►
From YouTube: Tutorial: A Kind Workflow for Contri... Benjamin Elder, Duffie Cooley, James Munnelly & Patrick Lang
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
Tutorial: A Kind Workflow for Contributing to Kubernetes - Benjamin Elder, Google; Duffie Cooley, VMware; James Munnelly, Jetstack; & Patrick Lang, Microsoft (Limited Available Seating; First-Come, First-Served Basis)
Please bring your laptop fully charged as we will have limited charging stations available in the room. How can you be confident that the change you make is functioning as you expect *before* you submit that PR? Many contributors to the Kubernetes code base want to increase the confidence they have in their code prior to pushing that code upstream. This workflow will simplify this process for you! Bring your Mac, Windows or Linux laptop to this session! We will show how to install the tools you need - Docker, golang and Kind. We will guide you through a great workflow for contributing and testing your code. We will be leveraging sigs.k8s.io/kind to show you how to build Kubernetes locally and test your code. Then we’ll use Kind to run e2e tests against your local build. Together, these new skills will enable you to feel more confident in the changes you are introducing to the existing code base and enable you to contribute more frequently!
https://sched.co/Uaek
A
Everybody
welcome
to
the
session.
It's
good
to
see
you
all
if
you
have
not
already
done
so.
I
just
need
you
to
download
this
image.
This
image
is
going
to
represent
all
of
the
resources
necessary
to
do
the
work
that
we're
gonna
do
in
this
session.
The
line
here
on
the
bottom
of
the
screen,
curl
L
capital,
L
kinda
case
I,
work,
slash
fetch
and
you
can
you
can
do
that
without
the
pipe
bash
and
see
what
the
command
is.
A
If
you've
not
already
done
that,
if
you
don't
already
have
that
image
or
if
you
didn't
see
any
of
the
communication
around
making
sure
that
you
have
already
done
like
docker
pull
queda
IO
kind
workshop,
trying
to
cache
that's
another
way
to
do
it.
Actually,
that
might
be
the
way
to
get
you
unstuck.
Doctor
pull
queda
io,
/
kind,
workshop,
/
kind
cash-
if
you
could
do
that,
could
be
good
yeah.
A
But
do
you
welcome
we're
gonna
get
started
here
before
too
long?
If
you
have
not
already
downloaded
the
kind
cache
image,
if
you
saw
the
tweet
and
just
did
a
docker
pull
of
it
before
you
got
here?
That's
great!
You
don't
need
to
do
this
again,
but
if
you
have
not
already
downloaded
that
image
or
made
it
available
on
your
machine,
the
quickest
way
to
do
so
is
right
here.
The
mas
slide
on
the
slide.
A
A
You
didn't
back,
oh
there
we
are
so
if
this
works
for
you,
then
what
will
happen
is
it'll
now
be
a
very
1.4
gig
image
in
your
temp
directory
it'll
be
called
kind.
Cached
rgz,
we're
gonna
load
that
in
to
docker
I
like
when
we
get
this
all
set
up,
so
go
ahead
and
grab
that
image
and
once
it's
downloaded
it'll
be
available
in
your
temp
directory.
So
that's
where
you're
looking
for.
B
B
A
A
I'm,
probably
gonna
get
started
here.
What
I'd
like
to
see
as
we're
as
we're
filing
in
here
I
want
to
see
you
raise
your
hand
if
you
have
not
already
started
downloading
the
image.
If
you
have
not
and
you're
not
planning
to,
then
don't
buy
all
means,
but
if
you
have
not
already
Stadler
started
downloading
the
image
this
line
at
the
bottom
of
the
screen
gives
you
instructions
on
how
to
do
it
right.
A
So
what
this
will
do
is
it'll,
it's
just
a
docker
command
underneath
it,
if
you
want
to
do
it
without
pipe
bash,
you'll,
be
able
to
see
the
output
of
that
command
and
what
it's
doing
is
basically
just
running
a
docker
container
that
will
download
a
1.4
gigabytes,
which
represents
all
of
the
resources
that
we're
going
to
use
for
this
resource
or
for
this
lab,
as
you're
walking
in
the
very
last
line
on
this
slide
very
important
one.
It's
going
to
be
the
way
you
download
all
the
resources
necessary
for
this
entire
lab.
A
So
if
you
have
not
already
downloaded
it
talk
to
your
neighbors,
they
have
already
done
it.
If
you
have
any
questions
or
raise
your
hand,
we'll
get
somebody
out
to
you,
if
you
do
not
already
have
docker
installed
on
your
laptop
you're,
probably
going
to
have
a
rough
time
again
talk
to
your
neighbors
everybody
here,
raise
your
hand
if
you've
installed
docker,
that's
what
I'm
talking
about
alright.
C
A
A
Welcome
everybody
this
week
probably
could
get
started
everybody
ready
to
get
started.
What
do
you
think
actually
Det
raise
your
hand
if
you
have
already
downloaded
the
image?
If
this
is
that's
pretty
awesome,
good
good
people
who
have
two
minutes
remaining
three
minutes
everything
cool
I,
don't
know
good
there.
You
have
any
questions.
I
guess
I
just
stand
up
and
look
and
we'll
get
somebody
out
to
you.
B
C
A
He
also
have
been
elder
who's,
whose
idea
this
was.
He
works
on
in
cig
testing,
a
lot
he's
gonna
come
up
and
do
a
little
intro
to
kind,
and
we
have
James
Nunnally
wild
man
about
town
yeah.
He
started.
Cert
manager
started
a
bunch
of
other
interesting
projects.
He
works
for
jet
stack.
One
of
the
amazing
things
about
this
session
is
that
your
it's
coming
to
you
live
from
four
people
from
four
different
companies.
If
that
doesn't
say,
community
I
don't
know
what
it
does.
You
know
what
I
mean.
A
There
are
microphones
walking
around
in
the
in
the
audience
if
you
have
questions
or
if
there's
anything,
you'd
like
to
say
just
a
flag,
somebody
down
I'm
hard
of
hearing.
So
if
you're
not
trying
to
talk
to
me
with
a
microphone
I,
probably
here
you,
so
this
is
the
command.
If
you
were
to
look
at
kind
gates
that
work,
slash
fetch,
this
is
actually
what
the
command
is
and
what
it's
doing
is
basically
just
running
a
download
manager
to
download
a
much
larger
file.
A
All
right
so,
as
I
said,
the
commands
still
here
at
the
bottom,
so
if
you're
walking
in
right
now,
the
command
is
still
right
there
at
the
bottom
of
the
screen.
Please
go
ahead
and
run
that
command
it
scroll,
L
kind
of
caged
I
work,
slash
fetch
pipe
bash
I'm,
a
Linux
users.
For
this.
The
way
I
would
type
it
in
if
you're,
a
Windows
user,
you're
Patrick
and
help
you
out.
There
are
probably
other
Windows
users.
A
B
A
A
Alright
welcome
to
the
session.
Alright,
let's
get
started,
we
got
about
90
minutes
to
get
through
this
I
think
we
can
do
it.
It'll
be
really
fun
to
talk
through,
but
this
session
in
this
session,
we're
going
to
do
some
really
interesting
stuff,
but
I'm
gonna,
I'm
gonna
tell
you
what
we're
gonna
be
up
to
we're
gonna
go
through
it
and
then
we're
gonna,
and
then
all
of
you
are
gonna
go
through
it.
A
You
acquisition
I
spend
a
lot
of
my
time,
educating
and
helping
and
architecting,
and
basically
just
making
sure
that
people
who
find
success
with
cognitive
technologies,
whether
that
means
figuring
out
how
to
do
kubernetes
or
whether
that
means
figure
out
how
to
repot
form
to
coober
do
this
or
whether
that
means
you
know
how
to
break
up
an
application
into
micro
services
or
when
to
do
that
or
why
do
that
like?
This
is
kind
of
all
the
stuff
I
do
so
every
week.
