►
From YouTube: TGI Kubernetes 118: kubectl plugins and Krew
Description
UPDATE: Change of plans! Joe had a meeting conflict pop up last minute so Duffie is going to jump in and get started on this topic. We will be experimenting to see if Joe can join live as soon as he is done with his meeting.
Come hang out with Joe Beda as he does a bit of hands on hacking of Kubernetes and related topics. Some of this will be Joe talking about the things he knows. Some of this will be Joe exploring something new with the audience. Come join the fun, ask questions, comment, and participate in the live chat!
This week we'll be digging into kubectl plugins and how they work. We'll look at Krew, the plugin manager and try to find time to look into some (audience) favorite plugins.
A
A
B
C
B
C
B
B
B
You
know
it's
it's!
This!
It's
I
was
actually
just
talking
to
a
friend
Paul
Zukowski
earlier
about
this,
because
it
was
like
one
of
those
challenges
where
I
realized
that,
like
you
know,
in
every
operating
system,
for
some
reason,
I,
don't
think
we've
really
solved
a
sound
and
that
sort
of
thing-
that's
always
there's
always
something
funky
happening
somewhere.
That's
kind
of
weird
anyway
good
to
see
you
all.
So.
B
Thank
you
very
much
for
letting
me
know
that
audio
was
not
working
and
then
it's
working
now
so
I
started
off
by
basically
saying
that,
like
you
know,
I
was
really
looking
forward
to
seeing
Joe.
Do
this
episode,
but
he
unfortunately
got
pulled
into
some
other
highly
important
business
ii
thing,
and
so
it's
you're
stuck
with
me
again
this
week,
so
we're
gonna
explore
the
we're
gonna
explore
crew
and
keep
kettle
plugins
and
we're
gonna
get
into
that
kind
of
thing.
B
So
let
me
go
back
to
my
hellos
good
to
see
you
all
Suresh
Linate,
Rory,
Abhishek
and
Ahmed
good.
Thank
you
very
much
for
being
here
got
Martin
and
Enver.
We
got
Roger
ohm
and
Dinesh
and
Vivian
telling
me
I,
muted,
a
bunch
of
people
tell
me
I'm,
muted.
Thank
you
so
much
for
the
feedback.
It
really
helps
I.
Think
we've
got
it
all
figured
out.
Now,
though,
so
that's
good
yeah,
the
laptop
speaker
without
the
without
the
sock
is
probably
way
too
windy.
Mr.
B
Joe
Searcy
saying
hello
and
Ed
saying
the
sound
is
cutting
in
and
out.
That
was
probably
the
issue
looks
like
we're
all
in
good
shape.
Now,
though,
Angra
Gore
saying
thank
you
for
all
the
TGI
k-series,
did
you
guys
serious
I
really
appreciate
doing
them?
It's
like
it's
a
super
fun
thing
super
fun
part
of
my
week
and
it's
awesome.
You
know
just
to
be
a
part
of
this
community
and
like
give
back
some
of
that.
So
that's
awesome.
B
Federico
from
saying
hello
from
Sweden
Bogdan
say
hello
from
Bucharest
and
Marko:
hey
Marko,
good,
to
see
you
and
he's
in
Belgrade
all
right,
let's
get
to
the
third
screen,
which
should
be
the
screen
and
the
cam.
So,
let's
make
sure
that's
working
and
it
is
alright
cool.
So
this
week
we're
gonna
do
cube
kettle.
Plugins
and
crew
is
episode
118.
B
If
you
want
to
participate
in
the
in
the
chat
in
the
in
the
notes,
you
can
find
it
online
at
TTI,
K,
dot,
io
/
notes,
just
like
always,
let's
take
a
look
at
what's
happening
in
the
space,
we're
still
on
119,
alpha
3
node
and
changes
to
report
this
week.
Enhancement
freeze
is
next
week.
So
if
you're
trying
to
get
something
in
to
our
kubernetes
next
week
is
the
freeze
for
that
this
week
in
community
we
have
a
few
updates,
which
will
be
pretty
exciting.
B
We've
got
harbor
2.0,
taking
a
giant
leap
in
expanding
supported,
artifacts
with
OCI
support,
meaning
that
we
can
support
more
than
just
docker
containers,
pretty
exciting
stuff.
It's
a
no
CI
compliant
registry.
Now
so,
if
you're
interested
in
harbor
or
playing
with
it
or
checking
it
out,
there's
pretty
exciting
stuff
happening
in
the
space.
B
You
can
now
actually
host
helm,
charts,
get
home,
3
we're
shipping,
oh
cool
I.
Remember
this
actually
started
as
a
tweet,
harbor
and
aqua
and
aqua
trivia
working
together.
I
think
I
think
this.
Actually
this
project
started
as
a
tweet.
If
I
remember
correctly
could
be
wrong,
but
I
think
this
4
comes
from.
So
this
is
now
using
aquas
trivia
scanner
piece,
that's
pretty
exciting.
B
They
look
at
the
upgrade
path
from
1
1
X
2
to
X
I
had
I,
certainly
haven't
I
know
that
it's
I
suspect
that
it's
possible
big
community
check
out
shoutouts
to
Luke
Perkins
for
the
new
snazzy
go
harbor
website.
We
got
a
bunch
of
other
folks
defending
Udall
for
nothing
phew,
a
bunch
of
other
folks
in
the
community
really
helping
out
with
the
project.
This
is
a
CN
CF,
hosted
project.
So
it's
really
great
to
see
the
community
engaging
on
it.
Other
notable
features
we
got.
Oh
now
we
have
a
dark
mode.
B
B
B
Can't
hear
clearly
mm-hmm
everybody
else.
Sound
okay,
charity
check
everybody.
Alright
I
can
turn
this
up
a
little
more
here.
How
about
that?
It's
not
any
better,
all
right
cool.
Just
let
me
know
if
I'm
coming
in
too
hot,
alright
great
article
from
mr.
Alex
Ellis,
one
of
the
one
of
the
awesome
one
of
the
awesome
folks
in
our
community.
B
He
wrote
up
a
bare-metal
in
a
cloud
native
world,
pretty
good
article
describing
like
what
do
you
think
of
when
you
think
of
gar
metal
and
like
some
of
the
work
that
he's
done,
he's
actually
written
a
bunch
of
stuff
on
actually
leveraging
raspberry
PI's.
