►
From YouTube: GitOps Guide to the Galaxy (Ep 30) | GitOps From Scratch
Description
How do you deploy Argo CD in a way that conforms to the OpenGitOps principles? With the excitement around Argo CD and GitOps, many people rush to implement Argo CD in production. Unfortunately, this can lead to a lot of trial and error when it comes to directory structures. In this episode, we'll look at how you can set up GitOps repos from scratch, which can help deploy Argo CD in a way that adheres to OpenGitOps principles. If you're just getting started with GitOps, this is a great episode to help make you comfortable and follow the OpenGitOps principles, so don’t miss it.
Follow us:
Christian Hernandez https://twitter.com/christianh814
What is GitOps Guide to the Galaxy?
Every other Thursday at 3pm ET hosts Christian Hernandez and Chris Short sit down to discuss everything in the GitOps universe, from end-to-end CICD pipelines to creating Git workflows.
Subscribe to Red Hat's YouTube channel: https://www.youtube.com/redhat/?sub_confirmation=1
About OpenShift:
Red Hat OpenShift is an open source container application platform based on the Kubernetes container orchestrator for enterprise application development and deployment.
Openshift.com https://red.ht/33Fh6uV
OpenShift on github https://github.com/openshift/
OpenShift on redhat.com https://red.ht/2o4ccDk
#RedHat #GitOps #Kubernetes
A
A
A
All
right
and
we
are
live.
We
are
live
cool,
so
welcome
everyone
to
get
ops
guide
to
the
galaxy.
I
will
be
your
captain
in
this
journey.
Christian
hernandez,
technical
marketing
here
at
red
hat
again,
co-captain.
Is
you
this
time
we're
back
from
an
extended
break,
and
not
only
are
we
back
from
an
extended
break?
We,
we
came
out
a
little
a
little
an
hour
earlier.
A
A
We
have
a
big
topic
today,
we're
talking
about
originally
when,
when
I
decided
to
do
this,
this
stream,
that
this
topic,
I
was
thinking
about,
doing
it
in
terms
of
argo
cd
and
how
to
do
your
directory
structure
for
argo
cd
in
a
way
that's
getups
compliant,
and
I
figured
you
know
what
how
about.
I
just
show
you,
instead
of
just
talking
about
it
right
and
so
that's
why
I
kind
of
just
you
know,
re
renamed.
A
This
show
or
kind
of
re-thought
this
topic
and
say
how
many
I'm
going
to
do
a
directory
structure
from
scratch.
So
I'm
going
to
get
right
to
it.
I
do
have
a
heart
stop
today,
unfortunately,
so
this
will
be
a
shorter
show,
if
I'm
going
too
fast
that
that
is
why,
if
I
don't
get
to
your
questions,
that's
that's
that's
why,
but
and
again
forward
apologizing.
I'm
gonna
go
really
really
fast.
A
So
for
those
of
you
who
are
new
to
not
only
to
the
show,
but
also
to
get
ops,
this
is
gonna.
This
might
go
over
your
head
a
little
bit
so
just
kind
of
a
fair
warning.
It's
gonna
be
a
lot.
Don't
worry!
I
do
have
resources
that
you
guys
can
take
a
look
at
afterwards.
So
let
me
start
here
by
sharing
my
screen
share
my
screen,
which
one
this
guy
here
oops.
A
A
So,
right
before
I
I
jump
in
a
little
little
things.
I
I
want
you
guys
to
realize
that
a
one
git
ops
con
is
is
coming,
we're
doing
it
again.
We
loved
it
so
much
we're
doing
adopticon
a
second
time
again
in
person.
This
will
be
a
in-person
event
in
a
co-located
with
kubecon
in
valencia,
spain,
so
get
up.
Con
is
happening
in
17
weeks.
Look
looks
like
there's
a
there's,
a
countdown
here
saying
you
know
it's
17
weeks
away,
which
just
seems
crazy.
A
A
We
had
to
make
difficult
decisions
last
time
for
get
ops
con
about
what
we
will
accept,
what
we
won't
accept,
but
this
time
we're
casting
a
wider
net,
because
cncf
gave
us
two
rooms,
that's
how
popular
it
was
here
in
north
america.
So
we
hopefully
will
we
we
recapture
that
right,
hopefully
we'll
make
it
even
bigger.
We
we
hope
it'll
be
better
because
we
got
two
rooms
so
get
up
con.