I
am
a
part
of
a
video
blog
series
at
GG.
A
A
We
also
have
been
elder,
who
worked?
Who
works
on
the
kubernetes
release?
Team
is
actually
pretty
focused
on
a
lot
of
the
cig
testing
stuff.
He
originally
developed
kind
to
help
reduce
the
amount
of
time
it
takes
to
run
and
and
tests
on
stuff
when
presented
inside,
of
like
a
PR
inside
of
the
Kirby's
codebase,
and
that
has
been
wildly
successful
and
he'll
come
up
here
in
just
a
minute
to
talk
about
what
what's
happening
there
and
like
what's
been
done.
James
Nunnally
is
a
solutions
engineer
for
jet
stack.
A
James
is
actually
the
author
of
cert
manager,
and
then
there
was
another
one
before
that
that
I
can't
think
of
the
name
of,
but
he's
here
we
have
Patrick
Lang
from
Microsoft.
He
spends
a
lot
of
his
time
teaching
and
playing
with
containers
on
Windows
working
with
working
with
cubelet
on
Windows.
Doing
all
of
that
stuff
he's
been
working
on
all
this
stuff.
I
think
we're
actually
longer
certainly
longer
than
I
have
I.
A
Think
Ben
has
too
I've
been
working
on
kubernetes
for
about
three
years
now,
so
welcome
everybody
I
see
a
lot
of
laptops
out.
That's
good.
This
is
an
interactive
session
you're,
going
to
be
doing
some
hard
work
on
your
laptop,
but
you're
going
to
be
building
the
kubernetes
code
based
locally
on
your
machine
against
a
commit
and
then
we're
going
to
show
a
broken
functionality
of
version
116
of
Coover
Gnaeus.
Then
we're
going
to
fix
that
functionality
and
then
we're
going
to
build
it
again.
We're
going
to
iterate
show
the
new
built.
A
The
new
build
was
actually
got
everything
fixed
and
we're
going
to
be
able
to
evaluate
all
of
that
using
local
resources
on
your
machine.
Why
are
we
doing
that?
So
I
came
up
with
this
idea.
The
idea
of
this
talk
because
I
found
a
number
of
people
that
I
had
worked
with
were
nervous
about
contributing
to
the
kubernetes
codebase
and
it's
understandable.
It's
a
distributed
system
right,
there's
a
lot
of
moving
parts
and
when
you're
actually
trying
to
put
together
your
first
pull
request
against
a
kubernetes.
A
The
awesome
thing
about
kind
is
that
it
kind
of
changes
that
game
pretty
significantly.
You
can
bring
up
a
coupie
cluster
and
configure
it
in
lots
of
interesting
ways
right
there
on
your
laptop
and
because
of
that,
we
actually
even
kind
even
allows
you
to
build
a
note
image
which
represents
all
of
the
bits
built
from
your
local
check
out
of
kubernetes
and
we're
going
to
work
through
that
in
this
session
right.
A
So
really
what
I
want
is
I
want
people
to
feel
comfortable
engaging
with
the
community
like
I've
I've.
Definitely
had
people
tell
me
that
they
feel
like
they
don't
want
to
contribute
because
it
feels
like
if
they
make
a
mistake-
or
it's
like
some
really
obvious
mistake-
that
they
should
have
known
or
noticed
and
I'm
like
I,
don't
get
how
you
would
have
noticed
that
if
you're
just
you
know
developing
blindly
like
if
you
don't
have
a
way
of
actually
like
evaluating
this
stuff
locally
I,
don't
know
how
you
would
have
noticed
right.
A
I
have
no
expectation,
you're,
gonna
figure
that
out
magically,
and
so
here
I
am
giving
you
tools
so
that
you
can
evaluate
it
locally.
So
you
can
learn.
So
you
can
actually,
like
you
know,
play
with
this
stuff
locally
and
that's
gonna,
be
the
killer
feature
here.
How
many
people
have
contributed
in
kubernetes
before
awesome?
A
So
this
is
our
agenda.
We're
gonna,
get
that
or
if
you
I'm
hoping
all
of
you
have
docker
installed.
If
you
don't,
then
you're
gonna
kind
of
be
following
along
watching
me,
do
it
or
watching
somebody
who's
sitting?
Next
to
you,
do
it
one
work
out?
The
other
we've
started
with
get
the
myths
right.
So
if
you're
just
walking
in
now,
you're
gonna
want
to
hook,
go
ahead
and
make
sure
you
get
those
bits.
A
You're
gonna
use
the
curl
command
to
go
ahead
and
download
an
image
that
will
make
sure
that
we
get
all
the
resources
necessary.
We'll
talk
about
that
image
here
in
a
second,
but
we're
gonna
go
ahead
and
get
all
the
bits
necessary
to
run
this.
It's
actually
somewhat
of
a
contrived
example,
and
it
represents
all
of
the
bits
necessary
to
run
this
lab.
All
of
these
pieces
actually
do
apply
to
kind
of
like
your
normal
workflow,
but
they
are,
you
know,
encapsulated
in
a
way
that
we
can
work
through
a
lab
here.
A
So
we're
gonna
walk
through
like
some
of
the
differences
between
what
we're
doing
we're
going
to
do
an
over
kind
of
the
cat.
We're
gonna
do
an
overview
of
the
kind
project.
We're
gonna
do
a
quick
overview
of
contributing
to
kubernetes.
Just
gonna.
Take
a
look
at
like
how
to
sign
the
CLA
where
to
enter
into
this
whole
thing.
While
people
are
downloading
the
image,
we're
gonna,
do
we're.
Gonna,
look
at
s.
A
Sample
commits
we're
going
to
build
from
before
the
sample
commits
and
we're
gonna
then
build
after
and
then
we're
gonna
verify
the
functionality
of
the
change,
and
this
is
actually
gonna
be
pretty
fun.
So
that's
our
goal
and
then
finally,
we're
gonna
be
super
successful
everybody's
gonna
get
it
done,
it'll
be
what
we're
working
on.
If
you
don't
already
have
docker
installed,
there's
some
instructions
in
the
deck.
If
you
want
to
see
this
deck
on
your
machine,
if
you
go
to
kind
k,
eights
dot
work,
that's
where
this
deck
is
hosted.
A
A
A
So
if
you
had
docker,
you
had
everything
you
needed
and
so
I'm
not
actually
going
to
have
you
install
kind
locally,
we're
gonna,
do
it
in
it.
We're
gonna
do
it
inside
of
a
container,
because
that
way,
all
of
it
can
just
work
for
you.
But
if
you
want
to
do
this
for
real
definitely
check
out
kind.
Cigs
Kay
it's
to
I/o
has
instructions.
A
I
actually
interact
with
that
stuff
on
WS
l2,
on
where
on
Mac
and
Linux,
all
of
the
stuff
is
available,
and
one
of
the
great
one
of
the
cool
things
that
happened
in
the
0.60
r
dot
zero.
Now
it's
actually
available
in
brew,
and
it's
also
available
available
in
choco
and
if
you're
running
Linux,
you
can
just
download
it
from
the
releases
page
on
the
on
the
on
the
github
page,
as
I
said,
everything's
can
everything's
locally
available.
A
What
that
means
is
that
I've
already
checked
out
Cabrini's
for
you
and
I
wanted
to
make
this
point,
because
I
personally
found
it
a
little
confusing.
So
when
you
actually
have
gone
ahead
and
done
the
installation
of
golang
on
your
local
machine
when
you're
actually
checking
out
the
kubernetes
source,
you
don't
actually
pull
it
from
github.
To
really
discover
do
this,
although
you
can
find
it
there,
you
pull
it
from
khz
o
slash
chord.
It
is,
and
that
might
be
confusing
at
first.
Is
that
it's
an
idiom
of
go
itself
right.