So
if
you're
interested
in
raspberry,
PI's
and
and
that
sort
of
work
definitely
check
out
Alex
Alice's
stuff,
he
also
works
on
open,
faz
and
a
bunch
other
kind
of
crazy
stuff.
So
really
cool
stuff
talks
about
functions
as
a
service
and
talks
about
the
surveillance
project.
B
Kind
of
describing
some
of
the
general
terms
that
we
unfortunate
operators
think
about
when
in
terms
of
bare
metal
describing
some
of
the
components
that
we
have
to
deal
with
on
a
daily
basis,
pretty
cool
stuff
talks
about
nut,
booting,
so
yeah,
if
you're
interested
in
bare
metal
and
that
whole
space
check
that
out,
there's
actually
another
project
that
I've
been
starting
to
explore
and
the
bare
metal
space
that
might
be
worth
checking
out.
Also,
if
you're
interested
in
tinkerbell
org
I
was
actually
thinking
about
doing
a
TGI
cap.
B
Cjk
episode
on
this
as
well,
but
T
Tinkerbell
is
a
set
of
services
that
can
be
used
as
a
as
an
API
driven
provider
for
bare
metal
machines,
right
and
so
I
I
was
thinking
about
setting
up
something
like
this,
using
like
a
virtual
box
or
vSphere
workstation
environment,
in
which
I
can
actually
have
these
virtual
machines,
like
just
booting
off
of
the
network
and
then
using
Tinkerbell
to
kind
of
explore
that
stuff.
So
that's
definitely
one
of
the
things
I'm
very
interested
in
checking
out.
That's
something!
That's
interesting
to
you.
B
Just
let
me
know
in
the
chat
Michael
Lerner
has
written
up
understanding
rough
consensus.
We
talked
a
little
bit
about
raft
in
the
sed
episode
this
last
time.
We
kind
of
explored
how
it
works
a
little
bit,
but
this
seems
like
a
pretty
good
collection
of
information
around
raft
and
how
it
works.
So,
if
you're
interested
in
understanding
a
little
more
about
it,
this
I
don't
think,
gets
into
the
detail
of
how
sed
itself
implements
it.
B
B
B
Talks
about
the
different
templates
for
what
he's
doing
he's
written
up
a
retrospective
on
it,
pretty
good
article.
You
know
kind
of
describing
like
how
you
can
get
from
one
to
the
other,
so
this
might
work
for
some
people,
kind
of
exploring
a
different
method,
I'm
a
different
method
for
how
you
can
actually
make
a
jump
from
your
existing
hosting
solution
to
this
one
CD
kate's.
What
is
Kate's
die?
Oh
she
decay
for
kubernetes,
define
kubernetes
Afton
components
using
familiar
languages.
B
The
cloud
development
kit:
why
would
you
use
cdk
harness
for
pipe
I?
Don't
actually
know
what
this
one
is
works.
Recruitment
is
everywhere.
You
can
use
CD
Kate's
to
define
any
application
for
any
kubernetes
cluster
running
anywhere,
including
cloud
or
on-premise
thing
has
a
language
support.
What
are
we
looking
at
here.
B
B
So
they're
building
out
tooling
that
allows
you
to
describe
pods
and
services
and
first-class
kubernetes
objects
and
other
objects
within
the
world.
So
here's
an
example
of
a
basic
Kubuntu.
This
object
written
in
typescript,
so
importing
the
deployment
object
and
then
defining
this
stuff
as
code.
So
this
I
mean
it'd
be
interesting.
I
imagine
this
is
something
of
a
competitor
to
folks
like
Kalu
me
and
some
of
the
stuff
that
terraform
is
working
on
lately.
Like
those
folks,
I
imagine
probably
view
this
as
a
competitor
as
competition.
B
B
C
B
B
B
B
B
I
see
yeah.
This
is
actually
a
very
good
point.
So
what
they're
describing
is
you
may
want
to
know
that
there
is
a
resource
that
you've
deployed
into
your
kubernetes
cluster,
that
at
deploy
time
was
defined
at
a
previous
version,
and
this
is
really
useful
if
you're
doing
things
like
migrating
from
1/16
to
117,
and
you
want
to
make
sure
that
all
of
your
source
manifests
are
able
to
make
that
jump
right
and
the
way
that
you
can
detect
that
in
some
cases,
not
in
every
case.
B
B
B
So,
let's
bring
up
a
cluster,
a
quick
and
I'll
show
you
what
I'll
show
you
kind
of
what
I'm
talking
about
here.
It's
actually
some
interesting
changes
in
the
1/18
resource
piece
anyway,
but
while
that's
happening,
let's
keep
reading
so
basically
they're,
just
basically
they're
trying
to
find
a
way
to
like
expose
some
of
the
some
of
the
configuration.
That's
happened
inside
of
your
existing
cluster
I.
B
Think
I
pointed
this
out
before,
but
if
I
haven't
another
one,
that's
really
good,
for
this
is
Pluto
from
Fairwinds
ops,
and
this
was
I
think
we
talked
about
this
one
in
the
last
episode.
But
Pluto
is
another
one
of
the
tools
that
is
actually
out
there.
That
can
help.
You
understand
whether
the
resources
that
you
have
on
disk
before
you
apply
them
to
the
cluster
are
actually
using
deprecated
versions
or
not
so
like
having
something
in
this
like
this
in
your
pipeline,
I.
B
C
B
B
Jet
stack
and
verify
forces
to
bring
join
forces
to
bring
machine
identity
protection
to
the
cloud
native
stack,
so
I
was
actually
just
talking
to
mr.
James
Nunnally.
He
was
a
good
friend
in
the
community
and
I
was
telling
him
congrats.
This
seems
like
it
has
to
be
just
a
huge
validation
of
all
the
work
that
you
and
the
community
have
put
into
things
like
cert
manager,
so
to
see
Ben
if
I
acquiring
jet
stack
is
awesome
and
it's
you
know
it's
it's.
B
It
sounds
like
it's
a
really
good
situation
for
everybody,
so
congrats
all
around
it
was
started
by
Christian
Simon.
Our
first
employee
I
was
cube,
Lego
I,
remember,
cube,
Lego,
cube,
Lego
came
before
cute,
assert
manager,
yeah,
cert
manager,
success
driven
and
pioneered
by
James
Nunnally,
oh
I,
think
was
also
involved
in
the
in
the
in
cube
Lego
at
that
time.
So
yeah!