There's
the
link.
I
dropped
it
in
the
chat.
A
Their
cfp
is
open,
so
feel
free
to
submit
as
much
as
you
want.
On
top
of
that,
another
event
we're
having
in
february
so
february.
9Th
we
are
having
red
hat.
Is
sponsoring
openshift
commons?
So
for
those
of
you
who
don't
know,
bishop
commons
is
where
the
the
community
of
users
of
openshift
and
okd
get
together
it's
kind
of
an
upstream
community
that
is
sponsored
by
red
hat.
A
We
are
having
a
get
ups
day
right,
so
get
ups
commons
gathering
on
openshift
commons
gathering
on
get
ops
so
and
in
there
you
can
see
we're
gonna.
Have
you
know
our
friends,
cornelia
and
dan
right
from
amazon
and
quote
fresh
they've,
been
on
the
show
before
they're
they're
they're
coming
back
members
of
the
openshift
get
ops.
Sorry
of
the
open
get
ops
community
will
be
there
as
well.
It's
gonna
be
a
on
february
9th.
A
So
it's
it's
open
virtual
right,
so
it's
virtual
got
to
say
that
right,
unlike
githubscon,
the
this
gathering
is
virtual.
So
for
wherever
you
are
it's
it's
free
to
attend.
It's
free
online,
get
to
know
the
community
I'll,
be
there
right,
I'll,
be
I'll,
be
kind
of
doing
a
sub.
An
expansion
of
what
I'm
doing
today
is
how
to
get
started
with
get
ups
so
be
sure
to
check
that
out,
moving
right
along
at
10
000
miles
per
hour
pace.
A
So
the
idea
here
is,
I
I
get
a
lot
of
questions.
Not
only
do
I
get
a
lot
of
questions,
but
also
our
here
at
red
hat
are
solutions,
architects
and
consultants,
and
even
in
the
upstream
right,
even
in
the
openshift,
get
up
sorry
openshift
get
out.
There's
too
many
words
with
open,
open,
github's
community
right
the
opens
get
up
working
group.
We
get
this
question
a
lot.
Is
the
question:
is
okay,
I'm
buying
into
get
ops?
A
You
know
I
played
around
with
argo
city.
I
played
on
flux
I
played
around
with
acm.
How
do
I?
How
should
I
structure
my
repo
right?
And
so
you
know,
there's
there's
a
lot
of
opinions
out
there
there's
a
lot
of
kind
of
best
practices,
and
I
figured
that
I
would
do
a
show
on
like
how
do
I
create
a
repo
when
you
know
when
when
doing
getups
right
and
so
the
idea
is
here
is
let
me
drop
down
to
my
cli
here.
A
If
I
do
an
oc
get
nodes
right,
I
have
a
kubernetes
cluster
in
this
case,
openshift
right.
I
just
installed
this
day,
one
right.
So
if
I
do
oc
get
pods
a
argo,
there's
no
argo
or
any
any
getups
controller
on
there.
So
the
idea
being
that
you
know
I
have
a
cluster.
How
does
how
does
christian
get
get
started
and
I
figure
that
I
will
kind
of
give
my
best
practices
slash?
How
do
I
do
things
here
in
in
a
way
that
is
repeatable
right?
A
So
at
any
time
I
do
something
I
like
you
know
you
as
your
devops
practitioners.
You
us
get
ups
folk,
you
want
to
do
things,
that's
repeatable
right
and
so
that's
kind
of
like
the
idea
I
have
also
I
want
to
do.
The
idea
of
that
is
agnostic
to
the
gitops
controller
right,
so
here
here
at
red
hat,
we
we're
all
in
on
argo
cd.
A
You
know,
openshift
git
ops
is
based
on
argo
cd,
but
people
like
using
flux,
there's
tons
of
flux
users
out
there
flux
is
a
certified
operator
on
openshift.
So
I
want
to
you
know
I
don't
want
to
like
box
myself
into
any
particular
git
ops
controller.
So
this
is
how
I
get
started.
This
is
specific
to
argo
city.
Openshift
get
ups,
but
it's
essentially
you
can
do
it,
modify
this
to
use
it
for
flux
as
well.
A
I've
done
with
flux
the
work
I've
done
with
argo
autopilot
and
I
kind
of
just
mash
it
all
together
to
make
it
work
for
me
specifically,
one
one
caveat
is
this:
this:
this
design
is
specific
to
one
cluster,
meaning
that
this
is
a
one-to-one
relationship.