A
That's
a
vanity
URL,
and
the
thing
that
actually
makes
that
makes
sense
is,
if
you
look
inside
of
the
code
at
the
imports
you'll
find
that
inside
of
that
import
list,
there
are
a
bunch
of
references
to
kHz,
o
/
kubernetes
right
and
that
way,
if
we
for
some
mysterious
reason
needed
to
move
off
of
github,
or
you
know
anything
else
that
has
to
happen
or
if
we
wanted
to
do
like
local
caching
of
modules
or
any
of
that
stuff.
It
gives
us
the
ability
to
use
that
URL
for
that
trick
right.
A
So
when
you
were,
if
you
were
to
download
kubernetes
locally
on
your
laptop
in
real
life,
you
would
just
do.
Go
get
Steve
for
download.
Okay,
it's
the
I/o,
slash
kubernetes,
and
it
would
download
it
locally.
Tear
get
to
your
your
go
pass.
That's
how
you
would
do
it.
You
wouldn't
check
out
github
doc,
who
knew
the
scooby-doo's
just
want
to
make
that
point.
I
found
it
personally
a
little
confusing
at
first.
A
C
C
Without
without
any
online
dependencies
without
needing
cloud
credits,
you
know
you
can
just
do
this
on
your
laptop
when
I
first
worked
on
kubernetes
I
was
a
google
Summer
of
Code
student
and
I
was
just
working
on
my
Lake.
You
know
Mac
book
and
it
was
really
important
to
me
that
I
could
just
do
this
on
my
laptop
and
back
then
we
have
this
a
current
provider
and
long
story
short,
that's
gone.
C
We
built
this
and
we
hope
that
this
will
let
people
be
able
to
contribute
test
things
develop
just
with
a
laptop,
no
other
resources.
So
this
is
sort
of
what
it
looks
like
when
it's
running
we've
got
docker
on
your
host
and
the
kind
command
line,
and
then
each
node
is
a
container
in
which
we've
run
system,
the
container
D
cubelet,
the
entire
kubernetes,
and
we
can
run
multi
node
clusters
that
have
multiple
of
these
node
containers,
working
together
to
simulate
a
multi
node
cluster.
So
you
can
test
things
like
rollout.
C
But
I
said
criminalizing
docker
multi,
node
H
a
control
plane
very
important.
There
are
a
lot
of
local
kubernetes
cluster
options,
but
you
can,
with
this
one
very
easily
take
a
kerbin
at
a
source
check
out,
modify
some
things,
build
it
and
run
it
which
we
really
needed
for
the
kubernetes
project
itself,
and
we
wanted
it
to
be
really
fast
and
really
stable,
because
kubernetes
has
infamously
long
CI.
A
Thank
You
Ben's
kind
is
also
an
open-source
project.
It's
hosted
at
sigdoc,
ADA
Oh.
If
you
want
to
get
involved
or
if
you
want
to
extend
kind
to
support
other
other
functionality,
you
can
definitely
put
up
a
pull
request.
There
I'm
actually
going
to
pull
up
that
website,
see
if
it
all
could
see
it
real,
quick
because
I
think
it's
actually
a
pretty
good.
That's
the
repository.
A
So
this
is
actually
where
we
do
a
lot
of
the
documentation
of
the
project.
Also,
so
some
of
the
other
stuff
that's
changed
in
0.6,
oh
that
I
thought
was
actually
really
exciting.
Is
this
local
registry
functionality
right?
You
can
spin
up
our
local
registry
for
the
cluster
that
you're
using.
So
if
you're
doing
local
development
against
a
container,
you
want
to
use
it
in
kind
of
a
mini
queue
way
right.
This
is
a
way
of
actually
spinning
up
a
local
registry,
so
you
can
leverage
it
that
way,
which
is
actually
pretty
neat.
A
We
have
functionality
for
private
registries,
there's
a
lot
of
talk
here
in
the
website
about
the
design
and
how
it
works
and
there's
a
lot
of
really
great
capabilities.
You
can
mount
in
things
like
file
structures
that
are
like
on
your
local
machine
into
the
container.
You
can
disable
the
built-in
C&I
and
play
with
things
like
calico
or
psyllium
or
canal
or
final
or
any
of
those
things.
I
use
this
mainly
as
a
learning
the
teaching
tool
so
kind.
For
me,
it's
like
a
hugely
valuable
thing,
because
I
spend
a
lot
of
my
time.
A
Educating
people
on
kubernetes
and
all
of
the
topics
around
it,
and
so
with
this
you
can
actually
do
science
like
you
can
divide,
you
can
divide,
you
can
develop
a
theory
about
like
what
happens
when
all
of
your
certificates
expire
and
practice
that
and
see
what
happens
right.
You
can
do
things
like
you
know
like
we're
going
to
do
in
our
session,
where
we're
actually
going
to
build
a
session.
A
So
if
you
have
the
image
go
ahead
and
raise
your
hand,
that's
really
heartening.
It's
really
good
to
see
how
many
people
are
still
downloading.
We've
got
about
like
20
how
many
people
are
still
downloading,
and
it's
gonna
be
a
really
long
time
how
long
it's
a
really
long
time.
Eight
minutes
two
hours.
A
A
Yeah,
if
you
have
a
phone,
then
you
can
tether
to
it.
That
might
be
quicker.
I'd
be
kind
of
wild.
Oh
yeah
I
mean
yeah.
If
you
have
a
USB
stick
and
you
want
to
sneakernet
it
over.
That
would
be
great.
That
would
be
super
helpful,
but
it's
just
that
one
image:
that's
right
there
in
your
temp
directory,
those
of
you
who
have
it.
If
you
have
a
if
you
have
a
USB
stick
and
you
want
to
hand
it
to
the
other
guy.
That
would
also
be
tremendous.
A
A
Don't
do
this
on
purpose
and
we
don't
do
this
on
accident.
Do
not
purpose,
but
don't
do
it
on
accident,
but
you
like,
because
my
computer,
in
the
process
of
easing
your
you've,
seen
a
local
doctor
to
do
all
kinds
of
wild
things
and
even
in
our
case
like
what
we're
actually
doing
a
lot
of
building
of
creepiness,
it
can
take
up
a
lot
of
space,
and
so
this
is
a
way
of
cleaning
it
all
up.
So
I
have
my
time
cache
hard,
easy
and
if
I
do.
A
Ellis
here
see
it's
one
point
four
gig,
so
this
is
the
image
that
you're
downloading.
If
you
don't
already
have
it,
it's
probably
already
still
downloading
we're
gonna
go
through
this
I'm
gonna
continue
from
here.
But
if
you,
if
you
need
help
just
raise
your
hand
once
you
get
it
downloaded,
we
can
make
sure
that
you
have
the
next
step,
which
will
be.
A
If
you
look
at
the
output
of
that
command,
the
reason
I
did
the
short
link
is
because
it's
so
much
shorter
than
actually
typing.
How
having
you
type
in
all
this,
like
from
my
voice,
I,
want
you
just
to
be
able
to
like
hear
what
I'm
saying
with
curl
L
kinda
case.
I
work,
slash
cash
and
then
pipe
that
to
bash,
and
it
will
run
it.
So,
let's
go
ahead
and
run
that.
A
A
So
once
you
actually
have
the
temp,
the
image
intent
that'll
be
in
slash,
temp
kind,
cache
tar.gz
your
next
command
will
be
docker
load.
I
it
should
have
actually
prompted
you
for
that
at
the
bottom
of
the
download
docker
load
I
and
then
pointed
at
that
image,
and
it
will
load
it
into
your
into
your
space.
That's
your
next
command.
A
We
can
see
we
have
kind
cache
locally
and
we
also
have
some
other
stuff
that
I
had
downloaded
earlier.
I
had
I
guess
it
didn't
wipe
it
all
out,
but
I
have
this
kind
cache
image
and
then,
when
I
expand
that
image
I'm
going
to
see
a
whole
bunch
of
other
images
come
out
of
it
because
effectively
I'm
using
docker
and
docker
to
load
all
those
images
from
the
cache
image.