Well,
if
you
want
a
little
more
history
about
what's
happening
there
and
what
they're
gonna
do
definitely
check
out
the
get
to
Jed
stack
blog
but
super
exciting.
B
To
hear
does
that
happen,
I
mean-
and
it's
just
great,
to
see
like
validation
of
some
of
the
some
of
the
new
like
PKI
and
TLS
panels
that
are
emerging
from
this
whole
crew.
Bring
this
world
like
the
ability
to
you
know
like
obviously
inside
of
kubernetes,
especially
as
we're
defining
new
web
services
and
we're
defining
new
services
that
have
other
requirements
for
certificates
being
able
to
have
tooling,
like
cert
manager.
It's
pretty
critical
to
really
like
to
reducing
the
friction
around
PKI
and
TLS
in
general.
B
So,
if
that's
something
that
you're
interested
in,
I
hope
that
I
hope
that
you'll
check
out
the
blog-
and
I
just
want
to
say
it
congrats,
one
more
time
to
everybody
in
that
community
and
all
of
the
work
that
they've
done
seems
just
tremendous
the
next
one
up
on
our
list
is
from
mr.
Brad
D
Siemens
and
he
is
has
written
a
cubicle
plugin.
So
let's
kind
of
talk
about
this
and
play
with
it
a
little
bit
so
and
then
we're
gonna
and
then
we're
gonna
get
into
like
some
of
the
other
pieces
here.
B
We
can
look
at
crude
at
SIG's
to
IO.
We
can
look
at
this
presentation.
This
is
a
great
reference
link
actually
on
that
who's.
Part
of
our
chat
today
did
a
great
presentation,
I'm,
not
sure
where
you
presented
this
on
that,
but
I
just
actually
drifted
through
the
presentation,
and
it
looks
really
solid
about
effectively
like
why
about
making
cube
kettle
plugins
mainstream,
and
he
really
highlights
all
of
the
highlight
all
of
the
high
points
about
exactly.
Why
cube
kettle.
B
So,
if
you're
interested
in
that,
oh,
oh
and
San,
Diego
I,
somehow
missed
that
one
but
definitely
check
it
out,
and
then
he
also
has
a
few
links
to
keep
cattle
plug.
In
so
like
some
naming
guide
are
practicing
the
best
practices
for
developing
one
and
and
for
Duran
plug-in
development
itself.
We're
gonna
kind
of
explore
those
things
and
then
interesting
issues
to
follow
our
custom.
Plugin
indexes
I,
don't
know
what
that
one's
gonna
be.
B
But
let's
get
into
this
one,
because
first
I
kind
of
want
to
highlight
this
kind
of
what
I
think
is
a
pretty
important
point
right,
which
is
that
the
way
cube
kettle
plugins
work
in
general,
and
this
has
nothing
to
do
with
crew.
This
has
nothing
to
do
with
like
the
work
that
I'm
at
has
done
with
crew
or
any
of
that
stuff,
but
just
kind
of
a
caveat
emptor
around
plugins
in
general,
like
this
is
the
thing
that
I
really
want
to
make
sure
that
folks
understand
about.
B
A
C
B
B
B
Typically,
if
you're
applying
resources,
you
can
actually
see
the
way
that
that
was
defined
at
that
time
so
like
in
our
case,
we
can
see
that
it
was
a
kind
deployment
and
we
can
see
that
the
API
version
of
this
resource
what
it
was
defined
when
it
was
put
into
the
cluster
as
apps,
be
one
or
actually
you
know
when
it
was
applied
by
pubic
at
all.
It
was
apps
for
us.
B
So
this
is
one
of
the
resources
that
tooling
like
that,
can
actually
look
at
to
validate
whether
the
thing
defined
inside
the
cluster
originated
from
a
from
an
unsupported
version
or
an
unsupported,
unsupported.
Sorry,
this
is
a
way
that
we
can
define
whether
we
condition
the
way
that
we
can
determine
if,
when
the
resource
was
applied,
if
it
associated,
if
it
was
associated
with
an
API,
a
version
or
group
that
has
been
deprecated
because
we'll
be
able
to
see
the
stored
configuration
for
that
resource
inside
of
Etsy
D,
we
don't
keep
any
information
about
this.
B
Other
than
this
field
right,
so,
as
you
define
these
things
say,
we
could
bring
up
a
116
cluster
and
we
define
a
resource,
and
then
we
upgrade
that
cluster
to
118
or
117
and
then
118.
When
we
look
at
the
way
those
resources
are
stored,
they're
gonna
be
stored
exactly
the
same
way,
and
it
wouldn't
matter
at
all
that
they
that
we
upgraded
through
it
and
the
only
record
we
would
have
that
that
resource
was
created
under
a
deprecated
API
group,
which
would
be
this
would
be
this
field.
B
It
doesn't
actually
keep
that
record
in
the
way
that
it's
stored,
the
garden
is
pretty
awesome.
I
have
a
great
backyard
and
I'm
really
I'm
really
grateful
to
have
it.
So
that
was
one
of
the
things
I
wanted
to
show
you
so
like.
If
you
have
wondered
how
tooling,
like
that,
is
able
to
introspect
resources
to
find
inside
your
cluster
and
understand
that
those
resources
were
created
with
a
previous
version
or
a
previous
API
version,
or
maybe
a
deprecated
API
version.
C
B
So
the
way
that
cube
kettle
plugins
work
today.
This
wasn't
the
way
that
it
always
worked.
It
used
to
work
where
you
actually,
where
we
had
a
defined
sub
directory,
where
things
would
where
all
of
the
cube
piano
plugins
would
sit
and
then,
as
long
as
you're
binary
or
whatever
the
plugin
was
was
located
in
that
directory,
then
we'd
be
able
to.
B
That
can
probably
keep
me
honest
here,
but
though,
when
this
change
happened,
we
basically
started
following
a
different
pattern
in
which,
as
long
as
the
binary
is
in
your
path-
and
it
starts
with
cube
kettle
and
then
whatever
your
plugin
is,
then
it
is
discoverable
right
so
and
that's
actually
part
that's
built
into
cube
kettles
as
it
is
today
right.
So
we
just
downloaded
a
plug-in
and
let's
take
a
look
at
that
plug-in.
That's
called
cube
kettle
honk
and
this
is
a
simple
little
bash
script
right
and
inside
of
the
bash
script.