I
have
one
repo,
that's
pointing
to
one
cluster
right.
A
One
repo
is
managing
one
cluster,
the
design
changes
as
you
get
into
like
to
say,
mono
repo,
where
you're
doing
one
repo
for
many
clusters
or,
if
you're,
doing
like
poly
repo
sort
of
design,
where
it's
many
repos
to
many
clusters
and
you're
mix
and
matching.
So
I
decided
to
start
simple
a
one-to-one.
I
have
one
cluster.
A
How
do
I
do
my
git
ops
directory
structure
here
so
first,
things.
First,
as
always,
create
a
new
repository
right,
create
a
new
repo,
we're
doing
this
from
scratch,
right
so
with
with
warts
and
everything
with
mistakes,
and
all
this
is
how
I
do
it
right,
and
so
here
we'll
call
this
see
here
something
fun.
A
A
One
of
the
one
of
the
things
that
that
I
also
do
at
red
hat
is
that
I'm
also
I'm
not
only
the
get
ops
tmn,
but
I'm
also
the
windows
containers
tmm,
two
of
the
weirdest
things
unrelated,
but
you
know
so
that's
why
you
saw
me
do
dur,
because
I
just
spent
some
time
on
a
windows
server.
So
right,
it's
just
the
readme,
where
it's
square
one
great
things
right
so
when
where
or
how
to
get
started
is
or
how
to
initiate
something
onto
a
cluster.
A
How
to
get
something
onto
the
cluster
is
what's
something
that's
commonly
referred
to
as
a
bootstrap
right,
so
bootstrapping.
So
this
is,
which
is
why
I
always
create
a
directory
called
bootstrap
right.
So
I
have
a
directory
called
bootstrap
and
that's
kind
of
like
where
I'll
put
all
my
things
where
I,
how
I
bootstrap
a
cluster
or
how
I'm
gonna
bootstrap
this
cluster
is
that
this
is
where
the
all
the
artifacts
go
in
there.
So
in
the
bootstrap
directory
I
use
customize.
A
So
for
those
of
you
who
haven't
aren't
familiar
with
customize-
or
this
is
your
first
time
I'll,
put
a
link
here
in
the
chat
I
can
find
it
there
we
go
so
if
you've
missed
past
episodes
feel
free
to
check
that
out.
I
talk
about
customize
a
lot
and
the
reason
I
use
customize
is
I
do
I
use
the
dry
method,
don't
repeat
yourself,
chris
short,
which
who
was
an
original
host
here
for
get
ops
guy
to
the
galaxy
who's
since
left
red
hat?
A
Well,
we
he
actually
joked
he,
because
we
we
still
keep
in
touch.
He
actually
joked.
You
know,
don't
repeat
yourself,
it
says,
don't
repeat,
yaml,
either
way
the
the
acronym
works.
The
three-letter
acronym
works
is
don't
repeat
yourself,
so
I
have
so
I'm
gonna
use
customize
for
that
so
customize,
so
I'm
gonna
have
a
base
directory
here
and
I'm
gonna
have
overlays
as
well.
So,
let's
back
up
here,
let's
do
a
tree,
oh
warts,
and
all
I
messed
up.
I
actually
like
to
create
a
directory
above
this
directory.
A
So
let's
do
let's
do
that.
So
what
I
like
to
do
is
I
like
to
create
a
directory,
and
I
and
I
name
this
directory
the
same
as
the
cluster
name,
so,
for
example,
if
I
do
an
oc
get
machine
sets
right.
This
is
the
cluster
name
of
this
cluster
that
I'm,
that
I'm
doing
so
in
order
to
get
that,
I
just
grab
that
cluster
name,
which
is
that
guy
here
and
I
create
a
directory
there.
A
A
Let
me
do
touch,
even
though
I
can
just
vi
it
base
here
I
do
inside
of
the
bootstrap
base
directory.
I
do
everything
that
I
put
all
the
yamls
that
it
takes
to
install
the
get
ops
controller.
In
this
case
it's
going
to
be
openshift
get
ops,
but
this
doesn't
have
to
be
openshift
get
ups.
It
can
be.
You
know,
argo
cd
could
be
flux
whatever,
but
in
in
our
in
this
specific
use
case.
