
doctor
images,
grat
nose
lips
is
latest,
that's
not
going
to
show
anything.
A
B
A
It
says
an
issue
here:
copy
from
builder
workspace
bill,
pull
access
denied
for
builder,
pasta,
Tory's
not
exist.
Let
me
actually
go
back
to
my
source.
Oh
that's
the
problem,
so
third
time's
a
charm.
It
helps
whenever
you
actually
have
all
the
stage
lines
in
there.
So
remember
before
that
it
was
30
about
30
Meg's.
This
image
that
I'm
going
to
show
you
now
so
look
it
it's
actually
it's
about
the
same
size
to
tell
you
the
truth.
So
let
me
clear
this.
A
Oops
yeah,
this
is,
let
me
see
my
analysis
web
say
so
it's
about
the
same
size,
so
first,
the
first
pointer
out
first
thing
I
want
to
point
out
is
to
say
that
if
you're
a
goal
and
think
about
you
scratch,
if
you
really
must,
but
there
are
other
small
slim
images
and
actually
let
me
go
through
the
I
just
saw
something
off
the
corner
of
my
eye.
Someone
says
it
Avinash
says
he
likes
my
my
my
eggshell
setup,
interesting
story:
I
use
fish.
A
So
no,
no,
all
those
neat
things
that
you
see
and
in
in
like
zet,
shell
or
Z
shells,
whatever
you
want
to
call
it
all.
Those
things
were
copied
from
one
fish.
So
that's
what
I'm
doing
here
so
but
back
to
my
point,
is
that
thinking
about
using
so
using
the
dishes
from
Google
and
Google
cloud
is
actually
a
pretty
good
idea,
but
actually,
let's
think
about
something
else.
So
it's
2020
and
if
I
go
back
to
my
source
here
and
I'll,
make
it
move
this
back
over
here
and
I'll
make
this
slightly
bigger.
A
A
Things
like
doctor
files,
they're
fine,
but
that
knowledge
is
not
transferred
well
throughout
organizations.
Oh
another
thing,
I
wanted
to
introduce,
which
is
actually
been
covered
in
a
video
of
previous
TGI
K.
Is
this
whole
idea
of
cloud
native,
build
packs
and
what
cognitive
build
packs
are
there's
a
lot
of
words
on
this
page,
but
I'll
summarize
it
for
you
is
to
say
that,
given
the
type
of
project,
we
should
be
able
to
make
some
pretty
good
guesses
about
how
it
should
be
built,
and
we
should
be
able
to
just
build
them.
A
If
this
is
go,
we
should
be
able
to
build
it
in
a
certain
way.
If
this
is
java,
like
a
java
web
application,
we
should
be
able
to
build
it
or
a
ruby,
app
or
a
Python
or
something
javascript
or
typescript,
and
where
this
comes
from
is
actually
Heroku
Heroku
many
years
ago,
Wow
a
long
time
ago,
actually
pioneered
this
idea
build
packs.
So
you
could
just
take.
A
You
can
just
point
Heroku
at
your
directory
and
you
would
get
you
could
do
something
like
git
push
Heroku
master
and
it
would
work
what
glob
cloud
native
build
packs.
Do
is
something
very
similar,
but
on
the
outside,
instead
of
just
pushing
to
Heroku
what
it
actually
does
does
does
what
it
does
is
it.
It
actually
creates
a
container
image.
So
let
me
show
you
that,
and
let
me
show
you
how
I'm
gonna
actually
do
this
and
I'm
not
gonna,
have
any
any
setup.
A
There's
a
little
bit
of
mojo
going
on
here
and
I'll.
Explain
all
the
pieces
as
I'm
doing
it.
So,
within
the
cloud
native
bill
packs,
there
is
a
binary
called
pack,
and
pack
has
a
few
things
in
it.
It
can
build
and
I,
don't
even
know
what
rebase
does
but
I'm
sure
it's
pretty
cool.
They
can
inspect
an
image
and
show
you
what's
inside
of
it
cloud
native,
build
packs
is
actually
combined
with
these
ideas
of
builders
and
inside
of
builders.
A
There's
build
packs
and
what
you
can
do
is
you
can
basically
layer
things
on
top
and
you
can
create
the
type
of
image
that
maybe
your
admin
wants
to
do,
but
not
everybody
on
the
team
has
to
be
able
to
do
it.
They
should
just
be
able
to
run
run
your
your
image.
So
how
does
this
actually
work?
So
what
I'll
do
is
I'll
run
the
command
and
then
I'll
explain
to
you
what
it
does
pack
pack
that.
B
A
B
is
for
choose
a
builder,
and
in
this
case
what
I
have
is
a
builder
called.
That's
actually
pulled
from
a
container
from
a
registry
which
is
from
this
GC
r,
io,
/,
picado
/,
built
packs,
and
what
this
thing
is
is
it
builds
tiny,
go
images,
it
just
knows
how
to
build
time.
You
go
images,
and
the
next
thing
that
I
have
here
is
the
image
that
I
want
to
build
and
we'll
just
call
it
X
X
X
X,
just
because
I
don't
have
a
good
name
port
and
we'll
publish
it.
A
So
really
what
this
command
will
do
is
it
will
build
a
container
image
and
it
doesn't
need
root
and
it
does
not
need
a
docker.
It
does
not
need
docker
to
be
running,
which
is
actually
another
interesting
thing
about
the
dock
of
command
and
it'll
build
it
locally
for
us.
So
let
me
actually
run
this
for
you
all.
So
what
it's
doing
now
is
it's
pulling.
It
and
it's
doing
it's
actually
now
it's
expanding
itself
and
thinking
about
Oh.
What
do
I
need
to
do
so
it
says.
Oh,
this
is
a
go
application.
A
Well,
I'm
gonna,
compile
it
for
you,
because
and
in
the
go
community
and
just
like
every
programming
communities,
there's
lots
of
conventions
that
you
can
take
advantage
of.
So
one
convention
is
that
you
can
put
your
binaries
and
in
a
directory
called
slash,
CMD,
slash
something
else
we
can
handle
that.
But
in
this
case
this
application
actually
just
had
it
at
the
root.
So
it
said:
oh
there's
a
file
here,
a
main
dot
go
file.
Oh
I
know
how
to
build
that
and
I
didn't
tell
anything
else.