B
C
B
Now,
although
it
won't
show
up
in
the
in
the
auto-detect
right
won't
show
up
in
the
tab,
completion
I
can
actually
do
cube,
kadal,
honk
and,
if
I
wanted
to
I
could
also
just
put
the
dash
in
as
well,
because
it's
still
in
my
path
right.
So
this
is
going
to
be
true
of
the
way
that
plugins
work
in
general
as
long
as
you've
got
as
long
as
the
the
binary
itself
is
accessible
inside
of
your
path
like
you
can
actually
discover
anything
called
cuboidal
dash
x
and
then
those
things
will
work
now.
B
One
thing
I
do
want
to
there's
two
parts:
I
want
to
like
really
make
sure
that
we
have
an
understanding
on
one.
Is
that
like,
as
this
thing
points
out
right,
any
binary
that
you're
running
would
have
access
to
anything
else
on
your
system,
it's
not
jailed
or
or
isolated.
This
isn't
running
inside
of
a
container.
It
doesn't
have
its
own
file
system
that
is
secured
from
your
existing
file
system.
These
plugins
are
actually
fully
fledged
processes
in
your
and
your
in
the
system
where
your
or
your
cube
kettle
is
running
right.
B
So
if
you
have
cube
kettle
running
here,
this
means
that
anything
accessible
to
you
from
where
cube
kettle
is
running
is
also
accessible
to
the
plugin.
This
is
important
because
it
means
caveat
emptor.
You
know
if
you're
gonna,
download
and
play
with
a
plug-in,
be
aware
that
that
plug-in
can
has
access
to
resources,
perhaps
a
little
more
than
you
might
expect
now.
I'm
totally
like
I
prefer
a
world
of
trust
right,
like
I
mean
like
I,
think
we
all
trust
each
other,
but
definitely
be
aware
of
that.
I
don't
feel
sick,
yeah
I!
B
Think
that's
fair!
You
know,
Brad
is
a
Brad.
Brad
is
a
wily
individual.
If
you
haven't
had
a
chance
to
check
out
the
recent,
the
recent
talk
that
Brad
and
Ian
did
at
our
say,
I
definitely
recommend
checking
that
out.
It's
like
with
a
part.
It
was
a
really
good
talk
and
I
really
am
very
much
enjoyed
it,
but
so
this
is
your
note
from
the
security
goofs
and
honk
as
they
say
alright.
B
C
B
This
talk
here,
I'm
just
gonna
shoot
through
it
real,
quick
and
highlight
kind
of
what
I
think
are
the
important
points,
but
definitely
check
it
out.
If
this
is
something
you're
interested
in,
but
it
basically
talks
about
a
cube,
kiddo
plug-in
that
we've
already
covered.
We
talked
about.
Basically,
this
write
an
executive
or
program
named
it
cube
kit,
all
something
put
it
in
your
path,
just
remembered.
Ok,
so
one
of
the
questions
I
had
on
this
security
topic
before
we
go
down
in
the
path
any
farther
right.
It's
like
obviously
built
in.
B
To
cube
kettle
our
arguments,
things
like
cube,
pillow,
config
right
so
I
can
do
cube
kettle,
config
view
and
minify,
and
this
will
show
me
the
redacted
version
of
my
cube
config.
So
you
can
share
that
and
like
show
show
how
it
configs
and
if
you
wanted
to
see
the
whole
thing,
if
you
do
flatten,
it
will
show
you
the
encoded,
certs
and
everything,
and
this
becomes
a
thing
that
you
can
export
and
use
it
and
other
places.
But
this
is
just
a
simple
kind
config.
B
B
B
B
Bret,
if
you
want
to
put
the
link
into
the
chat
or
into
the
you
know
the
notes,
that's
probably
the
way
to
go
so
yeah
exactly
so
I
think.
That's
actually
you
know
a
good
thing
security
wise,
because,
obviously
you
can
think
of
some
some
pretty
interesting
ways
that
that
might
get
interesting,
and
so
you
can't
shut
up
built-ins.
That's
a
good
thing.
B
Let's
explore
the
next
part
of
this
now.
Remember.
I.
Also,
remember
that
your
pet,
your
your
path,
is
set
by
you
right.
So,
if
I
do
echo
path,
I
can
see
that
my
entire
path
is,
he
is,
is
defined
by
me
and
also
I
have
some
included
system
paths
that
you
know
came
as
part
of
my
system
came
as
part
of
my
set
up,
but,
more
importantly,
I
actually
have
a
dot
local
bin
and
I
have
been,
and
this
means
that
I
have
control
over
what
plugins
are
being
are
discoverable
I
mean
right.
B
So
let's
go
back
to
this
piece
and
we'll
keep
going
here
so,
where
do
I
find
plugins
github
some
list
package
manager?
This
is
where
we
get
to
the
fun
part
crew.
So
crew
is
a
project
that
mr.
I
met
put
together
and
it
started
I
think
this
was
your
you're
working
from
the
beginning.
Wasn't
it,
but,
but
basically
to
give
some
tooling
that
will
allow
you
to
discover,
install
update
and
remove
plugins.
This
was
started
in
2018
and
it
was
donated
to
kubernetes
project
in
May
2019,
and
you
can
find
it
today.
B
B
B
Actually
this
is
a
different
org,
so
you
have
the
kubernetes
org
and
you
have
the
kubernetes
things
orc
and
underneath
the
Cooper,
these
things
or
that's
where
we
have
things
like
kind
things
like
mini
cubed
and
customized
other
stuff.
That
has
been
specific.
That
is
very
specific
to
kubernetes,
but
it
is
actually
still
looking
for
a
contribution
from
the
community
that
is
owned
and
I
mean
it
is,
and
it
is
contributed
directly
to
that
to
the
community
itself
right.
B
B
So
what
does
crew
it
kind
of
talks
about
what
it
is?
It
talks
about
the
user
guide?
There's
a
quick
start,
let's
play
with
the
QuickStart.
Let's
play
with
this
thing.
So
let's
go
ahead
and
get
this
started.
Let's
do
an
install
of
crew
here,
I,
don't
think
I
have
it
already
installed
good
kit,
all
crew
I?
Don't
all
right!
So,
let's
get
this
started
here
so
run
this
command
make
sure
the
kid
is
installed.
What
is
this
command?
We're
going
to
make
a
temp
directory
we're
gonna
drop
the
latest
download
and
the
Emma
file.
