►
From YouTube: Speed Your Istio Development Environment With vCluster
Description
Do you have concerns with cost, CPU and networking for your Kubernetes cluster? What is a cluster within a cluster? In this livestream, Fabian and Rich from Loft Labs along with Antonio from Solo.io will join Lin to discuss what, why and when to use vCluster and live demo how to speed your Istio (or others) dev environment with vCluster to ease your cost, CPU and networking concerns.
#istio #vcluster #dev
00:03 welcome + speakers intro
4:00 news for the week
7:20 vCluster intro, what, why and how
25:45 Demo on multiple vClusters
42:00 Demo on multiple vCluster with Istio multi clusters
51:49 wrap up
A
B
Hey
lynn,
hey,
hey
everybody!
Thanks
for
having
me
yeah,
I'm
I'm
fabian,
I'm
one
of
the
co-founders
of
loft
labs,
which
is
essentially
the
company
behind
vcluster,
I'm
also
the
initiator,
and
currently
a
co
core
contributor
of
vcluster
and
yeah.
We
are
very
happy
here
to
present
a
little
bit
about
why
we
think
virtual
clusters
might
be
the
future
in
the
multi-tenancy
for
kubernetes.
C
Yeah,
I'm
rich
burroughs,
I'm
a
staff
developer
advocate
at
loft
labs.
I
work
with
fabian
there.
I've
been
working
with
vcluster
for
a
little
bit
over
a
year.
Now
it's
super
exciting.
I've
been
very
happy
to
see
more
folks
getting
interested
in
it.
Hey
mourinho
and
I've
been
before
working
as
a
developer
advocate.
I
was
in
infrastructure
myself,
so
I
had
many
roles
over
the
years.
C
A
D
A
A
So
the
first
news
I
want
to
share
is
the
slim
has
alarms
1.12
ga
with
psyllium
service
mesh.
So
congratulations
to
the
psyllium
team.
One
thing
I
do
want
to
point
out:
in
the
blog
there
was
something
that
was
a
little
bit
confusing
to
people
who
have
issue
background.
I
know
my
my
co-worker
john
harvard
from
google,
who
is
also
a
fellow
toc
member,
have
been
questioning
about
some
of
the
statements
in
this
blog
regarding
the
gateway
api.
A
A
A
It
does
require
to
run
with
the
sidecar
based
approach
with
istio
the
site
called
this
free.
A
data
path
to
istio
is
potentially
in
in
the
future,
but
based
on
interest.
So
it's
not
something
that
has
been
done,
because
there
was
a
lot
of
confusing
on
twitter
and
people
thought
this
might
already
be
implemented.
So
the
answer
is
no.
The.
C
Other
thing:
hey
lynn,
I
actually
don't
think
about
this,
so
I
I
I
host
a
podcast
called
cube
cuddle
and
I
had
with
was
rice
on
and
we
talked
about
this.
There
was
a
listener
question
about
it
and
and
yeah
that
was
basically
her
take
is
that
you
know
they
would
pursue
potentially
doing
the
side.
Carless
thing
yeah,
depending
on
the
demand
from
the
community
or
people
from
the
community,
were
willing
to
contribute.
You
know
and
and
help
make
it
work.
So
yeah.
A
Yeah
that
makes
total
sense
so
just
to
clarify
for
our
audience.
The
other
thing
I
want
to
clarify
for
the
audience
which
I
thought
jung
hao,
would
also
ask.
The
good
question
is
whistling
so
there's
a
statement
about
whistling
service
mesh.
You
have
those
options
available
in
your
platform
and
can
even
run
a
mix
of
the
two,
so
we're
still
trying
to
understand
from
the
ceiling
team.
How
do
you
run
the
mix
of
the
tute?
A
Your
knowledge
is
not
going
to
be
easily
implemented,
so
we're
trying
to
get
a
little
bit
more
clarification
on
that.
Oh
one,
last
thing
I
want
to
share.