A
This
is
where
I
put
in
how
to
install
the
the
get
ops
controller
and
in
this
case,
we're
using
olm
right
so
we're
using
an
operator.
So,
in
our
case,
that
means
that
we
just
have
to
subscribe
right.
We
have
to
add
a
subscription
so
clear
out
of
this
once
I
do
a
tree
again
right
so
in
base.
I
have
this
file,
so
let's
let's
modify
this
file
right
and
then
I
want
to
go
fast.
I
have
this.
A
I
was
here
off
to
the
side,
but
right
I
just
put
in
you
know-
and
you
can
find
this
in
the
official
openshift
documentation.
So
this
is
how
I
get
the
operator
installed
right
now,
openshift
get
off
to
operator
install
so
again,
I'm
gonna
do
this
a
lot
tree
right.
So
here
I
got
my
base
in
there,
and
this
is
how
this
is
the
installation
manifest
for
for
openshift
get
offs.
This
can
have
as
many
files
as
needed
right
for
argo
cd.
A
It
can
be,
you
know
a
few
dozen
files
or
you
can
have
one
file
with
all
your
yamls
in
them
right,
but
since
I'm
using
customize,
I
have
to
create
a
customization
file
so
that
I'm
gonna
go
over
under
bootstrap
overlays.
Actually,
I
have
to
create
a
directory
there.
First
and
I'll.
Tell
you
why,
in
a
second
overlay's
default
right.
A
Quick
recap:
right:
overlays
is
basically
where
you
start
your
deployment
from,
and
it
reads
from
base
and
you'll
see
how
this
works
in
a
little
bit.
If
you're
not
used
to
customize,
don't
worry.
This
will
make
sense
as
you
go
along
so
bootstrap
overlays
default,
and
then
here
you
have
to
have
a
customization
file.
A
A
A
Let
me
do
a
tree
and
I'll
talk
to
the
tree,
walking
to
trees.
We're
talking
in
nature
here
so
inside
overlays
inside
the
default
is
specific
configuration
for
the
get
ops
controller
for
this
cluster.
So
while
this
is
like
a
generic
bootstrap,
the
cluster
itself,
its
specific
configurations
for
the
get
ops
controller
will
go
into
default.
So
this
takes
two
things
so
one
I
do
the
let
me
touch
the
installation
file
hit
touch
cluster
bootstrap
overlays
default,
I'm
touching
the
openshift
get
ops.
A
A
A
A
I
always
said
some
people
disagree
with
me.
I
disagree
with
that
disagreement,
but
I
always
I'll
give
cluster
admin
to
argo
cd,
because
if
I'm
doing
get
ops,
everything
is
driven
via
git
workflows,
so
it
doesn't
matter
if
this
controller
has
cluster
admin
it
needs
like.
How
can
you
deploy
something
onto
a
cluster
if
it
doesn't
have
access
to
deploy?
So
disagree,
that's
cool
put
in
the
chat.
If
you
disagree,
I'm
I'm
happy
to
disagree
with.
A
A
I
don't
know
if
it's
like
1k
or
2..
It
looks
like
it's
just
1k,
okay
and
then
here
a
few
things
right.
So
first,
let's
put
the
header
the
gvk,
because
that's
always
important,
then
resources.
So
the
resources
just
go
back
resources,
basically
the
the
name
of
the
files
right.
These
two
files
that
I
want
to
include,
which
I
do
have
copied
over
here.
A
A
And
then
there
we
go
so
here
is
actually
this
is,
I
would
say,
if,
if
you
don't
leave,
if
you
with
anything
else
except
for
this,
this
is
actually
a
pretty
good
starting
point.
This
is
actually
viable
and
you
can
probably
just
start
here.
So
if
I
do
a
customize
a
build,
a
cluster
bootstrap
overlays
default,
this
will
basically
apply
everything
that
we
created.
So
we
created
the
subscription
which
you
see
is
down
here,
which
is
there.
A
We
created
the
our
back
policy,
which
is
here
and
then
the
argo
city
instance,
which
is
there,
and
so
this
is
actually
enough
to
get
started.
But
I
like
to
add
a
few
things
on
top
of
this
right,
and
this
is
kind
of
like
where
the
genesis
of
like
okay
cool
this
is
cool
to
get
started
but
like
where
you
know
how
do
how
do
the?
How
do
the
experts
do
it
right,
and
so
let's
do
a
quick
tree.