B
B
B
B
C
B
B
B
And
presumably
now
we
have
cute
cat
all
crew
and
we
do
looks
like
it's
I'm
able
to
call
it
individually
and
also
as
part
of
regular
cute
pillow
crew.
So,
let's
flip
back
over
to
the
terminal
where
it
makes
a
little
bit
more
sense
to
see
crew
installs
itself
as
a
plug-in.
That's
true,
it's
so
posted.
So
let's
take
a
look
inside
that
directory
that
we
saw.
B
We
kind
of
interesting
to
see
how
Bosch
completion
could
work
for
that
I'm,
not
sure
how
that
would
actually
function.
I
guess
you'd
have
to
do
it.
You'd
have
to
pick
up
this
part.
I,
don't
know
how
to
make
completion
like
a
polymorphic
like
that,
but
anyway,
so
if
we
did
cuckoo,
we
did
cube
kennel
crew
list.
B
B
B
B
We
should
be
pulling
that
stuff
down,
so
it
says
you
installed
plug
in
get
all
from
the
crew
index
plug-in
repository.
These
plugins
are
not
audited
for
security
by
the
crew.
Maintainer
x'
run
them
at
your
own
risk.
Caveat
emptor,
as
we
had
already
said,
I
love
that
that's
actually
a
part
of
the
crew
output,
though
so
that's
actually
really
solid.
I'm
gonna,
I'm
glad
you're
doing
that.
Cuz
I
think
that
it's
pretty
easy
to
get
missed
to
use
this
plugin.
B
It
gives
you
the
command
to
do
it,
and
then
it
tells
you
where
the
documentation
is.
So.
If
the
author
of
this,
who
in
this
case
is
Cornelius,
wag,
presumably
put
up
documentation,
then
we
can
go
see
that
documentation
of
the
link,
but
let's
play
with
it
a
little
bit,
because
you
know
why
not
I
trust,
Cornelius
he's
a
good
dude,
a
person,
I
should
say
I,
don't
know
if
he
to
do
it
or
not.
B
Wow,
this
really
is
get
off,
so
this
is
getting
every
resource
that
is
defined
inside
the
cluster,
whether
built-in
or
otherwise-
and
we
can
of
course
pass
some
arguments
right.
We
can
do
get
all
help
and
we
can
see
some
of
the
examples
so
they've
got
get
all
exclude
got
all
since
one
minute.
Oh
that's
interesting!
B
Nice,
that's
really
cool
okay,
so
we
can
actually
do
is
since
one
minute
we
can
see
what
has
changed
inside
of
our
cluster,
and
this
is
actually
kind
of
a
telling
thing
right.
So
we
round
a
1:18
cluster,
so
we
got
the
service
defined,
as
you
would
expect.
We
also
got
the
endpoints
defined,
which
is
currently
empty,
and
then
we
have
the
endpoint
slice
defined.
Also,
so
that's
actually
all
pretty
cool
stuff
yeah
keep
cattle
get
all
doesn't
get
all,
but
what
is
it
actually
good?
B
B
B
Obviously,
people
have
been
fighting
the
battle
that
cube
can
Oh
get
all
doesn't
do
what
it
says
on
the
tin
for
quite
some
time
upstream
and
I
would
say
that
it's
not
likely
to
be
fixed
anytime
soon,
but
as
long
as
you
have
the
ability
to
actually
code
this
in
to
a
binary
and
and
manage
that
part
of
it,
but
we
could
absolutely
do
it
from
that
perspective.
Right
like
we
could
actually
have
a
plug-in
that
gets
created
and
join
it
and
join
it
in
that
way.
B
Reason
for
that,
it's
all
as
an
API
category,
yeah!
That's
what
it
is.
Thank
you,
if
you
add
your
CRD
to
all,
it
would
work,
but
many.
Oh,
that's
a
kind
of
a
cool
hack,
but
many
Korra
API
is
don't
add
themselves
to
all
and
it
seems
like
they
are
unlikely
to
any
time
any
time
soon.
So
what
is
another
plug-in
that
y'all
are
interested
in
exploring
well,
one
of
the
other
ones.
I
was
curious
about
was
our
back
to
look
up
and
our
back
of
view.
So
let's
check
that
one
out.
B
B
B
C
B
C
B
So
this
is
a
pod,
not
a
deployment.
You
can
tell
because
there's
in
the
pod
name,
there's
no
there's
no
hash
right.
It
just
says
no!
No
to
admin,
it's
one
of
the
quickest
ways
to
tell
it's
not
the
only
way
inside
of
here
we
populated
the
node
name,
presumably
to
point
at
whatever
node
we
chose,
as
the
you
know,
kind
control
plane
so
in
the
menu
it
asked
me
which
node
I
wanted
to
target,
and
then
it
threw
me
into
a
node.
So
now,
if
I
do.
C
B
Is
a
complete
this
gives
me
a
complete
way
of
getting
in
so,
if
you're
like
in
AWS
or
GCP,
or
something
like
that
and
you
were
like
trying
to
get
access
on
a
host
without
having
to
SSH
in
this
is
one
of
the
ways
to
do
this.
I.
Do
this
in
a
different
way,
I
have
a
thing
called
KS
root,
which
does
it
almost
exactly?
This
is
pretty
clever,
so
yeah,
you
probably
see
me
talk
about
K
through
before.
B
And
then
it
gets
deleted
after
so
your
sniff
plugin
is
amazing
what
else
we
have
look
at
the
collection
of
plugin.
This
looks
to
be
a
collection
of
plugins
repo
I
know
it
that's
what
it
looks
to
me,
but
it
doesn't
seem
to
be
exposing
them
in
that
way.
So
let's
go
into
that
path
again
and
take
a
look,
so
them
keep
kid
all
know
that
min.
B
This
is
my
very
own:
node
admin
plug-in
function
and
it
basically
does
exactly
the
same
thing.
I'm
using
keep
title
run
to
do
it
and
I'm
passing
I'm,
taking
as
an
argument,
the
first
value
and
then
so
I
presume
that
you're
gonna
pass
me
the
name
of
the
node
as
the
first
argument,
and
it
will
basically
do
almost
exactly
the
same
thing
as
node
admin,
so
pretty
cool
tool,
good
for
Diagnostics
kind
of
scary,
from
a
security
perspective,
Thank
You
Rory.
Let's
take
a
look
at
our
backup
look.