A
It
didn't
even
use
a
docker
file
as
a
matter
of
fact,
RM
dash,
RF,
dockerfile
and
and
the
RF
was
definitely
superfluous,
but
I
want
to
show
you
all,
but
that's
I'm
how
serious
I
am
about
getting
rid
of
docker
files.
So
we're
gonna
build
us
again
and
doesn't
take
that
long.
So,
like
I
said,
it's
actually
figuring
itself
out.
It's
excellent
downloading
and
it's
it's
downloading.
So
let
me
go
to
the
comments
here.
So
I
have
two
screens,
so
this
is
where
it
gets
weird
ma
says:
is
it
possible
to
build?
A
Is
it
possible
to
codify
build
pack
config
command?
Yes,
actually,
if
you
go
to
the
bill
Peck
site,
they
actually
do
have
someone
else
up
there
and
and
I
do
know
this
so
pivotal,
which
was
just
required
by
VMware
they've
gone
through
and
they
have
a
whole
build
service
based
on
this.
So
if
you
have
an
application
and
a
tree
absolutely
whenever
you
check
it
and
dog
silly
automatically
build
it
for
you,
so
it
can
always
be
available
and
even
better.
A
Yet
they
can
look
upstream
to
look
at
your
dependencies
and
they
can
automatically
build
those
and
whenever
those
dependencies
get
built,
everything
down
the
tree
gets
built
as
well.
It's
actually
pretty
cool
whenever
you
think
about
it
and
definitely
part
of
this
modern
software
supply
chain
idea
where
we
are
going
from
source
to
running
on
servers
and
do
a
moving
from
all
the
steps
in
a
secure
fashion.
So,
back
to
that
back
to
this,
we
built
this
image,
and
now
we
have
this
image
locally.
A
So
what
I
want
to
do
is
actually
look
at
this
and
see
how
big
this
image
is
and
I'm
gonna
need
to
make
this
a
little
bit
bigger.
So
this
image
that
I
just
built
is
6
megabytes,
bigger,
I'm,
not
gonna,
go
look
inside
of
it
and
see
what's
in
there,
but
still
not
not
too
much.
You
didn't
add
too
much
on
to
where
we
are
so
that
was
just
what
I'm
going
to
talk
to
you
about
building
cloud
native
bill.
A
Paxton
from
the
rest
of
us
I
will
not
be
doing
docker
files
anymore
I'm,
actually
going
to
do
this
whole
CMB,
which
is
short
for
cloud
native,
build
packs.
So
next,
so
we
have
an
image
now.
I
need
a
cluster,
and
this
is
an
interesting
place.
I
used
to
work
for
a
cloud
company
help
build
a
cloud
company
many
years
ago.
Digital
ocean,
so
I
use
digital
ocean
for
everything.
Now
I
work
at
VMware
and
we
compete
and
partner
with
everybody,
it's
crazy.
So
what
clouds
do
I
use?
A
I
use
AWS,
I
use,
Azure
and
I
use,
Google
and
I
use
digital
ocean
sometimes
and
I
may
have
been
known
to
use
IBM
so
I'm
all
over
the
places
with
the
clouds.
But
today
so
I
don't
have
to
take
a
pic
of
what
cloud
I'm,
gonna
I'm,
just
gonna
use,
MIDI
and
I'm
gonna
boot
up
a
mini
cube
here
and
why
mini
cube
a
while
ago,
I
was
out
there
advocating
that
people
should
use
kind,
and
actually,
let
me
let
me
actually
go
and
look
at
this.
A
So
I
do
a
lot
of
Cooper
Nettie's
development
on
my
desktop
and
mrs.
and
mrs.
actually
as
a
desktops,
not
a
laptop
I
any
given
time
not
now
to
this
rebooted,
but
I
might
have
four
or
five
six.
Seven
eight
clusters
running
and
depending
on
what
project,
I'm
working
on
and
I
used
to
advocate
for
kind
and
kind
is
great
kind
runs
in
docker
is
basically
boots
of
kubernetes
cluster
inside
of
docker,
and
it
has
support
for
multiple
master
nodes
and
and
multiple
worker
nodes.
A
And
you
know
what
that's
great,
but
there
is
one
thing
about
kind
that
I
do
not
like,
especially
whenever
I'm
doing
applications
that
I
want
to
do
ingress
for
getting
actual
access
to
their
cluster
is
weird
whenever
it's
in
doctoring,
especially
if
you're
on
a
Mac,
because
what's
gonna
happen
is
well
I
want
to
go
to
port
80
on
this
thing
and
I
want
to
go
to
port
81
another
one.
What
port
80
is
that
and
I
don't
want
to
have
to
go
to
my
browser
and
type
in
well
cluster.
A
A
port
80
is
80
81
and
cluster
B
port
80
is
80
82,
so
I
know
it's
something
that
can
be
fixed
and,
and
you
know
what
shoutouts
to
Ben
elder
and
everybody
else
in
that
crew
are
actually
this
I
love
this
software.
But
in
this
case,
for
this
project,
I'm,
just
gonna
use,
mini
cube
and
mini
cube.
Is
a
venerable
mini
Q's
been
around
for
a
long
time
and
what
mini
cube
is
it
just
runs
kubernetes
locally?
A
The
cool
piece
is
that
I
do
work
with
VMware
and
I
have
a
nice
little
beefy
computer
here.
So
what
I
end
up
doing
is
I
run
the
I
run,
my
mini
keep
clusters
and
VMware
fusion,
so
I
can
actually
spend
them
up.
Keep
them
all
separate,
and
the
neat
part
is
is
because,
like
I
said,
this
is
a
beefy
machine.
A
I,
give
every
single
cluster
12
minutes
of
memory
and
eight
CPUs,
so
it
allows
me
to
not
simulate
full
production
workloads,
but
I
can
actually
get
a
few
things
on
there
and
actually
play
around
with
it.
So
this
thing
is
now
up
a
couple
neat
things
about
my
command
line.
I'm
sure
you
all
have
been
talking
about
it.
I
do
use,
I
do
use
fish.
Shell
I've
used
fish
shell.
A
A
Whenever
I
went
to
alternative
platforms,
where
I
had
to
do
Korn,
shell
and
TC
show
and
and
realized
well,
this
is
all
garbage
went
back
to
bash
and
then
oh
wow,
back
2006-2007
I
moved
over
to
seashell
and
loved
it
and
did
some
work
on
that
thing
and,
and
then
oh,
my
zio
sage
came
out,
and
that
was
amazing
and
then
Braca
that
I
discovered
fish,
which
did
all
these
things,
like,
literally
all
these
things
that
you
wanted
to
configure
your
Oh
my
fish
to
do.
