►
From YouTube: GitOps Happy Hour (Ep 2): Avoid YAML Duplication
Description
Description:Join Christian Hernandez, GitOps Extraordinaire, for a journey through how to achieve GitOps in any number of ways. The occasional Red Hatters and special guest will join us too.
A
A
A
A
A
B
C
A
B
B
Who
don't
know
langdon?
Is
he
actually
does
the
level
up
hour?
So
he
does.
He
actually
does
another
show.
So
I
don't
know
if
any
of
you
that
are
on
I'll
shall
watch
that
show
and
kind
of
see
if
we're
doing
a
little
little
cross
training
here.
So
but
that's
that's!
Okay!
That's
that's!
That's!
To
be
that's
to
be
expected
and
yes.
B
B
So
yeah,
so
thank
you.
Everyone
thank
you
for
for
joining
another
another
episode
of
the
the
get
offs
happy
hour.
I
understand.
There's
there's
there's
meeting
fatigue
now
I
think
they're
like
or
like
virtual
event
fatigue.
I
know
I
feel
it
like
everyone
else,
so
I
actually
do
appreciate
everyone
coming
coming
on
and
tuning
into
the
stream.
So
so.
B
Everything
is
in
git
and
your
cluster
syncs
everything
from
this
git
repository.
So
the
first
thing,
though,
is
it's
like
okay?
Well,
I'm
going
to
store
everything
in
get
I'm
going
to
be.
You
know,
storing
all
my
cluster
manifest
as
you
can
get
all
my
my
deployment
manifesting
it
and
but,
like
you
start
noticing
it's
like
wow,
I'm
like
I'm,
really
duplicating
a
lot
of
yammer
because
it's
like
I
have
yaml.
B
Yaml
is
just
basically
everywhere
and
one
one
of
the
things
that
that's
some
people
come
to
realize
really
quickly
is
okay,
like
how
do
I
stop
duplicating
all
this
yam
all
right,
and
so
I
decided
to
take
a
very
you
know
very,
very
simple
use
case
and,
like
basically
just
start
like.
How
would
I
build.
B
You
know
a
deployment,
I
guess
get
repo
and
and
me
not
having
to
you
know,
duplicate
so
much
yaml
right
and
I'm
going
to.
We
talked
about
in
the
first
episode
the
the
different
templating
engines.
Right
customize
is
the
one
I'm
going
to
be
using
right.
Yeah,
it's
very
really
popular
helm,
I
guess
is,
is
the
other
reigning
champion
there's
also
a
case
on
it
and
other
things
like
that,
but
I
think
those
are
like
the
two
big
ones.
A
B
B
C
A
A
D
A
B
Yeah,
that's
that's.
I
guess
that's
a
good
topic
for
another
for
another
episode,
but
argo.
Really,
although
I'm
going
to
be
using
argo,
argo
is
not
the
ideal
ideal
tool
for
that
right.
You
can
make
it
work.
You
could
do
things
like
use,
use,
techton
and
do
pipelining
to
where
techcon
can
fire
up
something
like
openshift
hide
where
it'll
spin
up
a
cluster
for
you
or
you
know
you
have
some
sort
of
a
cube
adm.
B
B
That'll,
you
know
spin
up
a
cluster
for
you
or
or
whatever
right,
whatever
automation
process,
but
that's
like
and
then
use
argo
cd
to
sync.
All
of
that
for
you,
but
that's
not
really
our
go
cd's
job
right,
roci
doesn't
install
or
instantiate
a
cluster
for
you
right.
It's
it's
purely
for
deployment
purposes.
So
with
that
said,
cross
your
crusher
cross,
whatever
fingers
toes
whatever.
A
B
B
B
B
B
B
Going
like
from
the
beginning
right
like
how
do
I
start
like
I'm,
starting
basically
from
the
ground
up,
I'm
gonna
start
literally,
I'm
gonna
be
going
a
new
repository
right,
so
we're
we're
literally
going
from
nothing.
Let's
call
this
test
oops.
Let's
call
this
test
ops.