I
did
publish
your
blog
on
psyllium
layer,
7
capabilities
compiled
with
istio
in
different
categories.
So,
if
you're
interested
in
layer,
7
policies
and
the
comparison
between
psyllium
and
istio
definitely
check
my
blog
out
on
cncf,
so
that's
the
news
of
the
week
with
that.
Let's
turn
our
attention
to
virtual
cluster
and
how
does
it
help
so
I'd
like
to
ask
the
loft
team?
A
C
C
Yeah,
no,
no!
It's!
Okay!
Yeah!
It's
yeah!
Folks!
You
know
want
more
information.
If
you
go
to
vcluster.com,
it's
got
a
link
to
the
github
and
the
docs
and
all
of
that,
but
but
yeah
the
the
virtual
clusters.
Basically,
it's
it's
a
way
to
help
with
multi-tenancy
pain
and
think
of
it
as
having
a
a
namespace
on
a
shared
host
cluster.
B
Yeah
sure
so,
essentially
like
like
rich
mentioned
already,
what
virtual
cluster
is
doing
is
essentially
spinning
up
a
small
virtual,
a
small
kubernetes
cluster
by
default,
it's
k3s
inside
a
namespace
and
that's
small,
which
kubernetes
cluster
essentially
does
not
have
any
like
connected
nodes
or
any
network
plug-in
or
anything
inside
it.
It's
just
like
the
control
plane.
B
So
it
connects
the
pure
virtual
kubernetes
cluster
to
the
actual
worker
nodes
and
networking
of
the
host
cluster,
and
it
does
this
by
essentially,
they
are
synchronizing
resources,
so
lower
level
resources
like
pods
services,
secrets
and
so
on,
into
the
host
cluster
in
the
single
name
space.
So
essentially,
you
can
install
like
full-blown
operators
such
as
istio
or
like
search
manager,
any
ingress
controller
monitoring
solutions
inside
now,
just
like
a
single
name
space.
A
C
I
was
just
going
to
say
like
in
in
terms
of
like
multi-tenancy,
you
know
in
the
past.
I
think
there
have
been
at
least
a
couple
of
big
models
that
have
been
in
use.
You
know,
and
one
of
them
has
been
the
name
space
isolation
and
you
know,
like
fabian,
said
some
of
the
problems
with
that
are
that
you
know
people
need
to
suddenly
start
making
exceptions
and
stuff.
C
If
you
want
to
use
multiple
namespaces,
it's
you
can't
manage
global
objects
like
crds,
and
so
you
know,
namespace
isolation
is
kind
of
painful,
because
a
lot
of
times
you
end
up
with
the
platform
team
having
to
get
more
involved
because
they're,
the
ones
who
can
you
know,
administer
those
global
objects
and
things.
And
then
the
other
option
is
kind
of
like
every
tenant
gets
their
own
cluster
right
and
and
that's
obviously,
really
expensive
and
really
kind
of
a
management
headache.
A
Yeah
makes
sense
so
if
I
can
try
to
answer
why
virtual
cluster,
based
on
what
you
guys
were
just
mentioning,
so
I
guess
one
of
the.
Why
is
if
the
name
space
isolation
is
not
sufficient
for
this
purpose
and
the
second?
Why
is
the
cost
right
to
save
me
from
spin
up
10
cluster?
I
could
have
a
large
cluster
with
10
virtual
cluster
inside
of
it,
which
should
cost
a
lot
less.
C
Yeah,
no
you're
you're,
absolutely
right.
I
think
the
other.
The
other
big
one
is
just
that.
It's
really
fast
and
easy
right
so
so
say
that
you're
using
eks,
you
know,
and
you
want
to
provision
a
new
cluster.
I
I
think
they've
sped
it
up
some,
but
the
it's.
It's
still
really
slow
right.
It
takes
like.
C
Many
minutes
to
get
a
new
cluster,
it
takes
seconds
to
spin
up
a
v
cluster
and
because
they
are
like
really
ephemeral
and
really
fast
and
easy
to
like,
create
and
destroy
it.