And
if
we
look
back
at
the
slide,
we
can
see
what
we're
expecting
to
see
from
the
output
here
all
right.
A
A
A
Kind
thought:
okay,
it's
not
work
dyslexics
on
top
right
now
my
image
is
already
loaded
and
the
first
thing
it
does
is
it
does
a
docker
PS
just
to
make
sure
that
it's
connected
to
docker.
If
you
try
to
run
this
command
by
hand,
you
may
be
like
skip
to
the
point
of
actually
mounting
in
the
docker
socket,
that's
not
going
to
work
and
then
it's
going
to
go
ahead
and
load
all
of
the
images
necessary.
So
here
we
see
our
build
them
loading
up.
A
We
see
the
pause
container,
accordion
s
a
couple
of
Debian
containers
which
actually
make
up
kind
of
the
base
image
for
some,
some
of
those,
some
of
the
build
parts
of
it
and
then
the
big
one.
It's
kind
cross.
My
remember
Hercule,
but
we'll
see
you
here
in
a
second,
so
this
is
actually
the
output
that
you
should
see
on
your
screen.
A
You've
already
done
docker
load
and
now
you're,
actually
doing
curl
L
kind
cache,
so
I
kind,
CAD
start
work,
slash
cache
and
it
will
actually
expand
that
image
locally
and
you'll
see
all
of
these
images
come
out.
So
I'm
gonna
go
ahead
and
pair
here,
so
we
can
see
the
command.
This
is
the
these
are
the
two
commands
that
you're
gonna
run
next
right,
so
you're
gonna
do.
A
A
Once
you
have
that
image
loaded
into
docker,
then
you
can
actually
expand
that
image.
It
will
load
a
bunch
of
other
images
into
your
machine
that
will
be
curl,
L
kind,
Kate's,
not
work,
slash
cash
and
you
can
pipe
that
to
bash
or
you
can
just
copy
and
paste
it
whatever
you
feel
more
comfortable
with,
and
that
will
load
all
of
the
images
necessary
to
work
through
the
rest
of
this
lab
right
there
on
your
machine
when
you're
done.
A
A
There's
anybody
struggling
with
this.
Please
help
your
neighbor
right
like
y'all
are
in
this
together
we
got
90
minutes.
I
want
to
see
all
of
us.
Do
it.
I
would
love
to
see,
like
a
bunch
of
us,
actually
achieve
this
at
the
end
of
the
session,
so
once
you've
actually
downloaded
the
image
and
again,
if
you,
you
know
what
I'm
gonna
do
just
to
make
this
easy
I'm
gonna
do
this.
B
A
Are
the
three
commands
that
you
should
have
run
since
you've
come
in
the
door
all
right?
The
first
one
will
actually
help
us
download
the
image
that
you
need.
The
second
one
will
load
that
image
that
you've
downloaded
into
docker.
The
third
one
will
expand.
It
I
realized
that
some
of
you
are
still
downloading
I
apologize
for
that
I
hope.
That's.
It
gets
done
sooner
than
later.
A
What
we
talked
about
earlier
was,
if
you
have
a
USB,
stick
with
you
that
you
trust,
perhaps
it
might
be
worth
like
handing
it
over
to
somebody
who
already
has
the
image.
It's
the
same
image
for
everybody.
So
if
you
wanted
a
sneakernet
it
that
might
be
a
little
quicker
download
speeds
from
usb
or
generally
always
going
to
be
faster
than
trying
to
download
it
from
the
internet.
B
C
A
With
you,
mush,
mush
I
know,
I
want
to
hear
about
what
I
actually
did
to
make.
This
work
is
kind
of
wild,
so
this
image
what's
happening
inside
of
this
image.
When
you
look
at
the
output
I'm
using
a
Linux
download
manager
to
download
that
1.4
Giga
Mitch
in
AWS
uswest,
one
I
have
1025
gigabit
interface
VMs
that
are
hosting
exactly
one
file,
and
that
is
the
file
that
we
downloaded
and
what
Aria
does
is
I
can
provide
all
of
those
hosts
and
it
will
basically
chunk
download
across
all
ten
right.
A
So
the
goal
was
to
kind
of
get
the
ability
to
make
sure
that
you
could
actually
download
from
ten
hosts.
You
could
pull
it
down
bit
by
bit.
You
know
in
little
pieces
rather
than
one
big
piece,
so
if
it
gets
interrupted
or
whatever
we
won't
get
messed
up
right,
it'll
validate
that
the
chunk
of
got
is
the
right
chunk,
put
it
all
back
together
at
the
end
of
the
day
in
your
temp
directory,
this
seems
to
be
working
okay,
but
sometimes
for
something
out:
okay,
I!
Don't
get
the
two-hour
thing
that
blows
my
mind.
A
A
After
the
session,
if
you're
interested
I
have
a
whole
bunch
of
kind
stickers,
I
have
a
bunch
of
stickers
that
actually
represent
the
kind
logo.
They're
gonna
be
here
on
the
stage.
So
if
you
you
want
to
come,
get
a
sticker
while
you're
waiting
feel
free
to
do
so
come
on
up,
get
a
sticker
or
get
some
stickers.
I
have
a
bunch
of
them.
They're
tiny
they're,
like
one
inch
by
one
inch,
so
pretty
fun.
You
know.
A
All
right,
hoping
we
get
everybody
to
this
point.
Somebody
was
asking.
So
why
are
we
doing
the
curl
thing
and
the
reason
is
because
I
don't
expect
that
anybody's
gonna
stay
with
me
if
I
were
to
say
I
want
you
to
type
docker,
run
RM,
but
now
I'm
not
doing
it
right.
So
instead,
what
I
did
was
I
just
basically
used
Jekyll
inside
of
my
github
repository
and
I
made
a
file
called
cache
and
I
served
it
with.
A
You
know,
github
Docs
and
what
that's
doing
is
basically
giving
me
a
way
to
express
a
command
in
a
way
that
you
might
be
able
to
pipe
it
to
bash,
makes
it
very
simple:
you
can
copy
and
paste
it.
It's
actually
on
an
HTTP
secured
website,
I'm
using
let's
encrypt
for
the
for
the
encryption,
but
that's
why
it
curl.
That's
why
it's
a
remote
call
right,
I
just
want
you
to
be
able
to
pull
that
line
and
apply
it.
I!
A
Alright,
how
many
of
y'all
have
all
of
the
things
loaded
how
many
of
y'all
do
not
have
all
the
things
loaded
wow?
That's
like
ten
people,
I
thought
that
number
would
be
way
higher
this
hour
in
the
in
this
session.
I
have
to
say,
I'm,
actually
pretty
stoked
that
it's
a
low
number
I
noticed
that
some
of
y'all
are
sitting
together.
Maybe
that's
just
a
bad
luck,
kind
of
part
of
the
area
you
know,
joking.
A
A
Again,
as
we
mentioned
before,
if
you
have
a
USB
stick
in
your
stuff
and
you
want
to
just
grab
this
image
from
your
neighbor,
he
probably
has
it.
So
if
your,
if
that's
a
thing
that
you
could
do,
that
might
be
faster
to
get
you
back
online
I'm
going
to
move
forward
here
in
just
a
minute.
I
can't
wait
for,
unfortunately,
I
can't
wait
for
everybody
of
those
who
won't
get
through
it.
So
it
may.
A
Again,
the
first
command
is
basically
just
downloading
a
fetcher
if
you're
walking
in
I
think
we
have
everybody
already
here.
So
all
right,
so
I'm
gonna
move
on
here,
because
I
don't
think
about
how
much
time
I
have
left
I
think
we're
doing.
Okay,
three
ten!
So
let's
get
this
thing
kicked
off:
let's
do
the
fun
stuff.
So
what
we're
going
to
do
next
is
we're
going
to
start
up
one
of
those
images
that
is
in
the
image
repository
here.