B
B
B
B
So
you
can
look
up
by
user
what
access
they
have
like
what
what
like,
where
in
all
of
the
are
back
bindings
cluster
row,
bindings
or
regular
bindings,
that
particular
user
shows
up
in
the
list
the
subject
list,
and
in
that
way
you
could
actually
identify
what
a
user
by
that
subject
name
has
access
to
the
clip
has
access
to
within
the
cluster,
at
least
from
the
perspective
of
of
how
our
back
is
configured.
It's
kind
of
an
interesting
one.
I
really
dig
this
actually
and
here's.
Why?
Alright?
B
B
B
B
This
view
shows
me
what
this
particular
service
account
can
do
inside
of
the
inside
of
the
cluster,
with
the
permissions
that
have
been
granted
them,
but
it
does
not
describe
how
those
permissions
were
granted
them
right.
So
that
view
role,
maybe
not
me,
maybe
like
somewhat
extraneous
right,
because
obviously
I've,
given
them
I've,
given
them
an
admin
view
and
I've
also
go
and
I've.
Also
given
them
view
and
everything-
and
this
is
a
view
as
a
subset
of
admin,
but
I-
wouldn't
know
that
right,
I
wouldn't
know
that
they
they
conflict
like
that.
B
So
that's
pretty
wild.
How
are
you
doing
mr.
Joe
good
to
see
you
I
have
an
idea
of
pulling
Joe
into
the
session
if
you're
down,
but
but
yeah
we're
just
kind
of
exploring
the
different
plugins
thing.
What's
out
there
and
kind
of
take
it
easy.
This
episode
we're
looking
at
our
back
lookup,
which
is
a
really
cool
tool
for
showing
us
the.
B
Yeah
we've
looked
at
some
of
the
undercover
stuff
exactly
so
the
first
one
that
we
showed
actually
was
a
was
a
bash
one,
which
is
neat
because
it
indicates
that,
like
obviously
it
doesn't
matter
what
language
you
want
to
write
your
your
plugin
in
it
can
be
bash.
It
could
be
Python,
it
could
be
well
Python.
It
may
be
a
little
trick
here,
but
like
it
can
be
bash
as
long
as
it's
portable
right.
So
you
have
a
bash
crafter.
B
You
have
a
go
binary
what-have-you
as
long
as
it's
an
executive
all
and
you
can
kind
of
assume
that
it
will
work
and
most
of
these
in
most
of
the
target
environments,
then
you're
in
good
shape,
shout
out
to
go
for
being
really
awesome
about
that.
But
yeah
like
I.
Imagine
the
cube,
cuddle
plug-in
game.
Isn't
super
strong
in
Windows
Joe.
Would
you
be?
Would
you
be
down
to
join
a
a
zoom
and
Sheriff's
and
big
into
one
with
me.
B
B
C
A
B
B
B
B
So
I
think
that
we're
just
gonna
explore
this
next
week.
I
thought
it
would
be
worth
to
try,
but
sadly
doesn't
look
like
it
is
all
right.
Can
you
all
hear
me
better
now
and
we're
back
no
feedback
that
was
completely
nuts
and
I
apologize
sincerely
to
everyone
who
was
wearing
headphones
when
that
happened,
FaceTime
when
you're
home
and
hold
up
your
phone
to
the
webcam,
it's
possibility.
B
B
B
C
B
A
B
I
think
I
need
a
better
speaker,
because,
unfortunately,
it
looks
like
my
month.
My
laptop
speakers
just
not
doing
a
terribly
good
job
of
getting
sound
out
and
so
I'd
have
to
loop.
The
I'd
have
to
take
my
input
off
and
make
it
so.
The
input
was
coming
from
the
output
of
the
zoom,
but
there
is
Joe
at
least
he's
here
with
us
saying
hello
and
we've
learned
some
interesting
stuff
on
a
Mac.
You
can
create
a
virtual
yeah
exactly
that
would
be
the
next
part
I
need
to
explore.
B
B
B
Just
now,
with
the
whole
feedback
thing,
when
I
was
when
I
did
earlier
to
try
and
make
this
loop
is
in
in
pulseaudio,
you
have
a
loopback
device
and
you
can
actually
make
it.
So
the
two
inputs
go
to
the
same
output,
but
I
think
what
I
need
to
do
is
create
a
virtual
device
and
have
my
input
and
and
my
application
input
go
to
that
same
virtual
device
and
then
have
OBS
take
that
device,
which
is
not
something
I'm,
I'm
gonna
explore,
live
I,
don't
think
so.
B
B
Let's
see
what
y'all
are
looking
at,
you
are
looking
at
my
screen:
cool,
alright,
so
I'm
it
capturing
desktop
ID
with
euopean
yeah.
It
can
be
tricky.
The
screen
is
blurry.
The
screen
is
blurry
rather
because,
probably
because
I
had
just
reconnected
to
the
stream,
and
so,
if
you
look
at
the
bottom
right
hand
corner
down
I
guess
it
would
probably
be
down
there
down
there.
One
of
the
two
there'll
be
a
little
gear
icon
and
you
can
select
that
gear
icon
and
pick
a
better
resolution.
But
anytime
I
stop
my
stream
and
reconnect.
B
B
Interesting
issues,
custom
plug-in
indexes.
This
was
an
intern.
This
is
an
interesting
issue
from
I
met,
Oh
actually
against
crew.
So
that's
a
good
one.
This
is
talking
about
what,
if
you
could
actually
add
different
different
indexes
like
right
now,
I
think
you
can
use
crew
to
search
info
upgrade
and
list
and
that
sort
of
stuff
but
I,
guess
what?
B
If
you
would
want
to
what
if
you
wanted
to
host
your
own
index,
this
is
where
I
would
love
to
be
able
to
have
like
comet,
be
a
part
of
it
too
right
or
like
any
of
y'all.
It
would
be
great
till
I
kind
of
kind
of
expand,
our
our
ability
to
kind
of
bring
more
people
in
and
kind
of
explore
some
of
this
stuff
live.
B
B
Interesting
introduction
to
plug-in
development
if
you're
looking
at
some
interesting
ones,
consider
the
CLI
runtime
look
at
the
sample,
see
like
let's,
actually,
let's
go
play
with
that
that'll
be
fun,
so
this
is
actually
available
online.
It's
called
the
sample
seal
I
plug
in
and
it
looks
like
what
they're
gonna
do.