How
about
that
everything
with
ops
right
test,
ops,
twitch,
testing,
all
right
public,
oh
fine,
I'll,
add
a
readme.
Just
I'm
gonna
delete
it
anyway,
but
you
know
all
right.
So
I
start
what
what
my
goal
is.
B
Is
it's
gonna
be
a
deployment
repository,
so
the
application's
already
built.
This
is
just
how
I'm
gonna
deploy
it
right.
So
we
talked
about
having
separating
your
your
repos
to
where
you
have
your
application
code
in
one
repo,
you
have
its
deployment
artifacts
and
another
repo
and
all
the
infrastructure
components
is
in
yet
another
repo.
So
so,
let's
get
to
it.
What
I'm
going
to
be
doing
here
is,
let
me
do
a
t-mux
and
then,
what's
my,
I
always
change
my
binding,
and
hopefully
I
have
it.
Okay,
cool.
A
D
B
The
last
thing
you
see
it's
a
it's
a
it's
a
learning
thing
right,
yeah,
let's
get
this
little,
I
don't
know
how,
let's
see
if
it
shows
up
what
I'm
going
to
be
doing.
I
don't
want
to
move
the
window
hello,
just
one
of
those
days
escape
there.
We
go
all
right
up
there.
I
want
test
ops
right
and
so
what
I'm
going
to
be
doing
is
I'm
just
going
to
create
a
directory
called
apps
or
apps
right.
B
B
B
There's
a
there's,
this
concept
of
a
base
right,
so
there's
like
you'll,
see
a
lot
of
people
who
have
customized
the
directory
structures
is
like
they'll
have
a
base,
although
base
in
terms
of
like
in
the
in
the
yaml
is
deprecated,
but
I
still
like
to
use
the
base
concept
because
it's
going
because
it's
a
because
it
just
makes
sense
to
me
right.
It's
like
the
this
is
the
base
configurations.
This
is
meaning
that
this
these
yamls
are
going
to
be
applied
like
to
all
my
clusters
right.
C
B
So
so
you
notice
that
like
keep
track
of
the
left-hand
side
right,
so
I
thought
this
would
be
all
right.
So
the
first
thing
when
you
run
customize,
the
first
thing
it
looks
for
is
a
a
customized
customization.yaml
file.
This
yaml
file
basically
dictates
what
it
reads
in
and
what
and
what
it
does
right.
B
It's
see
if
I
spell
it
right
how
about
I
just
copy
and
paste
it
so,
let's,
let's
bend
this
guy
and
and
then
set
it
not
so
here,
resources
right,
so
you
get
resources
and
essentially
you're
just
going
to
list
the
yamls
that
you
want
loaded
in.
So
you
have
things
like
namespace.yaml
right
and
we're
gonna
have
to
name
these
things
accordingly,
employment.yaml
service.org
and
then
route
right
so
like
these
are
like
the
basic
components
of
an
application
right.
A
B
B
B
C
A
B
B
B
Yeah
yeah
json
path
is
another
thing,
another
language
right
that
you
have
to
know.
So
I
just
I
just
want
to
call
the
attention
line.
27
I
have
a
value
actually
26
and
27
right,
oops
2627
right.
I
have
an
environment
variable
color
and
the
value
this
application
takes
blue
and
it'll.
Show
you
a
blue
square
or
green
I'll.
Show
you
a
green
square,
any
other
value.
B
It'll
just
show
you
a
blue
square,
but
here
I
have
a
value
called
patch
me,
because
what
we're
going
to
be
doing
is
we're
going
to
take
this
base
yaml
file
and
apply
it
to
different
clusters
and,
depending
on
the
cluster,
I'm
at
I'm
going
to
do
something
different,
so
one
cluster,
I'm
gonna
have
blue
and
another
cluster.
I'm
gonna
have
green
and
and
but
we're
just
gonna
have
one
version
of
this
yaml
file
right.
B
So
this
is
kind
of
where
we're
customizing
templating
becomes
very
powerful,
especially
if
you're
doing
get
ops
right.