It's
the
build
environment,
we're
gonna
kick
that
off
and
then
we're
going
to
do
that.
A
A
That's
one
of
the
things
we're
going
to
do
we're
also
mounting
in
the
temp
directory.
The
reason
we're
doing.
That
is
because,
when
we
build
all
of
the
bits
all
of
the
kubernetes
parts,
part
of
the
build
process
is
to
our
sink
all
of
those
parts
out
of
a
volume
into
the
temp
directory.
And
then
we
actually
are
sync
all
of
those
bits
back
into
the
con.
A
You
hear
me
now
awesome
all
right,
so
that's
why
we
needed
to
mount
the
temp
directory
in
was
just
to
make
sure
that
the
bits
when
we
are
finished
compiling
them
are
available
to
you.
When
we
do
the
node
image,
build
we're
gonna
walk
through
that
here
in
just
a
second.
So
so
I
can
get
our
build
environment
started
up
here.
Some
of
you,
curl,
L
kind,
Kate's,
dot,
work,
a
domain
I'm
actually
pretty
happy
to
have
build
em.
A
A
B
A
A
So
these
are
the
commands.
You're
gonna
run
you're
gonna
go
ahead
and
do
your
curl
command
you're
gonna
go
ahead
and
exec
into
this
and
then
once
you're
in
there.
This
is
actually
gonna
where
we're
gonna
spend
pretty
much
the
rest
of
the
workshop,
so
I
want
to
get
as
many
people
in
the
room
as
I
can
to
this
place.
A
How
many
people
are
there?
Oh
yeah,
how
many
people
are
not
there?
How
many
people
are
not
there,
because
they're
still
downloading
all
right?
Okay,
y'all
are
awesome,
both
the
ones
that
are
still
downloading
and
the
ones
that
you
know
are
actually
in
the
session.
So
the
next
thing
I'm
going
to
do,
is
I'm
going
to
kind
of
give
you
a
little
overview
of
what's
happening
inside
of
this
container.
A
Now
you
may
have
noticed
that
the
process
that's
running
is
the
pause
container
and
that's
because
it's
a
container
right,
the
container
is
really
just
process
isolation,
so
I
needed
something
that
would
continue
to
run
all
the
time.
So
I
went
ahead
and
threw
the
pause
binary
in
there,
and
that
is
the
actual
running
command
for
this
container
image,
and
it
will
just
stay
running
forever.
Won't
it
won't
die
until
you
stop
it.
A
We
are
in
the
go
directory
that
is
expressed
by
the
go
path
right
so
inside
of
here
we
have
kubernetes
checked
out.
If
I
do
CD
source
kubernetes,
k8c,
Oh
Corrine,
it
is
I-
can
see
that
this
is
I,
have
a
check
out
of
kubernetes.
Now
this
is
a
shallow
check
out
so
fight
the
power
of
doing
git
fetch
or
any
of
that
other
stuff
that
you
might
be
thinking.
Oh
I'm
gonna
do
this
thing
because
it's
muscle
memory
don't
do
it.
This
is
all
already
cached
for
us.
A
So
if
I
do
get,
log,
I
can
see
that
I'm
on
master
and
that
master
is
tagged
and
if
I
do
get
branch,
I
can
see
that
there
are
two
branches.
There's
the
master
branch
and
the
fixed
branch.
We're
gonna
build
both
of
these
branches
in
this
session,
the
first
time
we're
going
to
build
master
branch.
So
let's
go
ahead
and
get
that
started.
A
A
A
So
now,
what's
going
to
happen,
first
is
we're
going
to
actually
build
a
an
environment
in
which
to
build
kubernetes,
it's
gonna
actually
bring
up.
Another
container
I
build
a
build
image
in
which
we're
actually
going
to
compile
all
of
the
bits
for
communities
and
we're
gonna,
compile
that
against
our
local
checkout.
So
if
we
were
to
look
at
some
of
the
values
for
this
cube,
build
container
we're
mounting
in
to
go
path,
we're
mounting
in
that
code
we're
going
to
build
that
code
inside
of
that
container.
A
This
is
going
to
take
a
second
and
we're
gonna,
see
it
go
through
a
few
different
permutations
of
stuff
right,
so
we're
one
of
the
first
things
it's
going
to
do
is
actually
take
a
look
at
the
other
containers.
It
may
already
be
running
inside
of
the
system
and
if
it
sees
any
of
them,
it's
just
gonna,
say
I'm,
not
I'm,
not
touching
this
one
we'll
leave
it
alone,
and
that
way
it
can
keep
track
of
like
what
it
is
reduced
and
what
you
have.
That's.
Why
there
that's?
Why
it's
going
through
that
process?
A
Then
we
go
ahead
and
we
sync
the
sources.
This
is
what
we're
talking
about
we're
gonna
our
sync
that
go
path
up
into
the
container
and
then
we're
gonna
go
ahead
and
do
a
build
anybody.
Who's
built
kubernetes
before
has
probably
seen
this
before
all
right.
I
do
a
lot
of
copy
we're
gonna
do
a
lot
of
cogeneration
for
a
lot
of
the
API
stuff,
that's
within
kubernetes
and
then
we're
gonna
start
building
targets.
A
We're
gonna
build
all
of
the
bits
necessary
from
the
source
that
you
have
locally
checked
out,
so
we're
a
DM,
cubelet,
peu
proxy
cube,
API
server,
all
of
those
pieces
and
then
we're
actually
gonna.
Do
it
in
two
phases.
In
the
first
phase,
we're
gonna
build
the
binary
parts
and
those
binary
parts
I
mean
like
these
are
the
bits
that
you're
actually
gonna
put
on
the
server
you're
gonna
operate
these
with
things
like
system
D
or
just
from
the
command
line.
A
So
like
cube
kettle
is
a
command
line
tool
but
cube
ADM
is
a
command
and
cubed
cubed
kettle
and
cube.
Idiom
are
both
like
command
line
tools,
but
cubelet
is
gonna,
be
probably
operated
by
system
deep
as
part
of
building
this
we're
gonna
actually
take
that
base
image
that
we're
that
we've
identified
earlier
in
this
process
and
we're
going
to
put
all
of
these
bits
in
a
new
layer
on
top
of
that
base
image,
and
this
is
going
to
give
us
everything
we
need
to
create
a
kubernetes
cluster
from
your
local
arena.
This
check
out.
A
So
how
I
got
here,
if
you
already
in
the
play,
if
you're
already
in
the
place
where
you
have
the
build
environment,
start
it
up,
then
you
would
be
able
to
exec
into
it
and
then
you
would
be
able
to
run
kind.
Build
node
image
image.
Local
master
I
know
that
I'm
moving
forward
for
some
people.
If
you
want
to
catch
up,
you
can
ask
your
neighbor
to
ask
people
around
you
or
if
you
raise
your
hand,
we're
gonna
come
around
and
help
you
out.
A
A
A
Just
to
reiterate,
what's
happening
here:
I'm,
actually
building
all
of
communities
against
a
local
shallow
clone
of
Cabrini
this
and
if
I,
were
to
modify
that
code.
Like
add
some
error
statements
or
some
or
some
debugging
statements
or
whatever,
because
I
want
to
understand
better
how
things
work,
I
could
actually
just
modify
the
code
directly
inside
of
that
directory
and
I
would
be
able
to
build
against
that.
So
the
reason
I
wanted
to
show
you
this
is.
A
This
is
actually
how
you
can
iterate
right
if
you're
trying
to
develop
or
modify
that
functionality
of
a
particular
component
within
the
kubernetes
database
or
between
the
coop
earning
this
code
base.
This
is
this
is
exactly
what
I
mean
by
the
power
of
this
particular
session
right.
You
could
modify
that
code,
build
it
see
if
it
actually
builds.