A
A
Very
right,
it's
just
right
now.
I
just
have
bootstrap
on
the
top
on
the
top
level
directory,
so
I'm
going
to
create
something
else,
and
if
gerald,
if
you're
watching
this
here,
you'll
recognize
it
you'll
recognize
this
create
directory
cluster
and
I'm
going
to
create
a
directory
called
components
right
and
so
in
the
components
directory.
If
I
can
find
my
notes
here,
yeah
so
in
components,
components
means
different
things
to
different
people
right
and
to
probably
to
gerald.
It
probably
means
something
different
to
him
than
to
me.
A
To
me,
components
are
things
that
you
need
in
order
for
the
git
ops
controller,
to
work
or
things
that
you
do
in
order
for
the
get
ops
controller
to
to
function.
The
way
you
want
it
to
function
right,
so
that's
it's
so
for
me.
I
like
using
application
sets.
So
where
do
I
store
my
application?
Sets?
I
store
them
in
components.
A
A
I
do
other
things
that
could
go
in
here.
Are
things
like
argo
cd
projects
right?
So,
as
you
know,
our
city
app
projects
are
independent
of
openshift
projects
there
so
that
the
way
that
looks
like
it's
like
this
here,
so
that's
where
those
individual
components
go,
and
so,
for
instance,
here
let's
have
an
example
of
let's
oops.
A
A
A
This
was
a
this
is
a
workshop
in
in
or
or
training
or
a
class
or
something
I
would.
This
is
where
everyone
out
loud
says
customization.yaml
file
right.
So
this
is,
you
need
a
customization.yaml
file
because
we're
using
customize
all
the
way
through
an
episode.
I
called
customization
files
rat
droppings
because
because
that's
kind
of
how
it
is
in
your
repo
you're
just
going
to
see
this
all
over
your
repo
get
used
to
it.
It's
it's
fun.
A
A
A
A
A
So,
as
you
can
see,
the
more
and
more
we
go
through
the
more
and
more
you're
gonna
see
that
a
lot
of
these
steps
are
repetitive
right
and
so
because
it's
it's
essentially
you're
gonna
create
a
directory
with
all
the
manifests
and
you're
gonna,
just
reference
it
from
something
else
and
in
front
of
something
else
right.
It's
going
to
chain
together,
there
will
be
a
point
of
entry
where
you
make
changes
and
everything
else
basically
chains
off
of
that,
and
so
that's
kind
of
how
you
have
the.
A
A
Here
I
do
have,
I
wonder
what
would
be
the
best
way
to
like
just
do
the
application
sets
and
then
show
you
why
or
do
the
other
directories
and
then
go
back
to
the
application
set?
So,
let's,
let's
do
let's
do
the
ladder
right,
so
let's
do
so
the
next
directory,
so
we're
gonna
come
back
to
components.
Application
sets
we're
gonna
all
right,
so
here
we
have
bootstrapping
components.
The
next
directory
I
like
to
create
is
something
that
I've
kind
of
went
back
and
forth
of
what
to
call.
A
The
reason
I
call
this
core
is
because
this
is
where
all
the
core
components
for
for
this
cluster
to
run
right.
So,
like
all
the
core
components,
all
the
cluster
configurations,
all
the
you
know
all
this.
You
know
sort
of
thing
like
idp
configurations,
right,
ldap,
configurations,
right
things
for
like
admin,
e
stuff.
A
I
put
it
in
core
and
this
I'm
actually
not
100
married
to
this
name,
but
it's
basically
there
almost
as
a
placeholder
for
for
in
case
someone
else
gets
it's
a
better
idea
right
or
I
find
something
better,
so
yeah
so
gerald
right.
Thanks
for
joining
us,
he
says:
do
you
always
use
the
actual
cluster
name
in
the
repo
name?
Do
I
always
know
so?
If
I
want
something,
that's
like
completely.
Repeatable
means
something
that's
completely
agnostic
to
the
cluster.
A
I
just
call
it
cluster
and,
and
so,
and
the
reason
for
that
is
yeah.
So
the
reason
for
that
is
for
what
you
said
in
case
the
cluster
gets
redeployed.
I
can
just
you
know.
I
don't
have
to
do
any
weird
renaming.
If
so,
yeah
you
don't
absolutely
have
to
right.
I
just
I
just
did
it
for
for
this
demo,
but
sometimes
I'll
just
call
it
cluster
right.