Oh,
my,
oh,
my
Z
shell.
A
A
There
is
a
cluster
here
called
mini
cube,
mini
cube!
Well
that
that's
it's
just
saying!
Here's
the
cluster
here
called
mini
cube.
I
can
change
my
name
space
to
defaults,
and
this
will
now
change
from
this
is
just
shows
how
long
I
mean
what
what
cluster
and
what
name
space
is
my
d4,
because
I
touch
a
lot
of
clusters
and
then
right
here.
If
something
takes
a
long
time,
my
she'll
tell
me
that
it
took
a
long
time
to
run.
A
It's
just
things
need
I,
think
you
need
to
know,
and
another
neat
thing
is
that
there's
an
eye
here:
I'm
a
heavy,
heavy,
heavy,
heavy,
heavy
beam
user,
and
because
of
that,
even
on
my
command
line.
You
all
can't
see
this,
but
you
can't
see
my
hands,
but
I
use.
Ben
motion
commands
on
my
command
line,
so
you
notice
this
thing
changed
from
I
to
n.
If
I
go
back
to
insert
mode,
it
goes
back
to
I
and
then
delete
and
and
that's
it
so,
but
that's
not
we're
here
to
talk
about
today.
A
A
We
got
to
get
this
thing,
so
we
have
this
image
now
and
we
want
to
get
in
our
cluster
and
kubernetes.
Does
some
really
interesting
things
here?
So
some
of
you
all
are
thinking
I'm,
just
gonna
go
fire
up
a
command
I'm
editor
and
I'm,
going
to
go
ahead
and
write
out
my
deployments
by
hand
and
I'm
saying
no
I.
Don't
do
that!
There's
a
few
things!
I've
refused
to
commit
to
memory
my
ATM
PIN
code,
even
though
it's
only
four
numbers
and
also
how
to
write
deployments
by
hand
I
just
fulfill
that
there's
computers.
B
A
This
asked
Brian:
do
you
really
do
you
normally
run
minicab
with
more
nodes,
so
here's
an
interesting
thing
about
mini
cube
is
going
to
a
transition.
What
I'm,
not
what
I'm
waiting
for
is
on
many
cubes.
Actually,
where
tell
you
said
it,
it's
right,
many
cubes,
only
a
single
node.
They
are
working
on
it
and
also
what
they're
working
on
that
I
heard.
Is
that
they're
working
on
cluster
API,
because
what
I
really
want
is
a
main
mini
cube.
Oh
my
oh,
my
on
my
Mac
here
and
then
I
would
like
to
go.
A
Do
cluster
API
and
I
would
like
to
create
clusters
and
add
machines,
so
I
could
dynamically
create
these
things
and
I
could
just
have
it
all
configured
from
one
place
using
the
of
kubernetes,
a
declarative
configuration
to
make
that
all
work
not
there
yet
so
I
blew
them
up
by
hand.
Oh
and
actually
just
another
neat
thing
with
mini
cube.
You
can
have
profiles.
So
if
you
do
mini
cube,
P,
which
I
did
before
mini
cube,
I,
actually
so
right
now,
there's
there's
four
of
them.
There's
eight
of
them
on
here.
A
I
can't
run
all
of
them.
I'm
not
running
I'm,
only
running
one
right
now,
because
I
reboot
it
so
s
is
kind
of
a
neat
thing
about
mini
cube,
but
about
what
we
were
doing,
we're
creating
deployments
and,
like
I,
said,
I,
don't
create
these
by
hand.
There's
there's
actually
two
ways
that
I
do
this,
and
this
is
like
this
is
like
pro
tip
right
here.
How
do
you
create
a
deployment
of
kubernetes
each
search
for
nginx
kubernetes
deployment
example,
and
this
is
number
one
way
I
do
it?
A
I
come
right
to
this
page
and
I
copy
it
off
just
hit
this
little
button
here
and
I
copy
it
off.
That's
number
one.
The
next
thing
I
do
and
there's
something
else
that
you
can
do,
which
is
not
so
basic
is
that
is
that
everyone
has
you
patrol
in
the
machine.
Pig
Patrol
is
literally
it
does
too
much
and
then
sometimes
not
enough,
but
in
this
case
it
actually
can
do
some
interesting
things
here.
A
So
we
know
about
tube
control,
create
I,
mean
stop
talking
so
fast
because
to
control
create
and
you
can
create
with
the
dash
F,
but
there's
also
cube
control.
Creative
I
just
hit
enter
I
scroll
up,
there's
a
commands,
and
these
commands
right
here
things
that
cheap
control
can
create.
So
one
of
these
items
is
a
deployment,
so
what
we'll
do
is
we'll
go
to
control,
create
deployment
and
hit
enter,
and
now
it's
going
to
tell
us.
This
is
how
I
create
a
deployment.
A
B
A
A
If
I
do
okay
get
deployment,
you'll
notice
that
it's
there
it's
up
and
running
if
I
get
pods
there
are
there
and
they're
up
and
running,
but
it
and
I
have
to
do
anything.
So
if
I
do
to
get
deployment
web
say
that
oh
yeah,
no,
it
did
all
the
work.
It
did
all
the
hard
work
for
me.
Creating
deployments
and
services
and
service
accounts
and
secrets
should
not
be
hard.
Let
the
tools
do
things
for
you
and,
in
this
case,
I'm
using
these
tools.
A
The
problem
is:
is
that
and
as
we
will
see,
is
that
when
you
create
it
right
in
the
cluster,
no
one
does
that
in
production.
Please,
like
no.
You
should
not
do
that
in
production,
I'm
sure
someone
doesn't
in
production
somewhere
and
we
should
not
do
that
in
production.
So
let's
not
do
that
I'm,
just
showing
you
this
is
here
is
my
demo,
so
the
applications
running
the
next
thing
you
want
to
do
is
the
same
thing.
A
Is
we
want
to
create
a
service
so
I'm
sure,
okay,
I
have
create
perfect,
so
I
had
this
command
earlier
ia
leus
to
control,
2k
and
and
I
saw
someone
comment
on
big
props,
big
ups
for
cube
control.
Here's
the
thing
I'm
not
gonna,
be
really
a
gist
or
anti-ages
here,
but
control
a
sort.
Ctl
is
short
for
control.
You
have
tools
and
a
control
tools
and
control.