You
you're
gonna,
be
doing
everything
in
yaml
and
so
you're
gonna
you're
gonna
need
to
make
sure
you're.
Not
just
like
you
know,
constantly
copying
the
same
file
over
and
over
again
right.
You
have
one
version
and
it's
easy
to
take
care
of.
So
here
the
service
right
service
definition,
kind
of
the
oh
things
are
slowing
down
here.
For
me,
there
we
go.
B
B
Which
I
have
prepared
here,
the
thing
I
want
you
to
draw
draw
attention
to
you.
Guys
here
is
line
10
right
again
again:
host
patch
me
right,
I'm
in
different
clusters,
and
so
the
the
the
fqdn
is
gonna
be
different,
depending
on
which
cluster
I'm
at
so
so
this
is
another
thing
that
I'm
gonna
change.
I'm
gonna
have
one
route
definition,
but
I'm
gonna
change
it
as
I
load
it
in
right.
Don't
write
partial
buffer
all
right.
So
if
I
can't
customize
and
then
I
do
an
ls,
I
have
my
deploy.
B
Namespace
check
deployment,
check
service,
cool
route,
cool
all
right.
So
that's
my
base
right.
So
this
is
kind
of
straightforward.
There's,
nothing!
There's
nothing
get
upsy
about
this.
Just
yet
you
know
you
can
run
the
customize
and
it'll.
It's
basically
you're,
gonna
customize
says
it'll
read
this
like
okay,
when
when
you
call
me
I'm
gonna
load
up
all
these
manifests
that
you
told
me
to
tell
me
to
load
up.
So
I'm
going
back
here.
B
B
B
B
B
Under
resources,
same
idea
right
so
when
this
reads
the
customization
file,
it
says
it'll,
say:
okay,
we'll
read
everything
in
base
directory,
which
is
one
up
and
then
it'll
read
that
customization
and
then
it'll
load
up
all
those
resources
right.
So
with
this
one
line,
you're
loading
up
all
your
manifests
in
in
this
customized
file
right.
B
B
So
when
I
say
okay,
when
you
see
the
deployment
named
bgd
in
the
bgd,
namespace
patch
it
using
this
definition,
which
will
which
I'll
create
up
next
same
thing
with
the
round
right
so
line,
one
says:
okay,
go
up,
read
the
bases
and
then
that
in
turn
has
its
own
resources
list.
So
it'll
load
those
resources
in
when
it
reads
those
in
it'll
run
this
patch,
along
with
it
right.
So
you
can
do
all
kinds
of
stuff.
B
B
Hammer
so
here
yeah,
so
let's
go
back
and
take
a
look.
So
now
in
this
directory
I
have
the
deployments
and
the
customization
file
in
the
customization
file.
Remember.
I
said
that
when
you
re,
when
you
see
this
deployment
patch
it
with
whatever
you
have
in
this
yaml
file
right,
so
I
did
create
that
deployment.
B
So
if
I
go
to
base
and
do
a
deployment
right,
we
can
say
spec
right,
spec,
template,
spec
containers,
right,
environment
variable
the
first
one
value.
We
want
to
replace
the
value,
another
customization
right,
spec
templates
fact,
containers,
spec
templates
effect
containers.
So
I'm
here
the
zero
right,
which
is
like
the
first
one
right,
because
you
can
have
many
containers
in
a
pod
and
then
and
in
the
environment.
B
B
B
A
D
A
D
B
B
So
you
have
here
patches
right.
This
is
the
one
I
was
using,
there's
other
ones
that
you
can
can
use
and
the
other
other
things
how
how
you
can
manipulate
the
the
yaml
coming
in
right.
So
I'm
using
replace
right.
You
can
also
do
merge
right,
so
you
can
do
strategic
merges.
You
can
do
where
is
it
here?
B
Yeah
add
right.
So
I
want
to
add,
like
one
of
my
options
is
add,
I
want
to
add
to
the
like,
if
it's
an
array
right
like
I
want
to
add
two,
you
know
to
the
array,
something
or
other
I
want
to
replace,
there's
patches
strategic,
merge
right.