It
means
that
you
can
use
sort
of
different
patterns.
You
can
think
of
it
more
as
like
dynamic,
ephemeral
infrastructure
instead
of
like
I
have
this
one
cluster
and
I
have
to
keep
it
always
working,
and
you
know
if
something
goes
wrong
with
it.
I
need
to
spend
hours
troubleshooting
it
and
and
all
of
that
kind
of
stuff
you
can.
C
A
C
I
I
think,
definitely
going
to
vcluster.com,
so
I
don't
know
if
somebody
can
type
that
in
the
chat,
I
don't
think
I'm
able
to
yeah.
Let
me
do
that,
but
but
there's
a
bunch
of
info
there,
like
I
said,
there's,
there's
a
link
to
the
github
repo
there's
a
link
to
the
docs.
I
think
the
docs
are
pretty
good.
There's
a
there's,
a
getting
started
kind
of
quickstart
doc
there
and
it's
really
easy
to
do
it's
a
it's
a
go
binary.
C
You
know
you
download
it
and
install
it
on
your
system.
You
have
to
have
a
cube
context
pointing
towards
the
host
cluster
that
you
want
to
use,
but
that
could
be
just
about
anything
right
that
could
be
a
local
mini,
cube
or
docker
desktop.
There's
lots
of
different
things
that
you
could
that
you
could
use
for
that
or
it
could
be
an
eks
or
a
gke
cluster.
You
know,
depending
on
what
you're
doing
like.
C
Maybe
you
want
to
have
something
more
like
the
production
environment
that
you
operate
in,
you
know,
or
maybe
you've
been
given
access
to
a
shared
cluster
and
you
want
to
run
a
v
cluster
inside
of
it.
So
I,
but
in
terms
of
like
the
basics
like
like
how
to
fire
up
a
v
cluster
for
the
first
time,
it's
it's
really
really
easy
to
do.
Yeah.
B
Like
rich
said,
it's
essentially
just
downloading
a
binary
running
a
single
command,
and
then
you
have
a
new
virtual
cluster
up
and
running,
which
should
work
also
in
every
kubernetes
cluster.
So
essentially,
as
long
as
the
kubernetes
is
somewhat
like,
a
certified
distribution
of
of
kubernetes
vcluster
will
work
with
it.
B
You
don't
need
elevated
permissions
as
well,
so,
even
if
you
are
only
able
to
access
like
a
namespace
in
a
shared
cluster
and
you
you
need
somehow
the
capability
of
installing
an
operator
you,
you
also
can
install
it
there,
since
it's
not
requiring
you
to
like
change
any
low-level
things
on
notes
directly.
It's
rather
just
like
deploying
a
small
kubernetes
control
plane
and
then
connects
to
that
for
every
other
kubernetes
request.
B
A
B
A
Yeah,
it
makes
sense.
Actually
I
have
one
more
question
before
we
get
to
the
demo.
I
know
the
audience
probably
wants
to
see
your
demo
now,
so
I've
been
using
k3
as
myself,
so
educate
me.
How
is
virtual
cluster
different
than
I
just
spin
up
multiple
k3s
clusters,
myself
for
my
istio
or
whatever
development
I
want
to
do.
C
B
So
it's
not
too
much
different
than
spinning
up
kfcs,
however,
like
they're
in,
like
the
the
the
basic
idea,
is
very
different,
since,
if
you
spin
up
a
k3s
and
k3s
essentially,
is
just
a
certified
kubernetes
distribution
where
they
packed
all
the
different
binaries
into
a
single
one
and
what
it's
doing
it's.
It's
spinning
up
a
real
kubernetes
cluster,
similar
to
like.
B
For
example,
is
doing,
and
the
the
difference
here
is
that
when
you
run
k3s,
you
will
also
like
install
the
cubelet.
You
will
install
a
cni
plug-in.