If
it's
like,
so
generic
that
you
know
it
gets,
you
know,
deployed
multiple
times.
I
just
call
it
cluster.
A
So,
oh
and
and
gerald
says
he
uses
config
instead
of
core,
but
yeah
he's
not
100
on
it
either.
I'm
not
either
so
so
here
inside
of
core
I'll,
create
a
directory
per
thing,
and
I
don't
know
what
else
to
call
it.
Maybe
someone
else
will
figure
it
out.
It'll
make
sense.
Once
I
do
it
I
do
here.
I
do
a
get
ops
controller.
A
A
I
just
use
get
offs
controller,
so
here
I
do
get
ups
controller
inside
of
this
instead
of
this
directory,
and
in
here
this
is
a
very,
very
simple
there's
only
one
file
in
here
and
it's
let
me
touch
it
and
it's
you
guessed
it
a
customization
file,
and
so
in
here
it's
it's
a
very
simple
file.
It.
Actually
it's
actually
just
one
line,
or
you
know
a
few
one:
config
actually
is
basically
saying:
hey
load
everything
in
from
default.
A
So
the
reason
why
this
is
here
it
it
it's
gonna,
seem
you're.
Gonna,
think
this
seems
kind
of
redundant
christian,
and
it
is
a
few
reasons
right.
One
is
that
this
this
is
the
application
set,
is
gonna,
deploy
it's
gonna
point
to
this
directory
and
I'll
show
you
in
a
bit
how
that
works
and
and
that's
one
and
two,
the
this
is
how
argos
or
the
get-offs
controller.
In
our
case,
argo
cd
is
gonna
manage
itself
right.
A
So
you
you
want
not
only
argo
city
to
deploy,
but
you
want
argo
cd
to
manage
itself
as
well,
and
so
with
this
conforms
to
the
the
get
ops
principle
right.
You
don't
want
your
application
or
application
sets
definitions
to
be
independent
of
outside
of
the
gitops
workflow.
So,
and
so
you
can
test
this
by
doing
customize,
you
know
build
cluster
up
core
oops
yeah
cluster
core
get
off
the
controller
right,
and
you
see
that
it'll
it'll
build
just
fine
right,
because
it's
just
again,
let
me
clear
the
screen.
A
And
it's
just
pointing
up
to
the
bootstrap
directory
right,
and
so
this
seems
I
you
know.
100
agree.
This
seems
like
overkill
right
for
one
repo
and
the
reason
I'm
doing
it.
This
way
is
just
plain
and
simply
instill
good
habits,
because
if
I'm
doing
something
like
a
mono
repo
right,
where
I
have
one
repo
and
many
clusters
I'll
actually
move
bootstrap
and
overlays
outside
of
cluster
right
and
so
it'll
actually
live
one
directory
up
and
if
we
have
time
doesn't
seem
like
we
will.
A
But
if
we
have
time
I'll
show
you
I'll
show
you
how
that
looks
like,
but
in
this
oops,
so
here
you'll
see
cluster
and
then
you'll
see
bootstrap
living
outside
of
cluster.
So
and
then
that
way,
you'll
have
you
know
those
individual
configurations
so
so
again
running
out
of
time.
So
I'm
gonna
go
super
super
fast.
So
how
do
I
load
core
into
the
cluster
right?
So
this
goes
back
to
the
application
sets
so
in
the
application
set.
A
Application
sets
and
then
core
components
right
here
I
have
an
application
set
and
this
application
set
points
to
if
you're
not
familiar
with
application
sets,
we
did
a
show
about
application.
Sets
member
watch
if
you
haven't
seen
it
so
application
set.
Is
this
configuration?
The
the
important
thing
to
realize
here?
Is
that
one
this
this
configuration
here
line
12
is
basically
saying:
hey
every
directory
you
find
in
core
deploy
that
as
an
application,
and
so
a
few
changes
need
to
be
made
here.
A
A
A
Just
like
every
other
directory
that
we
have
on
here.
What
do
we
need
class?
What
do
we
need?
We
need
a
customization
file.
I
can
paste
it
correctly
go,
get
animation
file
and
then
we,
the
easiest
thing,
is
that
we
have
core
components
on
in
this
repo.
So
that's
one
thing
cool
so
now,
just
like:
let's
do
a
tree
so
see
where
we
are
so
there
we
got
a
core.
We
got.