This
repository
implements
a
single
cube
killer,
plug-in
for
switching
the
namespace
that
the
current
cube,
config
context,
points
to.
B
B
Because
it
looks
like
it's
actually
gonna
hit,
that
I
mean
so
if
anything,
OpenShift
is
a
superset
of
kubernetes
itself.
So
what
this
is
exposing
is
actually
the
relationship
between
things
inside
of
leveraging,
our
back
inside
of
kubernetes
itself,
and
so
because
open
ship
represents
a
superset
of
that,
like
all
of
the
included
stuff
and
some
other
things
that
are
open
ship
specific.
That
I
would
expect
you
would
at
least
get
useful
output
from
those
things
that
are
known
about
in
our
back
so
interesting
question.
Once
we
get
more
discoverable
keep
kettle
plugins
new
whisks.
B
But
it's
pointing
out
that,
like
any
of
these
plugins
have
access
to
like
everything
that
you
have
access
to,
and
so
you
have
to
kind
of
be
aware
a
little
bit
more
careful
about,
like
you
know
what
plugins
you're
operating
and
it's
actually
even
part
of
the
output
of
crew,
which
again
shout
out
to
minute
for
like
I'm,
making
sure
that
I
mean
for
putting
that
in
there.
That's
awesome.
B
B
But
this
is
a
plugin
that
we
ship
as
part
of
that
which
allows
us
to
authenticate
to
the
cluster
using
vSphere
primitives,
and
so
you
can
kind
of
authenticate
as
a
regular
vSphere
user,
and
this
in
this
ties
in
that
piece
of
it,
which
is
actually
pretty
cool.
But
it's
listing
all
the
plugins
that
are
known
right.
B
So
here's
the
links
to
plugins
that
have
been
installed
just
recently
and
it
actually
throws
a
warning
which
is
not
completely
accurate,
but
it
says
cue,
petal,
config,
overrides
existing
command,
cube,
petal,
config,
right,
I
would
say:
I,
don't
think
that
this
overrides
it
over
writes
it,
but
it
does
conflict
with
it.
So
perhaps
like
conflict
would
be
a
better
term,
but
regardless
that's
a
good
point.
Your
utilization,
you
know
somebody
was
actually
just
pointing
this,
putting
out
the
need
for
utilization
resources
and
that's
actually
kind
of
a
good
point.
B
B
So
this
is
actually
probably
not
as
interesting
as
I
was
hoping.
It
would
be
what
I
was
trying
to
figure
out.
It's
like
one
of
the
questions
I
had
for
plugins,
and
this
is
actually
something
that
might
be
interesting
to
hear
back
shattering
commands
as
a
feature
and
a
danger,
but
it
doesn't
actually
allow
you
to
shatter
them.
In
fact,
the
precedence
is
going
to
always
be
the
built
in
not
not
the
shadowed
bone
sandbox
to
some
degree.
That
would
be
cool.
I.
B
Think
that
the
way
to
do
that
would
be
to
actually
run
it
as
I
mean
in
some
form
of
container,
whether
you
run
it
as
a
docker
container
or
or
some
way
like
that.
So
one
of
the
questions
I
had
when
thinking
about
plugins
was
like.
How
can
you
extend
the
authentication
mechanism
that
cube
kettle
has
into
the
plug-in,
but
then
I
was
realizing.
That
I
was
thinking
about
this
wrong
and
that,
fundamentally,
the
permission
happens
before
the
plug-in
starts.
Is
that
right
omit
or
do
that?
But
do
we
have
any
feedback
on
that?
B
One
I'm
curious
how
that
would
work.
It's
like
in
this
example.
What's
happening
is
that
when
you
call
the
function
that
it's
a
part
of
this
plug-in
list
nodes
or
cube,
channel
node
admin,
then
this
function
is
actually
going
to
make
use
of
the
cube
channel
command
directly.
So
it's
kind
of
looping
itself
back
over
itself.
Right.
B
C
C
B
B
No
go
get
it
ass
see.
This
is
actually
something
I've
run
into
a
few
times
and
so
I'll
take
a
second
to
call
this
out.
Obviously
this
is
a
git
repository.
It
may
be
intuitive
to
you
to
think
that
you
could
just
do
go,
get
and
then
call
up
the
path
right.
Okay,
it's
not,
or
you
would
be
able
to
come,
you
go
get
github
Cooper
to
ditch
sample,
see,
lie
plug-in,
but
that
won't
work,
and
it's
actually
very
smart
of
them
to
call
this
out
here
down
below.
B
Like
a
note,
if
you're
gonna
vendor
or
go
get
this
package,
you're
gonna
want
to
get
it
from
this
particular
path.
This
is
actually
part
of
kind
of
like
the
vanity
URL
stuff
that
that
go
can
actually
support.
So,
if
you're
interested
in
that
kind
of
an
interesting
topic,
so
let's
go
ahead
and
kick
this
up.
Let's
play
with
this
sample,
plugin.
C
B
So
part
of
that
change
is
probably
like
a
facing
change
on
the
on
behalf
of
like
the
way
that
Cora
sed
is
now
part
of
sed
IO,
its
own
kind
of
its
own
github,
org,
and
so
part
of
that
means
that
it's
probably
just
recently
made
a
change.
It's
no
longer
available
at
the
old
repo,
but
it's
only
available
at
the
new
repo
and
one
of
my
dependencies
in
inherits
it
differently
than
the
way
that's
currently
described.
B
So
if
we
were
to
go
if
we're
to
go,
look
at
the
way
that
Etsy
the
repository
is
set
up
inside
of
github.
It
likely
requires
that
you
import
or
import
it
using
that
new
name
rather
than
the
old
name,
which
makes
me
also
wonder
if
we're
gonna
get
beat
up
on
that
in
the
in
the
kubernetes
repo,
because
there's
obviously
there's
a
in
the
API
server
we
actually
vendor
in
a
CD
as
well.
So
it's
actually
kind
of
an
interesting,
interesting
change.
I
wasn't
expecting
to
see.
B
B
C
C
B
Alright
we're
getting
somewhere,
let's
take
a
look
at
how
this
thing
works,
so
this
thing
is
pulling
in
PFLAG
it's
pulling
in
the
generic
seal,
I
opt
options
and
the
sample
C
and
I
plug
in
from
the
KS
repo.
Basically,
it's
kind
of
an
example:
implementation
of
itself.