I,
don't
have
cuddle
knobs
on
my
stove
I
have
control
knobs
on
my
stove.
A
So
this
is
why
I
use
it
that
way,
but
I
do
celebrate
everyone
who
likes
to
have
fun
with
it.
So
if
it's
cuddle
or
if
it's
cuddle
or
to
Bechdel
and
there's
my
friend
Chris,
know
that
likes
to
say
or
any
of
these
other
things,
you
know
what
that's
up
to
you
you're
wrong,
but
have
fun
with
it.
So
now
I'm
creating
a
service
same
thing.
A
So
now
I
have
this
service
here.
So
by
how
do
I
know
this
works
if
I
go
hey
kids
service
website,
that's
not
super
helpful.
I
can
tell
that
it's
in
the
cluster.
But
how
do
you
know
if
it's
working
well,
there
is
this
thing
called
end
points
and
if
I
look
at
the
end
points,
I
can
tell
that.
There's
one
that
my
website
my
website
service
is
pointed
to
at
least
one
in
point.
I
know:
I
only
had
one
thing
in
there,
so
I
know
it's
kind
of
working.
A
A
A
So
how
do
we
solve
this?
Well
remember
before
I
was
talking
about
my
old
FTO
days,
and
let
me
close
all
these
windows
I
have
this
problem.
You
can't
see
my
other
screen
where
I
have
another
hundred
tabs
going
open.
I've
been
on
this
thing
for,
like
40
minutes
and
I've,
already
opened
eight
of
these
things
so
yeah
we
don't
need
any
of
these
anymore.
A
Project
contour
project
contour,
so
project
contour,
like
I,
said
before.
Is
this
ingress
for
kubernetes?
What
does
that
mean?
It
allows
me
to
get
traffic
into
my
cluster,
so
let's
install
contour
and
there's
a
reason
why
I'm
using
contour
is
because
the
there
install
is
like
right
here,
two
clicks
from
the
way.
So
what
I'm
going
to
do
here
is
is
I'm
just
going
to
run
this
and
then
I'm
gonna.
Look
at
the
comments.
A
Real
quick,
so
Duffy
says
that
it's
really
hurting
his
brain,
that
my
pod
has
a
170
217
address
and
you
know
it
doesn't
matter.
Our
C
1918
addresses
as
long
as
it's
not
routable
does
not
bother
me
I'm.
Actually,
you
know
what,
for
my
home
network
has
actually
I,
don't
know
if
I
carved
out
a
16,
oh
I
carved
out
like
an
18
for
my
home
network,
which
doesn't
overlap
with
the
VMware
network,
which
you
know
it's
weird,
so
I'm
actually
glad
that
it's
172.
That
means
that
it's
not
touching
anything
else.
A
So
we've
installed
contour
and
notice.
The
farness
aren't
in
this.
In
this
whole
discussion
we
installed
our
web.
We
installed
deployment
and
then
we
created
a
service
and
then
that'll
be
installed.
Who
knows
what
inside
of
contour
here's
the
problem
that
I'm
having
let
me
go
up
a
directory?
Actually,
no
I'm,
not
gonna.
Do
it
here,
I'm
gonna.
Do
it
off-screen.
A
A
But
the
reason
I'm
here
is
because
there's
too
many
things
in
my
kubernetes
cluster
for
me
to
keep
track
of
in
my
head,
so
I'm
using
octant
to
basically
do
all
the
things
I
don't
want
it
to
do,
and
the
things
that
I
wanted
to
do
right
now
is
actually
see
if
contour
came
up
and
what
I'm
gonna
go.
Do
is
go.
A
Look
at
all
the
pods
and
looking
pretty
good
like
it's
up
this
one's
not
up
because
I
know
surge
in
it's
on
a
job
in
the
job
completed
successfully
it'll
go
away
in
a
little
bit,
but
I
know
that
project
contour
is
up
and
working
so
I
use
octant
like
it's
going
out
of
style.
Let
me
actually
since
I'm
here.
Let
me
show
you
some
neat
things
about
octant
that
are
that
are
new,
that
people
just
don't
know.
A
The
first
thing
is
that
I'm,
actually,
the
only
thing
is
that
octant
has
command
line
mode
and
what
you
can
do
is
I
hit,
control
and
enter
and
and
I
go
to
pods.
It
just
shows
me
my
pots.
I
only
have
one
pod,
that's
kind
of
neat
another
feature
that
I'm
putting
in
here
right
now
and
there's
a
reason
why
I
showed
you
I
showed
you
I
built
this
one
is
in
our
next
version
of
octant
we're
going
to
be
something
kind
of
neat
where
people
have
been
clamoring
for
this
forever.
A
A
Oh
read/write
and
I'm,
like
it's
hard,
dude
and
so
I
did
this
yesterday,
I
haven't
even
come
into
this
mastery
and
I
committed
it
to
a
branch
to
get
merged
into
master,
but
real
talk
I
just
put
that
up
there,
so
I
just
moved
up
to
three
and
you
notice.
It
says
three
right
here.
So
this
is
something
that's
kind
of
neat
just
wanted
to
show
it
off
kudos
to
the
team.
I
mean
I.
A
A
port
so
mini
cube
gives
me
that
running
on
a
virtual
machine
and
gives
me
all
that
good
stuff
that
I
like
here,
so
the
next
thing
I
want
to
actually
build
up
here,
is
a
an
ingress
to
allow
this
to
happen.
So
the
first
thing
I
will
do
is,
let
me
cut
and
paste
like
I
said:
I
am
NOT
a
typer,
so
in
my
work.
A
Or
work
we'll
just
call
this
ingress,
diamo
and
we'll
paste
this
thing
in
this
is
a
so
it's
interesting.
There's
an
ingress,
be
one
object
in
kubernetes.
What
has
been
found
out
is
that
it
actually
didn't
do
all
the
greatest
things
that
we're
doing
that
we
wanted
to
get
done.
So
there
is
now
work
going
on
at
ingress
v2,
but
in
the
meantime,
what
the
contour
team
has
done
and
they
actually
go
on
through
a
couple
of
iterations
of
this
now
this
thing
called
a
HTTP
proxy.
A
What
it
does
here
is,
is
it
creates
a
proxy
from
outside
to
inside
and
it
uses
on
envoy.
So
what
I
need
to
do
here
is
I
need
to
figure
out
two
things:
I
need
to
figure
out
what
I
want
to
call
it,
so
we're
gonna
call
that
website,
and
the
next
thing
I
want
to
do
is
I
want
to
figure
out
what
the
IP
of
my
boxes.