So
this
is
an
example
here
this
I
find
and
it's
the
cube
ctl
doc,
because
if
you
didn't
know,
customize
is
built
in
to
cube
ctl
and
and
by
extension,
oc.
B
B
B
Things
that
you
want
to
do
is
probably
in
this
in
this
in
the
stock.
So
so
I
recommend
reading
through
that
and
making
sure
you
know
you
you,
you
can
find
out
what
what
you
need
there
see
here
yeah.
So
I
was
curious
how
it
maps
the
helms,
delete,
key
option
and
the
new
post
render
flag
yeah.
So
I
don't
know
much
about
creating
helm
charts.
B
B
B
I'm
basically
replacing
a
cat
route
and
then
base
crowd.
Maybe
I
should've
done
it
the
other
way,
but
so
I'm
doing
I'm
gonna
replace
I'm
replacing
spec
host
and
then
patch
me
what
and
then
the
reason
I
have
patch
me
is
because,
because
I'm
doing
a
replace,
if
I
did
an
ad
I
would
I
would
have
to
have
that
removed.
B
B
B
B
A
C
B
So
yeah
I'm
gonna
use
argo
too,
so
I'm
just
gonna
create
application.
I'm
gonna
call
this
blue
green
deployment
project
default.
I'm
gonna
do
manual
sync
policy,
because
why
not
and
then
repository
url
good
question.
Luckily,
luckily
they
gave
me
that
option
because
I
forgot
to
edit
to
get
there's
no
get
offs
without
get
right,
so
this
will
get.
B
A
B
B
B
You
know
that
it's
gonna
read
all
those
because
up
there
sorry
based
customization
up
there.
It
basically
says
read
up
all
these
manifest
files
once
it
reads
that
going
back
up
to
blue.
It
says:
okay,
once
I
read
those
in
patch
it
with
everything
that
I
that
I
told
you
to
patch
it
with
a
route
and
a
deployment.
B
B
A
B
B
B
B
We
didn't
patch
the
namespace
or
the
service,
so
that
should
be
the
same.
The
deployment
right,
the
the
desired
manifest
here.
It
says,
oh
see
now,
instead
of
patch
me.
It
says
blue
nice
so
because
they
ran
that
customize
and
it
just
patched
that
that
configuration
file
and
then
the
other
thing
we
patched
was
the
router
right.
So
let's
look
at
the
desired
manifest
here
and
you
see
that
instead
of
patchme,
it
says
dot
app,
store
cluster,
one
oc
right,
so
this
is
still
syncing
here
come
on.
B
B
B
B
D
B
Yeah
only
then
like
that
the
deployment
I
actually
do
want
to
change
that
right.
So
I
said
I
wanted
green
right.
This
directory
says
green,
it
better
say
green.
So
I
basically
say
when
you
read
in
the
deployment
patch
this
to
green
cool
and
then
the
last
thing
is
the
route
right.
So
this
is
a
different
cluster.
There's
a
different
open
ship
cluster,
so
the
route
is
going
to
be
different.
Okay,
this
will
be
conveniently
conveniently
titled
oops,
a
green
right.
B
A
B
D
A
E
D
E
B
B
C
E
A
E
C
A
B
A
A
D
C
A
B
B
So
let's
take
a
look
and
see
what
happened
here,
the
deployment
right
so
the
deployment
in
the
cluster
two,
because
I'm
a
cluster
two.
Now
let's
go
to
the
desired
manifest
here
now
the
color
changed
to
green
right.
So
I
have
one
is
blue
one's
green,
but
only
only
one
deployment
yammel
that
I
needed
to
have
right,
so
I
didn't
need
to
have
two
separate
deployment
animals
to
get
this
going.
B
B
So
wait
till
this
comes
back
up,
but,
as
you
can
already
see,
if
you
go
here,
let
me
actually
go
back
and
do
this
here,
as
you
can
see
here,
on
the
left
hand
side
I
have
my
base
with
all
my
customization
says
that,
basically,
this
is
my
base
configuration
for
this
deployment
right,
so
everything
that's
common
for
my
deployment,
my
blue
directory.