It
sets
flags
for
a
new
flag
set
Cuban
s
I'm
pulling
in
the.
B
B
B
Recently,
rewrote
it
and
go
so
if
you're
interested
in
trying
out
the
go
rewrite
I,
don't
think
this
is
actually
part
of
like
the
way
that
cube
CTX
is
exposed
today
like
automatically
through
its
saw.
But
if
you
wanted
to
try
this
out
and
play
with
it,
I'm
actually
running
this
one
myself,
but
now
it
used
to
be
bash,
and
now
it
is
go
so
definitely
check
it
out
and
get
feedback
on
that.
Another
thing
that
ahmet
has
written
that
I
have
found
incredibly
helpful.
B
It's
this
network
policies,
repo
and
I,
don't
mean
to
say
that,
like
it's,
only
I'm
not
doing
this
work,
I
feel
I'm
sure
that
he
would
catch
me
out
on
that.
It
is
a
bunch
of
people
doing
it,
but
like
incredible
incredible
amount
of
like
really
incredibly
relevant
things
that
I've
seen
now
come
from
Ahmet
over
the
years.
So
this
kubernetes
network
policies,
repo,
is
also
tremendous
like
if
you're
trying
to
understand
network
policy
inside
of
kubernetes
I
highly
recommend
checking
it
out.
B
It's
been
tremendous,
so
I
mean
it's
great
to
see
like
all
the
work
from
and
the
community
surrounding
him.
You
know
like
it's:
tremendous
I've
opened
two
issues
on
a
number
of
different
repos
and
he's
always
been
very
responsive
and
I'm,
just
great
so
shout
out
to
I
meant
for
all
the
work,
including
on
on
crew,
and
there
were
policy,
QC
checks
and
all
the
things
that
are
not
top
of
my
head,
which
includes
a
ton
of
things
all
of
your
speaking
engagements.
That's
really
I'm
just
grateful
to
have
you
around.
B
B
B
B
I'm
very
much
looking
to
seeing
that
article
Steve
I'd
love
to
see
how
metal
does
sed.
It's
actually
kind
of
a
follow
on
of
the
chat
that
we
did
before
on
entity
and
I
do
plan
on
doing
a
second
one
of
those
so
stay
tuned
for
that
coming
soon
and
let's
take
a
look
at
the
gamma,
plugin
manifests
I'm
curious
your
criticism
around
that
the
again
will
plug-in
manifest.
Oh
I.
B
B
Yeah
other
plugins
probably
have
more
complex
ones.
Okay,
so
inside
of
your
dot
crew
directory,
you
have
for
direct,
you
have
four
sections,
and
this
is
actually
I
guess
how
crew
is
working
and
we've
looked
it
a
little
bit.
Look
at
this
a
little
bit
before
right.
So
if
we
go
into
the
bin
directory,
you
can
see
these
symlinks.
That
crew
manages
on
your
behalf
and
that
way
the
binary
that
it's
pulling
down
when
you
do
get
all,
for
example,
it
doesn't
have
to
be
that
this
doesn't
have
to
be
the
name
of
the
binary.
B
C
B
B
So
this
is
where
the
index
is
coming
from.
So
all
of
the
plugins
that
are
known
about
are
coming
from
here
and
I
believe
that
it's
covered
in
crew
that
you
can
actually
like
when
you
want
to
actually
register
a
plugin.
So
you
write
you
write
one
and
you're
interested
in
putting
one
up
how
you
go
about
submitting
plug-ins
to
crew.
If
you
want
it
to
actually
add
it
to
the
crew
index,
it's
a
pull
request
against
that.
B
B
B
B
This
is
a
it's
a
version
to
manifest,
which
is
interesting,
I
wonder
how
many
of
these
are
out
of,
like
our
you
know,
what's
in
the
same
particular
version
like
if
everybody's
using
to
be
one
fo,
two
or
if
they're,
multiple,
but
then
we
have
the
different
versions
of
the
resource
that
have
been
defined
right,
so
here's
the
access
matrix,
one
that
we
pulled
in
I,
think
if
we
don't
specify
it
just
grabs
the
most
recent
and
then
it's
also
tying
in
match
labels
OS.
So
whether
it's
an
the
OS
is
Darwin,
Windows
or
Mac.
C
B
B
B
B
Then
we
pop
into
the
receipts
directory
and
we
populate
and
we
see
what's
there
and
then
we
read
the
plug-in
and
then
we
see
that
it's
not
there
and
then
we
start
about
installing
it
created
a
temp
directory.
It
pulled
that
hunt
the
the
tar
ball
into
that
temp
directory.
This
is
very
well
annotated,
great
stuff.
The
log
is
tremendously
useful,
compares
to
checksum
with
the
signed
version
of
that
piece.
B
B
Yes,
okay,
so
as
part
of
processing
this
in
our
manifest,
we
specified
where
you
go
in
our
manifest.
We've
specified
that
we
wanted
to
process
the
license
file.
We
wanted
to
process
cube,
C
SSH
file
and
we
wanted
to
process
the
parse
arts
file.
The
tarball
might
own
might
contain
significantly
more
than
that,
but
these
are
the
three
files
that
we
want
to
have
stored
inside
of
that
path,
and
so
these
are
the
three
things
that
we're
just
exposed
directly
from
that
tar
ball.
B
B
Nice,
okay,
well,
obviously
I'm
using
a
kind
closer
which
doesn't
have
isn't
listening
on
SSH.
So
this
is
not
gonna,
be
useful
to
me,
but
very
cool
stuff
yeah.
Then
it's
really
neat
all
right.
Well,
that's
all
I'm
gonna
time
a
lot
of
time
to
explore
today
so
I
wanted
to
say.
Thank
you
very
much
for
being
here.
I,
look
forward
and
know
that
it's
important
to
us
and
to
to
kind
of
exploring
more
of
like
a
way
of
collaborating
on
some
of
these
things.
B
It
would
be
great
to
see
like
Joe
and
I
get
in
there
and
like
explore
some
project
it'd,
be
great
to
see
members
of
the
community
in
here
kind
of
exploring
some
stuff
with
us
like
for
this
one
with
Ahmet
I
would
love
to
actually
have
on
it
like
here
with
us,
exploring
this
stuff,
and
so
thank
you.
So
so
much
for
all
of
your
engagement
have
a
great
weekend
and
I'll
see
you
all
next
time.
So,
thanks
again,.