This
is
actually
like.
The
important
piece
so
look
it
out
of
here.
A
A
So
let's
actually
go
over
this
real,
quick,
so
I'm,
creating
a
virtual
host
and
I'm
calling
it
an
in
the
the
fqdn
I
used
to
pronounce
this
out
like
just
pronounce
it
out
like
it
was
a
word
I
guess
people
say:
fqdn
I
called
it
something
else,
but
I
won't
say
that
here
is
gonna,
be
app
dot,
192,
168,
98,
128,
X,
IP,
dot
IO.
Now,
why
do
I
use
X
IP
dot?
Io?
Let
me
show
you
all
this
real
quick.
A
A
Host
names
for
anything
so
in
this
case
I'm
just
creating
this,
hopefully
post,
host,
name
right
here
that
I
could
type
into
my
browser,
so
I
will
cut
and
paste.
This
and
I
will
apply
this
to
the
cluster,
and
it's
now
in
there
and
now
I
should
be
able
to
just
paste
this
in
and
guess
what?
What
does
a
cow
say?
It
says
hello,
so
this
is
kind
of
an
interesting
piece
here
just
to
make
it
easier.
Solo,
Maddie
says
anyone
happen
to
know
if
HTTP
proxy
will
be
the
basis
for
ingress
to
env2.
A
So
you
know
what
I
don't
know
the
answer
to
that.
But
I
do
know
that
there's
a
services
API,
there's
a
there's,
a
conversation
going
on
right
now
and
I
do
know
that
the
contour
team
I
know
they're
involved
in
it,
and
so
I
don't
know
what's
not
happen,
but
I
know
that
they're
going
to
definitely
influence
it.
And
if
anyone
hears
this
and
says
that
I've
got
this
wrong,
it
doesn't
matter
because
I'm
on
tape-
and
history
will
remember
me
so
all
right.
So
now
we've
created
our
app
and
we
got
it
deployed.
A
A
We've
created
an
application,
we've
packaged
it
up
in
an
image,
multiple
ways
and
we're
now
we're
going
to
use
cloud
native
built
packs.
We
now
we've
created
a
manifest,
which
includes
a
deployment
and
a
service
and
an
HTTP
proxy
from
contour,
and
now
we
have
it
showing
locally
and
here's
a
crazy
bit.
I've
been
talking
a
lot,
but
really
this
is
not.
This
was
only
a
few
minutes
of
work
and
and
what
we
need
to
do
is
first
of
all,
we
need
to
understand
this.
A
We
need
to
have
better
practices
at
this
and
then
what
we
need
to
do
is
automate
this.
So
no
one
thinks
about
this
anymore.
This
is
really
not
excited,
but
the
next
piece
that
I
want
to
talk
about
are:
oh
I,
know
what
I'm
gonna
talk
about
now.
I
want
I'll
talk
about
pot
security
policies
and
liveliness
checks,
but
I'm
actually
gonna
skip
over
that
and
I'm
actually
gonna
talk
about
is
another
thing
that
I
think
is
pretty
interesting.
A
Just
in
the
world
of
kubernetes,
so
remember
before,
when
I
I
started
off,
I
was
like
kubernetes
is
a
platform
for
creating
platforms
and
inside
of
kubernetes.
So
if
we
just
happen
to
look
at
this
namespace,
there
are
some
positive,
and
this
is
this
is
our
API.
This
is
a
visual
interpretation
of
our
API
and
what
does
the
api
and
kubernetes
look
like?
Well,
it
looks
like
this:
it's
not
it's,
not
your
standard
imperative,
REST
API!
It's
a
declarative,
API!
A
If
I
changed
it
before
it
got
to
100,
it
would
not
go
all
the
way
to
100,
because
it
would
look
it.
Although
the
current
state
of
the
world
should
be
five
and
if
it
was
below
five,
it
would
go
up
to
five.
If
it
was
above
five
it
would,
it
would
start
terminating
those
pods
to
actually
get
down
to
five.
So
this
declarative
world
is
an
interesting
place.
It's
a
kind
of
a
different
mindset,
but
it
allows
us
to
build
these
really
really
interesting
tools.
A
Actually,
how
we're
gonna
do
this
all
right
so
with
our
declarative,
API.
This
actually
is
a
there's,
a
there's.
A
video
about
this,
this
product
to
Blair
of
API
I.
Think
that
I
posit
that
you
can
build
anything
on
top
of
kubernetes
and
whenever
we're
building
these
tools,
and
then
we
run
Jenkins
outside
of
kubernetes.
You
know
that's
that's
interesting
and
then,
when
people
are
like
you
know
what
well
we
have
all
this
capacity
will
run
Jenkins
inside
of
kubernetes.
Well,
now,
I
find
now
I'm
offended.
People
are
like
well.
A
Why
are
you
offended
well,
I
said:
okay,
so
Jenkins
is
basically
just
the
workflow
engine
and
it
can
do
things
and
they
can
spin
up
jobs,
but
guess
what
kubernetes
can
do
all
those
things
and
spend
up
jobs,
whether
it
in
workloads,
whether
it
be
pods
or
whatever
else
so
now,
there's
all
this
duplication
of
effort
when
you're
running
Jenkins,
not
Jenkins
X,
but
Jenkins,
the
the
Java
thing
inside
of
kubernetes.
So
why
can't
we
take
advantage
of
of
kubernetes
to
build
whether
it
be
CI
and
CD?
A
A
It
was
not
great
for
like
a
year,
but
now
that's
not
that's,
not
actually
anything
to
do
with
the
project.
It's
just
that
one
page,
but
the
project
is
actually
pretty
dope
and
I'm
going
to
introduce
it
for
those
who
have
not
seen
it
before
so
like
it
says,
on
the
page,
it's
an
open-source
framework
for
creating
CI
and
CD
systems,
allowing
developers
to
do
things
and
what
we're
going
to
do
here
is
we're
going
to
really
do
something
interesting,
Mykel,
say
app.
B
A
App
I
push
this
out
to
push
this
out
to
github
earlier.
If
you
want
to
go,
look
at
it
or
take
my
code,
it's
all
good.
It's
all
open
source,
there's
no
license
file.
I,
probably
should
put
one
there,
but
it's
not
much
going
on
in
there.
So
we're
gonna
do
a
tech.