If
you
run
into
problems,
you'll
know
what
to
do,
because
it's
all
right
there
in
your
own
little
output,
you
don't
have
to
like
try
and
parse
what's
happening
in
the
end-to-end
tests.
A
A
A
But
the
build
command
if
you're
curious
about
it,
the
build
command
would
be
something
like
slash,
hack,
run,
sh
make
and
then
build,
and
then
what
there'd
be
like
a
capital?
What
and
then
the
actual
targets
for
the
things
that
we're
watching.
So
if
you
look
at
the
output,
I
came
here
if
it's
in
the
output
or
not,
but
it
is
actually
what's
happening
here,
we're
just
using
the
existing
build
tools
within
the
crew
breena's
codebase.
To
do
this
build
it
is
a
containerized
build.
A
So
it's
neat
because
I
think
in
all
you
need
is
docker.
You
don't
need
to
actually
all
the
other
stuff,
so
here
we're
building
all
of
the
other
bits,
we're
building
the
controller
manager,
we're
building
the
API
server,
we're
building
the
cloud
controller
manager,
the
scheduler
and
queue
proxy.
So
if
you're
developing
against
any
of
those
things
and
which
is
it's
just
what
we're
doing
in
this
session,
you'd
be
able
to
see
those
outputs.
A
A
This
is
that
the
bug
is
actually
in
queue
proxy
when
you're
running
IPV
s
mode
and
the
bug
is
that
if
you
are
in
queue
proxy
running
ipbs
mode
and
you
create
a
deployment
and
then
expose
that
deployment
with
a
service,
and
then
you
delete
the
service.
The
cluster
IP
associated
with
that
service
remains
bound.
Probably
not
what
you
want
right.
You
said:
delete
the
service.
The
cluster
IP
should
be
released,
so
the
neat
thing
about
that
is
that
the
fix
for
that
bug
is
actually
the
removal
of
code,
which
was
awesome
right.
A
The
rule
code
is
always
a
good
thing.
If
you
can
achieve
it,
and
so
we're
gonna
actually
in
we're
gonna
build
we're.
Gonna
show
the
fault,
we're
gonna,
understand
that
functionality
and
then
we're
going
to
build
again
with
the
new
fixed
version.
We're
gonna
bring
up
a
cluster
in
IP,
vs
mode
and
test
it
and
show
the
different
states
so.
A
A
So
all
that
code
generation
sticks
around,
but
a
lot
of
the
actual
building
of
binary
is,
does
not
stick
around
so
you're
gonna
rebuild
all
the
binary,
but
all
the
code
generation
remains
inside
the
cache
and
so
the,
but
but
to
pop
that
cache.
If
you
actually,
if
this
represents
a
new
git,
commit
the
new
hash,
then
that
means
that
the
cache
is
no
longer
viable
and
it
will
build
a
completely
new
build
container.
A
So
if
you're
just
modifying
code
on
the
fly,
you've
modified
a
bunch
of
files,
you
messed
with
some
things
and
you
just
do
another
cute
and
you
do
another
kind
to
build.
Node
image,
you're
gonna
leverage
that
existing
cache.
If
you,
if
you
do
a
git,
commit
against
that,
that
means
that
you're
gonna
get
a
new
commit
ID,
and
that
means
we're
gonna
killed
a
new
cache.
So
that's
one
of
the
things
that,
with
one
of
the
questions
now
we
got
our
image.
A
So
what
I'm
going
to
do
next
is
I'm
gonna
go
ahead
and
flip
to
the
other
branch
and
I'm
gonna
build
it
again,
so
that
we're
not
waiting
for
that
to
happen
so
I'm
gonna
go
into
Circe
case
communities
and
I'm
gonna
do
git
checkout
cuz
I'm
crazy,
like
that,
and
what
that
does.
Is
you
know
you
probably
are
aware
of
like
CD.
It
moves
you
back
to
another
directory,
get
checkout!
A
That's
that
I
learned
that
from
Chris
Nova
I
thought
that
was
really
cool
so
from
here
I
can
go
ahead
and
do
that
same
command.
That
I
ran
before
kind,
build
node
image,
I'm
gonna
give
it
a
different
tag,
because
otherwise
that
would
be
weird
so
I'm
going
to
go
ahead
and
again
get
check
out
into
the
other
branch
when
I
go
down
here
and
add
this
to
the
list
of
commands.
A
A
A
Let
me
get
somebody
over
to
you:
come
Ben
or
Antonio,
or
if
build
failed,
it
might
be
memory
or
CPU
like
if
you
don't
have
like
a
good
amount
of
memory
or
CPU
inside
of
your
docker
images,
come
out
wide
it
okay
should
do
it.
Yeah
did
you
have
more
than
one
CPU,
though
that's
surprising,
well
I
mean,
as
you
can
see
like,
it
works
pretty
well,
I
guess!
Maybe
you
get
bit
by
the
gremlins
who
knows.
A
So
once
we
have
the
master
tag
built,
we're
gonna
go
ahead
and
check
we're
going
to
move
into
source
kubernetes
our
k-style
kubernetes.
You
get
check
out.
That'll
move
us
to
the
other
branch.
You
could
also
do
get
check
out
fix
if
you
want
to
they'll
pull
us
into
that
other
branch
and
then
we're
gonna
go
ahead
and
build
another
node
image
with
the
new
functionality.
A
A
Yeah
I
just
wanted
to
get
them
build,
started.
I,
don't
want
to
wipe
out
the
command
so
soon
as
soon
as
this
build
is
not
able.
Definitely
look
at
that.
We're
also
gonna
look
at
the
commits
inside
of
the
kourounis
codebase
I'd,
like
in
the
PR
requests,
and
all
that
stuff
kind
of
look
at
that
feedback.
A
A
What
happened
here
is
that
this
this
engineer
determined
that
there
was
actually
kind
of
this
weird
behavior
right
that
when
they
created
a
kubernetes
deployment
to
have
a
service
and
they
deleted
the
service
that
the
IP
address
remained
around
and
they
started
out
by
creating
a
bug
that
was
associated
with
it.
And
so
the
bug
that
is
associated
with
it
I
think
is
actually
in
the
output.
For
some
reason,
I'm
not
seeing
it.
But.
A
The
bug
basically
highlights
the
the
actual
problem
and
describes
exactly
what
behavior
they
saw
so
in
queue
proxy
and
ipbs
mode.
It
describes
that
whole
behavior
and
then
the
engineer
went
on
to
actually
fix
it,
which
was
really
great,
and
so
you
can
see
in
this
pull
request
exactly
what
happened
throughout
the
lifecycle
of
this,
and
this
is
just
one
of
many
commits.
A
I
was
actually
trying
to
find
like
one
commit
that
I
could
easily
identify
and
show
you
like
all
the
changes
that
were
happening,
and
so
this
is
a
all
bundled
into
one
commit
makes
it
very
simple,
and
here
is
the
change
and
some
of
the
commentary
on
the
change
right.
So
this
is
a
this
represents
removing
19
lines
of
code.
A
We
can
see
that
it's
removing
19
lines
of
code,
not
adding
anything
except
for
better
functionality,
and
we
can
see
some
of
the
some
of
the
questions
that
have
that
happen,
or
some
of
the
conversation
that
happened
back
and
forth
between
Andrew.
Second,
who
I
work
with
at
VMware
and
the
engineer
in
question
right.
They
are
already
a
contributor
which
means
they've
contributed
to
the
codebase.
A
number
of
times
it's
unlikely
that
they
actually
follow
this
flow
to
go
through
it,
but
it
could
have
but
yeah.
They
actually
talked
about
like
why.
A
This
is
interesting,
why
it
happened,
how
to
get
how
it
got
fixed.
This
particular
issue,
if
you
go
to
issues
that
cage
to
I/o,
8
0,
9,
4,
2
that'll
be
the
actual
pull
request
associated
with
it.
So,
if
you're
curious
more
about
like
we
change
the
functionality
or
what
this
was
out,
what
this
is
about.