I
basically
reference
my
base
and
then
do
some
patching,
so
I
only
have
to
do
define
the
deployment
once,
but
then
I
can
just
patch
it.
B
However,
I'd
like
right
same
thing
with
green
right,
so
going
from
blue
to
green
yourselves
as
easy
as
copying
that
whole
directory
and
just
kind
of
changing
a
few
parameters.
Now
I
have
two
deployments
each
with
the
same
base
deployment
configuration
but
different
different
parameters
right
and
I'm
using
customize
for
this,
you
can
use
helm
for
this.
A
lot
of
people
use
helm
for
this
as
well.
Let
me
see
if
this
is
deployed
yeah.
B
So
if
now
this
is
deployed
come
on,
come
on
baby,
no
whammies,
no
whammies
green
right.
So
now
I
have
a
green
square.
This
guy
has
a
blue
square:
blue,
green
woohoo
right
yeah
success,
one
of
the
one
of
the
cool
things
and
like
the
the
last
kind
of
thing
I
I
kind
of
wanted
to
go
over.
I
know
we
have
a
little
time
because
yeah
we
started
later
due.
A
B
A
B
B
D
D
A
B
B
D
B
These
are
powers
for
good,
so
I'm
going
to
create
here
so
basically
the
same.
I
should
actually
do
this,
I'm
starting
to
run
out
of
room
here
twitch.
Let
me
call
this
test.
Ops,
app,
always
take
green
and
then
we'll
copy
it
here
all
right.
So
here
I'll
have
the
deployment
right
we'll
test.
My
theory
we'll
do
wood
color,
like
landon,
says
and
then
let's
bi
route
right
and
then
the
route
for
my
other
cluster
is
apps
and
then
oops
all
right.
B
B
B
E
B
A
B
B
D
B
B
B
B
B
So
here
let
me
sync
this
because
I
did
manual
sync
so
here
I
did
the
langdon
all
night
repo
here
and
I
just
have
app
wood
color,
and
I
only
have
the
three
manifests
right:
the
customize
the
deployment
and
route,
but
even
so
it
created
the
name
space.
It
created
the
service.
It
created
everything
just
because
in
that,
in
that
you
know,
wood,
color,
customization
file,
I
said
for
your
base
and
go
to
the
get
go
to
this
git
repo.
B
So
this
lends
itself
to
a
lot
of
interesting
use
cases
right.
You
can
have
like
a
base
deployment
of
your
application
and
then
others
can
then
just
take
that
and
patch
it
and
then
make
it
do
whatever
it
needs
to
do
in
other
in
other,
repos
or
or
you
can
have
like
a
a
centralized
configuration
so
like
on
on
the
admin
side.
Right
like
you,
can
have
a
centralized
configuration.
B
That's
the
same
across
all
your
clusters,
except
maybe
in
this
cluster.
You
want
to
add,
like
an
extra
user
or
an
extra
service
account
for
something
else
or
an
extra
application
right.
You
can
take
the
base
of
this
of
this
repo
and
further
and
further
customize
it
right,
literally
and
and
figure
to
make
with
the
play
on
words
and
everything
right.
You
can
customize
it
with
a
customizer,
just
customize
it
just.
B
However,
you
like
right,
you
just
have
that
that
same
base
config
and
you
just
reference
that
in
your
customization
in
your
resources
right.
So
here
I
can
actually
in
the
resources.
Assistance
is
an
array
I
can
add
another
entry,
and
I
say,
oh
and
also
load
a
job
right
like
if
I
want
to
create
a
job
like.
Let's
see
how
that
looks
like
how
to
sample
a
job,
I
should
probably
be
more
specific
sample
job
kubernetes.
B
B
There's
a
job
pending
right,
so
let's
sync
it
right!
So
then
in
the
customize
file.
Now
it's
now
it's
running
the
job
right.
So
in
the
customize
file
right
you
can
see
that
with
very
little
yaml
I
I
can
start.
You
know,
have
a
bassy
ammo
with
very
little
yama.