Time
is
what
we
want
to
do
is
we
want
to
say:
hey
I
want
a
new
version
of
this
app
I
want
you
to
build
it.
I
want
you
to
do
the
CI
process.
A
We
won't
talk
about
CD,
but
let's
talk
about
CI
well
soon,
nothing
was
a
change
and
I
want
to
be
able
to
trigger
that
change,
and
then
I
want
to
be
able
to
build.
My
image
and
I
want
to
push
it
off
to
to
docker
hub
and
why
am
I
not
using
GCR
or
a
CR
or
ECR
or
any
of
the
other
CRS?
Well,
you
know
what
I'm
logged
in
with
github
right,
I'm
logged
in
with
docker
hub
right
now
great
service.
Hopefully
it
stays
around
forever.
It's
actually
pretty
neat
and
there's
just
lower
friction.
A
A
Do
I
am
going
to
run
I'm
going
to
actually
run
this
come
and
here's
a
here's
a
crazy
bit,
and
this
is
something
that
you
should
be
very
conscious
about
whenever
you're
doing
this
with
kubernetes
the
reason
I'm
running
this
command
directly
is
because
I
just
ran
it
a
little
bit
ago
and
it
didn't
blow
McCluster
up,
but
in
production
there's
no
way.
So
if,
if
I
like,
let
me
show
you
this
thing
so
I
do
QT,
curl,
L
and
we'll
just
less.
What's
in
there,
there's
some
license
headers
and
just
lots
of
license.
A
Well,
that's
funny.
There's
license
headers
on
the
namespace
and
the
pot
security
policy,
but
they
didn't
choose
the
license
to
cluster
roll.
That's
very
interesting!
Alright,
so
they
license
most
of
the
stuff
in
here.
So
there's
quite
a
few
things
in
here
and
I'm
sure
I
could
go
through
and
tell
you
what's
in
there,
but
this
is
a
demo
and
it's
my
computer,
so
I'm
not
going
to
do
that
to
you
I'm
just
going
to
do
this
and
do
control
apply.
A
I
hope
that
this
thing
actually
creates
namespaces
it
does
it
created
this
Tecton
pipelines
and
then
they
created
a
whole
bunch
of
CR
DS.
We're
only
going
to
use
two
three
of
them
and
in
Crayola
has
a
validating
webhook
configuration
in
there
and
I'm
mutating.
Oh
it's!
Actually
it's
actually
pretty
cool,
so
Vivian,
Ricky
aronia,
says
tap
would
be
a
good
to
see
the
changes
before
you
control
apply.
Yes,
actually
you
are
a
hundred
percent
correct
and
here's.
The
interesting
piece
is
that
I
was
a
little
bit
time
constrained
on
this
demo.
A
A
In
the
second
session,
I
will
talk
about
cap
and
ytt
and
why
I
think
that
they
are
either
the
basis
for
a
lot
of
good
things
when
it
comes
to
templating
and
packaging
or
no
templating
and
like
workflow
in
kubernetes,
and
then
I
will
talk
about
another
thing
that
I
wrote
called
chief,
which
is
I,
think
when
I
think
about
packaging,
so
we're
baby
stepping
our
way
up
there
I'm
actually
trying
to
show
why
we
need
some
of
these
things.
Not
just
hey.
Look
at
this
cool
tool,
so
that's
gonna,
be
the
next
one.
A
Maybe
they'll
fight
back
to
it
next
week,
or
maybe
not
I,
don't
know.
But
let
me
let
me
move
on
so
probably
have
here.
Is
we
have
let's
go
back
like
I
said,
there's
too
much
stuff
in
our
cluster
I,
just
I
just
can't
keep
the
track
of
it.
So
I'm
look
I'm
using
octant
and
I'm.
A
Looking
at
my
one
that
project
contour
I
want
to
use
Tecton
pipelines,
I'm
looking
at
Exxon
pipelines
and
I'm
looking
at
deployments,
there's
two:
there
is
a
pipelines,
controller
and
then
there's
this
thing
for
web
hooks
backing
the
admission
controller
kind
of
dope
and
then
there's
some
pods
interrupt
a
cassette
all
right
there.
It's
not
that
it's
not
that
much.
But
what
we're
going
to
do
now
is
we're
going
to
actually
write
this
thing
that
builds
our
images
in
cluster
and
then
pushes
them
out
to
pushes
them
out
the
docker
hub.
A
So
let's
see
how
do
we
get
started
with
this?
Well,
if
we
go
to
github.com
size
detect
on
CD,
which
is
the,
what
do
you
call
it?
The
org
there
is
this
concept
of
Kellogg's
know:
maybe
it's
Kellogg
there.
It
is
catalog
and
inside
of
here
are
they
call
it
a
catalog,
but
this
is
really
just
like
recipes.
So
all
these
things
you
can
do
and
Tecton
so
this
Tecton
it's
basically
is
a
runner.
A
A
Build
kit
or
I
want
to
do
build
packs,
and
then
you
could
take
those
excuse
me
those
tasks
and
you
can
combine
them
together
in
pipelines,
and
you
can
say
all
right
when
someone
pushes
a
commit
up
to
github
and
instead
of
running
to
have
actions,
I
could
actually
have
it
go
out
at
the
Tecton
and
I
could
have
a
trigger
Tecton.
It
could
actually
do
the
build
and
then,
when
it's
done
it
can
signify
and
make
another
trigger.
A
So
what
we're
going
to
do
now
is
we're
going
to
do
some
cloud
native,
build
packs
and
then
we'll
probably,
what
will
will
say
I
do
this
will
be
fun.
So
what
are
we
doing
here
so,
like
I
was
saying
earlier,
this
is
low.
This
catalog
are
like
recipes
and
what
I
can
do
is
oh
I
can
do
some
more
to
control,
apply
random
stuff
to
my
cluster
I
have
what's
the
worst
that
could
happen
right.
So
what
I'm
going
to
do
here
is
I'm
just
going
to
do
this
and
other
window.
A
So
so,
if
I
go
back
over
here,
I
look
at
my
octant
and
I
go
to
Mike,
so
we're
just
using
the
default
environment
right
now,
yeah
and
I
three
pods
running
all
right.
So
if
I
look
in
here
in
this
list,
I
see
that
we
have
custom
resources
for
tech
time
and
we
have
this
bill.
Pax
v3.
So
go
look
at
the
animal
for
it.
There's
a
lot
of
stuff
in
here
and
actually
I'll
go
through
some
of
it
for
you.