Application
sets
right
that
had
that,
has
the
core
components
are
listed
in
the
customization
file.
A
A
A
A
So
next
component
final
component
is
something
that
gerald
got
me
into
right.
So
this
is
another
gerald
special.
Let's
cluster
right,
I
have
core
components:
the
next
top
level.
One
will
be
something
called
gerald.
Can
you
guess
it
you
guessed
it
before
tenants
right
so
tenants?
What's
the
difference
between
tenants
and
core
core?
Is
the
core
functionality
of
the
cluster?
Tenants?
Are
your
workloads
right?
You
can
call
this
app.
You
can
call
this
workloads,
you
call
us
whatever
you
want.
A
I
call
them
tenants
because
I
have
a
cluster
and
an
app
running.
An
app
running
on
my
cluster
could
be
short-lived
long
live,
so
it's
like
a
tenant
right.
You
can
think
of
it
as
a
kind
of
apartment
right.
So
the
comp
apartment
complex
is
the
core
right
cluster
and
the
tenants
are
just
your
tenants,
100
100
percent,
I'm
not
a
fastest
enough
typer.
That's
what
gerald
says
a
hundred
percent
same
functionality
as
as
core
meaning
that,
let's,
let's
give
a
let's
give
a
good
example.
A
A
All
right,
so
I
have
my
app
there,
and
this
is
just
a
you
know.
Just
a
dumb
welcome
hello
world
app.
So
in
order
to
load
tenants,
I
do
basically
do
the
same
thing
as
I
did
with
core.
Is
I
create
an
application
set
for
it?
So
if
I
do
cluster
components,
oops
application
set
and
this
application
said,
I
think
I
called
it
tenants.
Yes,
I
did
where's
my
notes
and
then
it's
app
set
beautiful.
This
app
set
looks
exactly
the
same.
A
A
It'll
create
an
application
because
of
the
application
set.
So
since
I'm
already
loading
in
the
application
set,
oh
wait.
The
one
thing
I
need
to
do,
I'm
pointing
at
the
screen,
as
if
you
know
what
I'm
printing
at
one
thing
I
do
need
to
do
is
I
need
to
add
the
tenant
app
set
to
the
customization
right,
just
because
nothing
happens
without
customization
all
right.
So
let's
do
a
build.
A
I
don't
want
to
do
a
build
of
that.
I
wanted
oh
actually
that
would
work
so
here.
Not
only
do
I
get
an
application
set
for
the
core
configuration
which
is
somewhere
here
yeah
for
the
core
configuration,
I
also
get
it
for
tenants
as
well
right,
and
so
that
is
in
a
nutshell,
we're
essentially
done
here.
Let's
go,
let's
do
a
tree.
A
A
This
is
it's
committed
right?
Is
it
it's
in
this
in
this
repo?
So
if
you
reload
this
repo,
it's
here
right
and
then
you
can
see
the
tree
outlined
here.
Is
this
there's
no
expand
all
right
here,
for
I
don't
know
if
you
guys
use
octotree,
anyways,
so
yeah,
so
gerald's
cool
initial
commit
we're
all
there
and
the
way
you
apply.
This
is
basically,
as
some
of
us
say
that
shell
is
the
make
sure
to
let
my
kid
know.
I'm
cool
yeah.
A
A
A
So
the
reason
you
have
to
do
an
into
an
an
until
loop
is
because
you
get
errors
like
this,
where
the
operator
needs
to
install
the
crds
right,
the
custom
resource
definitions
first,
in
order
for
the
crs
to
work
right-
and
this
will
take
a
little
bit
because
olm
takes
a
while
like
if
you're
doing
like
straight
flux
or
straight
argo
cd,
it
might
go
a
little
faster
because
we
don't
have
to
wait
for
olm,
but
this
basically
can
be
done
with
shell
right.
The
bootstrapping
process
can
be
done
with
shell.
A
It
can
be
done
with
like
terraform
ansible
right.
What
pick
your
favorite
tool
personally,
I
like
using
helm
for
bootstrapping,
because
you
can
make
it
pretty
agnostic,
and
it's
just
I
just
it's
just
one
of
the
one
of
the
better
tools
for
bootstrapping.
I
would
say
right
because
that's
what
it's
used
for
so
cool,
so
we're
all
here
so
that
was
applied.