I
can
actually
start
deploying
this
application
and
making
modifications
to
it.
So
so.
B
A
B
B
Right,
I'm
gonna
customize
build
everything
here
and
it
actually
does
it.
Does
it
for
you
here
see
there
you
go
it
just
spits,
it
just
fits
it
all
right,
so
you
can
just
do
it
locally.
You
do
customize
build
notice,
how
it
says:
spec
host
wood,
color
right,
and
so
you
can
just
do
it
locally
with
customize
right.
So
you
can
do
you
can
even
like
pipe
this
right
into
cube.
B
B
A
B
B
C
A
A
B
A
D
D
C
D
Then
then
at
least
I'll
have
constrained
choices.
You
know,
and
if
I,
if
I
run
into
a
problem
then
I
can
find
I
can
discover
the
problem
and
say:
oh,
I
really
need
this
thing
from
blah
blah
blah
and
let's
discuss
it,
and
we
can
give
you
three
months
as
an
admin
to
go
and
investigate
whether
it's
a
good
idea
or
not
without
you
know
so
so
I
think
anything
like
that
where
you
can
kind
of
say
here's
the
environment
in
which
you
can
play
that
is
clear
and
concise
and
is
modifiable.
B
You
you
actually
did
you
know,
you
touched
on
a
a
an
interesting
thing
and
I
think
it's
worth
it
iterate
reiterating
I'm
doing
it.
Maybe
every
episode
is
that
you
know
change
can
happen
right
and
you
do
have
that
communication
and
that
chain.
It's
it's
code,
review
really
really
right,
guess
get
ops
is
all
get.
C
B
And
a
lot
of
the
times
the
admins
don't
understand,
or
you
know,
or
they
just
don't
know
right
like
they're,
just
ignorant
like
you
know,
I
come
from
an
ops
background.
Is
that
you
know
no
there's
actually
like
a
process
that
happens
when
when
coders
right,
coders,
don't
just
throw
anything
up
there
right,
they
have
you
submit
a
pr
and
that
gets
you
know,
gets
reviewed
that
gets.
You
know
changed
multiple
times
like
how
many
times
did
you
submit
a
pr
and
the
what
you
submit
and
what
gets
merged
is
completely
different
right.
E
A
B
You
know
someone
can
make
a
pr
to
you
know
the
base
like
right.
Like
I
created
this
base,
you
know
the
test.
Ops.
Someone
can
submit
a
pr
to
permanently
change
the
scale
right
like
the
deployment
instead
of
deployment
one
to
ten,
but
I
can
look
at
that
as
a
as
a
maintainer
and
say
no
or
justify
it
or
no,
it
needs
review
or
you
need
to
go
through.
You
know
you
know
x,
y
z,
whatever
right
like
just
because
then
submitting
a
pr
doesn't
mean
it's
going
to
happen
right
so
well,.
D
B
B
C
D
And
the
and
the
change
may
be
actually
not
even
known
by
anybody
right,
so
you
know,
as
we
you
know,
as
I
can't
remember,
newton,
whoever
it
was.
That
said,
you
know
standing
on
the
shoulders
of
giants
right
I
mean
when
we
drop
a
modern
web
application
into
production,
there's
probably
literally
millions
of
lines
of
code
in
there
right
and
all
of
the
library
expansion
et
cetera.
It's
just
happening
kind
of
dynamically.
D
C
C
A
B
A
C
A
A
Yeah
yeah,
hopefully
these
the
power
outages
stay
away
from
christian,
but
stay
safe
out
there.
Everyone
seriously,
we
hope
you're.
Well
and
yes,
you
can
run
wordpress
on
openshift
at
scale
and
dynamically.
We
can
talk
about
that
more
in
chat,
so
yeah
join
our
discord
if
you
haven't
yet-
and
this
link
should
work
now,
better
folks
so
feel
free
to
join
the
discord
to
continue
the
conversation
and,
if
you're
interested
in,
what's
going
on
tomorrow,
check
out
the
calendar
we
have
a
couple
shows
looks
like
no.