A
So
with
the
with
the
tasks
have
per
amps,
and
these
prams
are
just
the
inputs
and
it's
kind
of
neat-
you
can
put
descriptions
for
them.
So
we
can
say
what
the
Builder
image
is.
We
can
say
the
user
ID
for
the
Builder
image,
so
we
can
configure
this
thing
to
work,
how
we
want
it
to
work
and
then
there's
this
thing
called
resources.
So
resources
are
other
things.
A
I,
don't
know:
Oh
PS
other
people's
stuff.
So
in
this
case
we
have
a
resource,
that's
called
source
and
it's
get
and
so
it'll
be
a
git
repository
and
then
what
we'll
produce
is
this
thing
will
produce
an
image
and
actually
everything
else
in
here
is
not
important
to
the
casual
observer,
but
the
neat
part
is
that
we
can
actually
see
using
the
steps
whenever
I
run
this
task.
This
is
what
will
happen
so
all
these
things.
So
it's
actually
a
lot.
So
I'm
not
gonna,
go
through
it.
A
A
One
build
image,
amo
and
it
says
uncommon
below
to
use
an
existing
cash.
So
whoops.
This
is,
is
that
the
builder
knows
they
can
cache.
If
you
have
a
heavy
build,
this
bill
doesn't
take
very
long,
so
I'll
get
rid
of
that,
and
we
will
do
like
this
to
see
what
else
is
in
here.
So
we
have
an
image
and
what
we
need
to
do
is
we
need
to
supply
it
with
an
image.
So
the
neat
part
is:
is
that
we
can't
just
supply
with
an
image.
There's
there's
more
things.
A
There
is
actually
an
image
from
github
and
the
sepulcro
posit
orys
is
easy
because
all
you
have
to
do
is
you
can
just
go
to
the
URI,
but
if
it's
in
a
private
posit
or
e,
there
might
be
keys
involved
or
it
might
just
come
from
somewhere
else,
so
we
don't
know,
but
we're
gonna
specify
it
here.
So
what
we're
going
to
do
is
we're
just
going
to
create
something
called
Brian,
L,
/
web,
say
/,
Tecton,
/
live
and
we're
going
to
get
rid
of
that
/
and
put
this
in
here.
So.
A
A
A
So
yes,
mr.
playstation4
zombie,
Zachary
5,
what
are
you
watching?
You're
watching
Bryan
Liles
create
a
Tecton
task.
Runner,
so
I
can
automatically
build
an
image
inside
of
a
kubernetes
cluster,
all
right.
So
now
we
we
have
this
builder
image
and
let's
see
what
happens
so
we're
going
to
sit,
because
this
is
all
kubernetes.
We
can
use
cube
control,
so
I'm
going
to
Kay
a
PI,
a
create.
A
See
I'm
going
to
I'm
going
to
actually
see
if
I
can
get
the
status
of
this
thing,
and
what
you're
gonna
see
is
that
my
example
run.
It
failed
and
here's
something
interesting.
Actually,
this
is
for
builders
who
are
using
kubernetes
notice.
Why
I
created
I,
just
typed
tube
control
get
Tecton
pipelines
and
it
gave
me
two
things.
A
So
here's
an
interesting
thing
about
kubernetes
and
let
people
know
so.
Whenever
you
type
view
control
get
all
tube
control
doesn't
really
know.
What
all
is
it's
whenever
the
objects
are
created,
they
are
defined
in
a
part
of
a
group
called
all
so,
whenever
you're,
creating,
if
you're,
if
you're,
building
controllers
and
CR
DS,
actually
in
the
CR
D
configuration,
you
can
say
what
group
something
applies
to
so
the
text
on
pipeline
authors
created
a
group
called
Tecton
pipelines
and
then
put
all
their
stuff
in
there.
That's
pretty
dope.
A
Everyone
should
do
that
whenever
you're
thinking
about
it,
because
how
many
times
you've
been
sitting
at
a
console
and
you're
like
Q
control,
get
what
any
type
keep
control
get
all.
And
then
you
don't
even
you
don't
even
know,
because
it's
not
all
to
tell
you
the
truth,
it's
just
all
the
things
that
were
in
a
group,
cobalt,
so
just
a
little
tidbit
there.
So
what
we're
doing
now
is
we're
gonna.
Do
keep
control,
describe
this
thing
and
see
what
it
says.
It's
going
to
say
man.
It
says
invalid
input.
A
Resources
tax
declared
required
resources
are
missing
from
a
task
force
from
tasks
run.
It
says
the
source
is
missing,
but
if
I
hold
on
here,
if
I
go
in
them
run,
build
image
and
all
I
have
it
in
my
workspace.
Why
is
that
not
working?
Well,
I
actually
do
know.
The
answer
here
and-
and
the
answer
is,
is
because
we're
not
dealing
with
workspaces,
so
we
need
to
actually
change
our
configuration
to
make
this
work.
So
what
I'm
going
to
do
is
is
actually
type
in
the
the
real
one
of
what
we
should
have
here.
A
A
How
we
should
do
this,
and
so
what
have
I
typed
in
here
so
before
I
remember,
I
was
talking
about,
we
had
inputs
and
outputs
and
there
was
an
call
source
well.
What
we
need
to
do
is
we
need
to
create
a
source
and
I,
don't
have
a
credit
yeah,
but
I'll
create
in
a
second
called
web,
say
underscore
of
what's
good.
What's
a
dash
get
and
that
just
describes
might
get
repository.
So
now,
if
I
delete
okay,
good
typed
on
pipelines,
if
I
delete
this
I'll
show
you
this
so
I
have
to
believe
it.
A
A
B
A
Saying
before
that,
pipeline
resources
had
more
configuration,
and
then
we
couldn't
just
have
a
git
repository.
That's
true!
So
what
I'm
going
to
do
here
is
I'm
going
to
create
this
website,
get
that
yeah,
Mille
and
I'm
cut
and
paste
some
more
Yale
in
here,
because
why
not?
And
but
what
really?
What
we're
doing
is
that
a
git
repository
is
just
not
a
URI
of
not
just
with
a
single
URI
there's,
also
a
revision
and
there's
also
a
type
because
it
doesn't
have
to
begin
necessarily.
A
A
A
Your
runner
like
this
is
the
instance
of
the
run
and,
if
I,
if
you
apply
it,
how
is
it
gonna
know
that
how
you're
gonna
separate
your
instances
so
I'm
using
create
and
delete
to
do
this?