A
A
A
So
one
last
thing
before
you
know
before
I
close
it
out,
I'm
actually
kind
of
late
for
my
other
meeting,
but
I
just
kind
of
want
to
do
one
more
thing
so
here
tree.
So
this
is
enough
to
get
started
right.
So
your
your
next
question
is
okay,
so
how
do
I
add
workloads?
So
the
so
let's
say
I'm
an
admin,
and
I
want
to
add
something
in
core
right.
I
wanted
to
add
something
in.
A
A
This
actually
doesn't.
Technically
speaking,
you
don't
actually
need
this
file,
because
application
sets
will
just
just
work
the
way
it
is,
but
I
like
to
I
like
to
be
consistent,
and
so
here
so
all
I
did
here
was
I
added
a
directory
and
I
added
two
files
right.
So
if
I
do
a
git,
diff
wait
status
right,
so
I
added
this
directory
and
I
added
that
file
in
the
customization
file
so
get
add.
A
A
So
all
I
did
was
I
added
a
directory
here.
It's
gonna
take
a
while
it
might
take
a
minute
to.
I
may
kick
it
a
little
bit,
but
because
the
reconciliation
loop
is
three
minutes,
so
hopefully
we
either
caught
it
right
before
the
three
minutes.
If
we
caught
it
right
after
we
have
to
wait
a
whole
three
minutes,
but
the
idea
here
is
that
I
just
commit
my
application
and
the
reconciliation
loop
will
then
the
application
set
will
then
pick
up
that
directory
and
deploy
the
application.
A
A
A
You
get
push,
you
wait
and
we
wait
okay,
yeah
so,
and
then
it
just
adds
right.
So
essentially
what
the
application
sets
gives
you
is
that
it'll
just
deploy
over
and
over
and
over
again.
So
someone
asked
very
good
question.
So
does
this
folder
structure
and
yaml
files
work
with
any
kubernetes
distro?
The
answer
is
yes,
kubernetes
because
customize
is
built
into
kubernetes,
and
so
since
customize
is
built
into
kubernetes
you'll.
Have
this
essentially
will
work
with
any
cluster
right?
You
can
just
do.
Customized
build
and
it'll
just
deploy.
A
Someone
says
they
have
a
rancher
cluster
should
work
with
rancher.
I
don't
know
what
rancher's
automation
is,
but
if
you're
using
argo
cd,
if
you're
using
flux
on
it,
it
should
just
work
right.
It
essentially
should
work
with
any
kubernetes
cluster.
I
do
have
examples
right
here.
This
example
is
going
to
go
away
because
this
is
kind
of
my
scratch
disk.
Let's
go
to.
A
Yeah,
let's
go
here:
let's
go
to
my
repositories,
I'm
just
going
to
paste
these
in.
I
have
an
example
for
openshift,
get
ops,
argo
and
flux
right.
So
if
you're
gonna
use
openshift,
there's
a
repo
to
get,
you
started,
you're
gonna
use
just
kubernetes
and
argo
there's
another
one,
and
just
to
show
you
because
I've
done
a
lot
of
work
with
flux.
Lately
this
one
flux,
you'll
notice,
although
I
have
like
three
separate
repos,
they're,
all
the
same
structure,
and
so
the
idea
is,
it's
just
a
basically
a
good
way
to
start.
A
So
this
is
how
to
start.
You
can
change
this.
However,
you
want
so
anyway,
I'm
running
late
for
my
next
meeting,
and
so
what
I'm
gonna
do
is
I'll,
follow
this
up,
probably
with
a
blog
post.
We'll
keep
talking
about
this.
Maybe
I'll
have
chris
short
back
on
as
as
a
guest,
ironically
right,
instead
of
instead
of
the
instead
of
the
host,
because
we've
done
a
lot
of
work
in
this
git
repo,
so
anyways.
Thank
you
all
for
joining.
A
I
appreciate
you
next
next
time
we're
gonna
start
at
our
usual
time.
So
again,
thank
you
for
all
your
patience.
We're
gonna
start
at
the
same
time
as
our
usual
time.
Right,
3
p.m.
Eastern
and
I
am
going
to
have
hashicorp,
come
on
and
talk
about
vault.
It's
actually
going
to
be
a
first
of
a
series
of
episodes
right.
So
it's
we're
not
going
to
have
actually
corporate
on
once
we're
going
to
have
a
multiple
times
to
talk
about
vault,
so
anyways
stay
safe,
stay,
safe
out
there!
Everyone!