This
is
where
you'd
find
it
now,
while
we're
here
and
talking
about
this
stuff,
I
want
to
go
to
kubernetes,
kundunese,
codebase,
real,
quick
and
point
out
here
on
the
bottom
and
the
readme.
A
A
Can
we
really
focus
on
just
those
things
that
they've
agreed
to
really
focus
on
instead
of
having
to
look
at
the
entire
code
base
like
Jordan
Liggett
does
all
the
time
you
can
actually
just
look
at
your
own
like
thing,
it's
kind
of
expecting
a
laugh.
That
guys
are
machines,
though,
like
it's
amazing,
so
there's
also
different
work
groups,
that
kind
of
like
look
into
the
different
functionality
so
yeah.
A
This
commits
basically
removes
19
lines
of
code
very
cool
stuff,
so
we're
almost
done
with
our
second
build
here
are
I
am
I
hope
that
most
of
you
are
as
well,
and
then
we're
gonna
go
ahead
and
build
a
kubernetes
cluster
or
two
and
show
off
this
functionality
and
show
the
fix,
and
then,
after
that,
we'll
be
at
the
profit
stage
of
our
talk.
We'll
have
actually
completed
it
and
I
hope
that
that
will
work
for
you
and
then
we
could
actually
answer
questions
and
get
everybody
else
through
the
session
as
well.
C
A
B
A
A
A
That
just
came
up
was:
can
you
elaborate
a
little
bit
more
about
what
kind
build
note
images
happen
it
is
doing,
and
so
what
happens
is
a
when
we
ship
kind
we
ship
a
base
image
with
every
release.
Right
now,
we're
on
like
oh
six,
oh
and
that
base
image
represents
everything
needed
all
the
way
up
until
all
the
kubernetes
parts.
It
doesn't
really
include
any
of
the
kubernetes
bits,
just
the
the
underlying
base
image.
A
So
it's
a
it's
a
debian
base
and
it
has
all
of
the
bits
all
the
way
up
until
the
the
kubernetes
bits
when
you
do
kind
build
node
image,
we're
basically
firing
off
a
make
command
against
the
discovery
to
go
path.
So
if
you
have
your
go
path,
environment
very
variable
set,
we
will
actually
our
sink.
The
code
found
inside
of
that
go
path
for
kubernetes.
So
if
it
were
slash,
go
/k,
h,
io,
/
kubernetes.
A
We
are
sink
that
entire
code
base,
with
whatever
check
out
you,
have
whatever
files
are
there
into
a
build
into
a
build
image
and
then
we'll
run
the
make
commands
make
for
for
each
of
the,
but
for
each
of
the
binary
bits.
The
cubelet
cube,
CTL,
cube,
ADM
and
also
the
container
images,
so
controller
manager,
API
server,
the
proxy
cog
controller
manager,
and
then
we
will
bundle
into
that
same
docker
container.
All
of
the
bits
like
etsy
d
and
cute
and
kind
net,
which
is
a
CNI
developed
by
antonio
who's.
A
Here
he's
actually
here
in
this
room,
be
developed.
It
has
a
very
simple
c
and
I
to
help
us
do
end-to-end
testing
it's
one
of
the
first
ones,
and
it's
actually
used
it's
one
of
the
first
ones
that
supports
kind
of
ipv6
in
a
really
neat
way,
and
it's
actually
also
one
the
one
that
we
use
to
actually
validate
ipv6
PRS
against
Cooper
needles,
which
is
pretty
neat.
A
A
So
when
you
do
kind,
build
node
image,
we're
basically
making
a
brand
new
docker
image,
all
that
base
stuff
and
all
of
the
bits
that
we
built
when
you
did
kind
of
build
image,
and
then
we
tagged
it
with
whatever
you
have
actually
chosen
to
target
with.
So
my
next
command
I'm
actually
going
to
create
a
cluster,
but
before
I
do
that
I'm
going
to
look
at
the
configuration
of
the
cluster
that
I'm
going
to
create.
A
Kind
config
dot
y
mo-
and
this
is
a
file-
that's
here
in
the
go
path
right.
So
if
you
take
a
look
at
kind,
config
ya,
mo
right
there
in
slash,
go
you'll
find
that
yeah
mo
file-
and
here
is
where
we're
actually
kind
of
expressing
some
of
the
power
of
kind
in
our
test.
What
we
want
to
show
is
I
cube
proxy
in
IP
vs
mode.
A
So
for
us
to
do
that,
we
need
to
enable
IP,
vs
mode
and
the
way
that
we
at
providing
a
cube,
ATM
patch,
that
turns
on
IP
vs
mode,
and
this
is
actually
leveraging
kind
of
the
customized
capability
and
I
think
to
modify
cube
ATM
at
instantiation
right.
So
when
you
actually
do
a
kind
create
cluster-
and
you
specified
this,
this
type
of
a
patch
then
what's
gonna
end
up
doing
is
we're
gonna
merge.
This
configuration
as
part
of
the
default
cube,
ATM
configuration
for
the
cluster
you're.
A
Creating
the
power
of
this
is
that
you
can
actually
modify
anything
that
cube.
Atm
can
modify
using
this.
So
that
means
that
you
could
you
could
do
anything.
I
mean
there.
I've
got
a
bunch
of
examples
up
on
my
github,
so
if
you
go
to
github.com,
slash
maui
lion
and
you
look
for
the
kind
repositories,
there's
a
ton
of
different
examples
for
how
to
do
this
stuff.
A
A
So
I've
got
a
command
here
that
says
kind
create
cluster
I'm,
a
passive,
the
config
file,
which
is
in
the
go
directory
right.
So
it's
gonna
be
time,
create
cluster
config
kind.
Config
yamo
image
equals
local
:
master,
that's
the
master,
build
that
I
had
and
then
I'm
gonna.
Let
that
get
started
up
and
we're
going
to
talk
through
what's
happening
here.
While
this
is
starting
up
right.
So
the
first
thing
we
do
is
we
create
the
docker
container
that
represents
that
node,
the
one
control
plane
node
that
we're
creating
the
second
thing
we
do.
A
Is
we
munge
any
of
those
patches
that
we
saw
in
that
qubit
in
that
kind
of
config,
with
the
cube
Adium
configuration
of
the
host,
and
then
we
actually
present
that
cube
adium
config
within
the
node,
so
that
we
can,
when
we
do
cube
ATM
in
it
all
that
configuration
is
made
available
to
the
to
the
instance.
The
next
thing
we
do
is
we
install
a
default.
A
After
that
we
go
ahead
and
install
the
default
C&I
and
then
we
default,
and
then
we
install
a
storage
class.
So
in
this
storage
class
is
just
using
a
dynamic
local
host
provisioner.
So
it's
actually
any
PB
that
you
create
it's
just
gonna
be
made
available
in
slash
temp,
so
it's
it
doesn't
do
cross
node
provisioning
yet,
but
at
the
same
time
you
could
use
it
to.
You
know,
evaluate
other
tools
for
that
sort
of
stuff.
A
So,
to
backtrack
a
little
bit
once
I
actually
have
the
the
second
image
built,
I've
looked
at
the
kind
of
config
and
then
I'm
going
to
create
the
first
cluster,
which
shows
the
broken
state
and
then
I'm
going
to
exec
into
it
and
show
that
broken
state
right.
So
first,
let's
do
docker
exact
yargh.
Actually
what
I'm
gonna
do
next
is
I'm
going
to
do
kind,
get
cube,
config.
A
A
Okay,
I'm
gonna
put
that
into
dot
cubed
config.
Just
to
make
my
life
easier
now
by
default.
Cube
kettle
will
look
for
a
file
at
dot
in
at
slash
home,
dot,
cube
config,
and
it
will
use
that
to
authenticate
to
your
cluster
so
once
that's
there,
I
should
be
able
to
do
get
pods.