Normally,
what
we
would
do
is
we
would
use
generate
name
as
part
actually
I'll
show
you
that
in
a
second
actually
I
said
right
now,
we
would.
B
A
We,
instead
of
using
name
I,
would
use
generate
name
and
I
would
call
it
example
Ryan,
and
it
would
actually
put
like
a
six
random
letters
up
to
this,
but
I'm
not
doing
that,
because
I'm
spending
on
deleting
it
a
whole
bunch
of
time.
So
it's
just
even
you're
having
one
so
ok
delete
is
that
in
the
cluster
nope.
A
B
A
On
so
I'm
gonna
show
you
this
other
command,
so
it
was
running.
So
this
is
a
Tecton
has
a
command-line
tool.
It's
called
TK
in
and
then
there
is
a
tool
here.
I
use
this
all
the
time
actually.
So
this
is
a
we're
using
a
Tecton
task
run
and
then
we're
just
showing
logs
for
everything
in
follow
mode.
So
really
what
this
is
doing
in
in
real
time
is
it's
following
the
bill,
so
it
downloaded
it
and
now
it's
actually
running
go
install
and
this
is
actually
happening
in
the
cluster
notice.
A
A
This
is
the
this
is
like
the
number
one
problem
of
being
a
gamble
engineer
and
if
you
work
on
kubernetes,
it's
it's
definitely
like
being
my
engineer
so
wow.
A
A
A
There's
no
way
you're
gonna
do
what
I'm
getting
ready,
I'm
getting
ready
to
cap
my
docker
config.
You
know
why,
because
there's
no
passwords
in
it.
If
you
were
on
Linux,
this
would
have
a
password
in
it,
and
this
is
the
difference
between
this
is
actually
one
of
the
problems
with
docker
for
Mac.
Is
that
because
I
use
these
handlers
to
make
it
easier
to
keep
passwords,
they
don't
keep
passwords
in
this
file.
A
So
what
we're
gonna
have
to
do
is
we're
have
to
create
a
service
account
and
a
secret
and
or
and
then
we're
gonna
have
to
well
create
the
service
account
and
then
we'll
know,
create
the
secret
and
then
we'll
create
the
service
account
and
then
we'll
attach
the
service
account
to
the
detect
the
task
Brian,
and
then
it
should
have
all
the
things
it
needs,
but
I'm.
This
is
just
that
this
tripped
me
up
for
hours
before
where
I'm
like.
A
So,
like
I
said,
I
am
out
here
cutting
and
pasting
all
sorts
of
things
today,
so
this
is
the
command
that
I'm
gonna
use
to
create
my
credential.
Something
I
want
you
to
note
is
first
of
all
my
fish.
Shell
is
pretty
awesome.
I'm
fish.
Shell
understands.
If
you
have
multiple
lines
that
you
want
to
edit
food,
you
want
to
actually
page
through
multiple
lines.
It's
pretty
dope
next
thing,
I'm
going
to
do
here
is
and
actually
what
the
record
are.
We
in.
B
A
B
A
B
B
A
Mary
so
quick
question
here
is
on
the
service
accounts.
What's
the
best
practice
for
the
registry
credentials,
stub
the
image
credentials
in
a
service
account
or
explicitly
specify
the
image
pull
seekers
in
and
appointments
wow,
that's
a
whole
new
book
conversation,
but
what
I
will
tell
you
to
do
is
I.
B
A
So
what
I'm
going
to
do
is
my
build
service
and
I'm.
Just
gonna.
Add
this
this
next
secret
from
that's
called
right
spray.
So
something
else
you
all
notice
is
one
of
my
bin
Mohd
I,
actually
get
blinded
by
by
yeah
Mille.
So
I
actually
have
my
yeah
Mille
whenever
it
loads
it
loads
it
in
decompress
mode.
A
So
that's
why
you'll
always
see
that
smile
to
file
up
I
don't
have
to
see
a
whole
bunch
of
things,
but
I
can
definitely
do
that
if
I,
if
I
choose
to
so
what
I
did
is
I
created,
my
service
account.
Alright
I
updated
my
service
account
and
now
my
service
account
has
two
secrets
in
it
and
there's
a
little
typo
here
and
octant,
but
that's
what
it
is.
A
There's
two
of
them
in
there
so
now
I
need
to
make
one
more
change.
So
I
want
to
go
to
my
run,
build
image.
A
And
we'll
go
service,
it's
called
service
account
name
and
we'll
call
it
built
circus
and
hello,
delete
and
will
delete.
This
thing
like
I,
said
because
I'm
only
running
one
test
and
I'm
the
pipeline
I'm
only
I'm
only
gonna
rent.
This
way.
Normally
you
would
not
do
a
create,
you
would
have
it
inside
of
a
pipeline
and
you
would
generate
one
for
each
run
and
then
I
want
to
create
it.
A
B
A
A
And
I
go
to
this
web
say
you
notice
that
this
Tecton
live
was
just
put
here
a
couple
seconds
ago,
and
now
with
that,
you
can
now
start
now.
You
have
a
CI
process
done
now.
You
can
start
working
you
on
a
get
ops
process,
so
I'm
actually
going
to
cut
it
here,
because
the
next
step
now
is
to
do
the
continuous.
We
started
on
the
like
the
the
parts
of
the
CI,
but
we
need
to
talk
about
packaging.
We
need
to
talk
about
templating.
A
We
need
to
talk
about
workflow,
like
someone
mentioned
cap,
before
we're
going
to
talk
about
all
that
next,
like
now
that
I
can
create
all
this
stuff.
How
do
I
really
get
it
in
my
cluster
and
then
how
do
I
work
in
multiple
environments?
So
that
is
where
we
are
going
to
end
it.
So
I
will
do
this
odo
in
here.
A
That's
right
and
it's
Friday,
so
we're
gonna
end
with
an
explosion,
and
thank
you
all
for
coming.
Maybe
it'll
be
next
Friday,
maybe
later
Friday
after,
but
it
will
be
soon.
Thank
you.
All
I
am
sweating
like
I,
don't
know
what
so
I
can
turn
my
fan
on,
but
thank
you
all
for
allowing
me
to
talk
to
myself
for
an
hour
and
a
half
and
I
hope
you
all
got
something
out
of
this.
Please
let
me
know
on
Twitter
or
Facebook
or
LinkedIn
or
tik-tok
or
whatever
else.
It's
all
good.