A-And
I
can
see
all
the
pods
running
inside
of
this
single
node,
kubernetes
cluster
and
I
can
see
that
core
DNS
has
started
up.
A
A
There
go
so
I
did
my
create
cluster
against
the
local
master
check
and
with
a
local
master
image,
and
then
I
went
ahead
and
did
a
cute
I
got
the
cute
config
for
it
and
they
threw
it
into
dot,
cube,
config
and
now
I'm
able
to
actually
interact
with
the
thing
so
I
can
actually
do
cube.
Get
I'll
get
pods.
I
can
see
all
of
the
pods
across
the
whole
thing.
Eh.
If
you
didn't
know,
it
is
actually
a
relatively
new
short
thing.
A
A
A
And
what
this
is
doing
is
it's
going
to
create
a
deployment
of
the
pause
image
just
to
give
me
something
to
surface
right,
I'm,
not
actually
trying
to
show
off
functionality
here,
I'm
trying
to
validate
ipbs
mode,
so
I'm
just
gonna
create
a
container.
It's
the
pause
container
it's
running
locally
inside
the
cluster
makes
it
simple
now
I'm
going
to
expose
that
keep
it'll
expose,
deploy
test
port
equals
80.
A
We
know
that
it's
not
listening
on
port
80,
but
you
know,
bear
with
me:
let's
not
create
a
deployment
and
I've
created
a
service
that
exposes
that
deployment
and
now
I'm
gonna
jump
into
the
container
image
and
take
a
look
around
so
docker
exec
T
I.
Remember:
all
of
these
commands
are
in
our
slide
deck.
So
if
you
want
to
follow
along
and
you're,
not
with
me
completely,
you
can
go
to
kind
K
at
work.
All
the
commands
are
in
a
slide
on
that
deck.
B
A
Okay,
so,
and
that
command
is
actually
I
should
have
made
a
short
command
for
that.
That
would
have
been
smarter
me,
but
you
know
how
it
is
anyway,
so
that
doctoring
section
is
actually
on
the
slide,
though,
and
you
can
see
it
and
basically
what
I'm
trying
to
see
in
this
output
is
the
cluster
IPS
that
are
associated
with
the
ipbs
interface,
and
the
thing
about
this
is
that
the
output
tells
me
one
I
have
succeeded
in
running
ipbs
mode,
two
that
I
can
see
the
cluster
IPS
that
are
associated
with
it.
A
If
you
ever
wondered
why
you
can
actually
suddenly
ping
service
IP
addresses
if
you're
an
IPS
mode,
because
you
know
how
many
people
know
that
you
can't
ping
service
IP
addresses
if
you're
not
in
IP,
vs
mode
cool,
you
can't
because
some
IP
v-
yes,
it's
an
IP
tables
trick.
So
by
default
you
can't
ping
a
service
IP.
You
can
only
ping
on
pot
IP,
because
really
we
didn't
ever
like
create
IP
tables
rules
to
actually
forward
ICMP
to
like
those
knock
back,
endpoints
kind
of
a
fun-packed
alright.
A
So
now
what
I'm
gonna
do
is
I'm
gonna.
Go
ahead
and
delete
that
service
cube
kettle,
delete,
SVC
test
and
what
I
am
expecting
to
see.
What
I
want
to
see
is
that
the
IP
address
associated
with
that
service
becomes
unbound
from
the
ipbs
interface
right.
So
let's
take
a
look
at
that.
If
I
still
see
this
IP
address
associated
with
it,
that
I
know
that
it
didn't
work
right,
the
up
arrow
and
run
that
command
again
and
sure
enough.
I
still
see
that
work.
That's
that
cluster
IP
working,
even
though
it's
not
defined
right.
A
The
service
that
associated
that
IP
address
with
that
interface
has
been
deleted
so
that
IP
address
should
be
released,
but
it
hasn't
been
that's
the
bug,
so
our
fix
is
to
remove
the
19
lines
of
code
that
was
keeping
it
around
now.
What
we're
gonna
do
is
we're
gonna,
create
a
new
cluster
and
show
the
change
in
functionality.
How
many
of
you
all
are
with
me?
A
Alright,
I'm
gonna
do
a
kind
delete
cluster
broken
anyway
and
then
I'm
going
to
go
ahead
and
do
my
kind
build
cluster
or
kind
create
cluster
again.
Only
this
time,
I'm
going
to
use
my
new
tag,
which
represents
the
new
change
right,
so
I'm
gonna,
say
kind,
create
cluster
config
same
config
because
it
still
puts
it
in
IBBs
mode.
The
only
difference
in
this
command
and
the
reason
and
the
last
one
is
they're
going
to
point
to
that
other
tag.
The
fixed
tag,
then
we're
going
to
kick
this
up.
A
Now,
when
you
build
a
we'll
talk
about
it
here,
a
little
bit
more
in
just
a
second,
but
when
you
build
these
things,
there's
a
bunch
of
stuff!
That's
left
on
the
disk
and
I'll
talk
to
you
about
how
to
clean
all
that
stuff
up.
But
so
let's
go
ahead
and
fix
our
cube
config,
just
like
we
did
before
no
change
in
the
command,
but
we
need
to
update
it
because,
obviously,
all
the
certificates
from
the
old
cluster
are
now
gone
because
we
really
did
it.
So
we
need
to
create
a
new
one.
A
So
let's
go
ahead
and
create
our
deployment,
just
like
we
did
before,
and
our
service
I'm
just
up
arrowing
into
my
history
here.
So
you
should
be
able
to
do
that
part
easily
as
well,
but
I
think
I
went
too
far.
There
we
go
and
then
I'm
gonna
go
ahead
and
delete
I'm
gonna
do
go
ahead
and
validate
that
the
configuration
is
actually
on
the
node.
So
let's
take
a
look
and
then
we
see
again
that
there
is
another
IP
address
associated
with
the
IPPS
interface
and
if
we
delete
that
service.
A
You
see
the
IP
address,
get
released
all
right,
so
this
is
the
punch
line
you
can
iterate
over
this
stuff.
You
don't
have
to
get
it
all
right,
all
the
time
you
can
just
continue
to
iterate.
Until
you
get
it
right,
you
can
get
it
working
locally.
You
can
feel
good
about
it.
You
feel
confident
in
your
change
and
you
can
ship
that
change
in
a
PR.
A
We're
at
time,
okay,
so
the
last
thing
I
wanted
to
show
you
was
the
cleanup
right
so
and
basically
it's
in
my
slides
as
well.
So
if
we
wanted
to
do
it,
go
ahead
and
docker
stop
to
build
them.
That
will
remove
the
image
and
then,
if
we
do
cube
kettle
or
sorry
docker
images,
we
can
see
cuz.
You
know
some
of
the
stuff
that's
left
behind.
So
there's
some
dangling
volumes,
that's
what
these
things
are.
There's
some
bingling
build
images
and
these
things
all
take
up
a
good
amount
of
space
right.
A
So
if
you
want
to
clean
that
stuff
up
the
commands
to
cleaning
it
up
are
locally,
for
my
part,
I'm
just
going
to
do
a
docker
system
prune
all
volumes,
and
that
removes
everything
and
that
cleans
up
everything
after
behind
it
and
we
can
see
how
much
data
we
used.
We
basically
used
18
gig
worth
of
space
to
do
all
of
this
testing.
So
I
hope
you
all
learned
something.
A
Thank
you
so
much
for
your
time.
I
know
it
was
a
lot
of
time
and
there
was
a
lot
of
quick
pace,
stuff
and
lots
of
CLI
stuff
check
out
the
sides
we're
actually
going
to
build
a
full-on
tutorial
on
kind
6kh,
Daioh
much
like
some
of
the
other
tutorials
that
I
built
before
so
you'll,
see
that
there
after
Cuba
Kahn
is
over
and
we've
had
a
chance
to
breathe.
Thank
you.
Thank
you.
Thank
you.