You
will
essentially
need
a
full
vm
to
run
this
k3s,
while
with
virtual
cluster,
you
are
reusing,
an
already
existing
kubernetes
cluster
to
essentially
separate
it
or
partition
it.
If
you
want
so
into
different
ones.
So,
essentially,
you
can
think
of
it
as
like,
a
virtual
machine
compared
to
a
physical
machine.
B
It's
much
more
efficient
since
are
essentially
running
k3s,
with
50
disabled
and
using
another
kubernetes
to
schedule
the
workloads
as
well
the
the
disadvantage
of
a
virtual
cluster
compared
to
k3s.
Is
that
like,
if
you
require
hardware,
boundaries
or
hardware
isolation,
for
example,
on
the
networking
site
or
on
the
workload
site
itself,
then
it
might
make
sense
to
have
like
different
physical
clusters
and
and
not
virtual
clusters,
but
usually
so,
for
example,
for
local
development.
C
B
C
A
Yeah
makes
sense
so
that,
as
in
summarize,
I
would
capture,
is
you
only
use
50
of
k3s
so
instead
of
the
full
blown
k3s
and
in
order
for
me
to
work
with
k3s
directly,
I
would
be
working
with
virtual
machine
well
with
vcluster.
I
would
just
work
with
kubernetes
cluster
and
then
use
that
to
spin
re
virtual
cluster,
so
it
could
potentially
be
a
lot
easier
because
I
don't
have
to
worry
about
vms
and
yeah
and
some
more
lightweight.
I
guess
the
benefits
comes
with
50,
yes
and
speed.
D
I
just
want
to
mention
that
here
the
picture
that
you
are
you
can
see
now
it's
my
local
environment
and,
as
you
can
see,
if
I
want
to
do
something
similar,
I
would
use.
I
would
need
to
use
something
like
a
k3d
which
is
the
multi-cluster
version
of
k3s,
and
that
brought
me
a
lot
of
complexity,
a
lot
of
complexity
that
I
didn't
need.
So
this
is
why
I
personally
I
found
v
cluster
because
to
solve
that
problem,
I
just
needed
some
main
cluster
in
this
case.
D
C
C
A
D
Yeah,
as
you
can
see
here
since
I'm
working
with
mac,
I
need
to
have
some
lima
vm,
which
is
provisioning,
the
k3s,
the
main
cluster.
And
after
that
point
I
just
keep
it
up
and
running
it's
very
lightweight,
so
I
don't
mean
to
lose
it.
But
since
I'm
working,
I'm
just
switching
between
other
v
clusters,
switching
totally
the
environment,
then
in
the
demo
weekly.
So
we
will
show
it.
A
D
All
right,
so
I'm
going
to
start
saying
that
the
gods
of
the
demos
they
are
not
with
me
today,
so
this
picture
crosses
right
before
the
meeting
and
I
was
like
screaming
out
like
oh
jesus,
oh
jesus,
so
I
just
use
an
aws,
an
eks
cluster
and
the
demo
is
gonna
run
on
eks
and
we're
gonna
spin
up
a
v
cluster
on
on
an
eks,
but
in
general
this
is
the
picture
I
use
in
local
to
give
a
reference
about
how
I
work
in
a
daily
basis.
We
are
like
a
team.
D
A
very
nice
team,
say
hello
to
my
to
my
team
in
mia.
We
are
under
danny
jannot
and
well.
He's
driving
us
to
give
pre-sales
presents
means
that
our
users
they
are
coming
to
us
with
hey.
I
have
this
crazy
idea
and
we
we
just
need
to
make
it
possible
and
matching
our
products
which,
in
this
case,
is
based
on
istio
and
multi-cluster
environments
on
istio.
Well,
it's
gloomesh!
So
it's
everything
about
service
mesh.
So
the
picture
is
like
that
today
I
was
working
here
we
go
and
I've
got
my.
D
D
D
D
D
And
we've
got
a
simple
deployment
of
istio
with
a
with
hello
world,
as
you
can
see,
the
picture
is
different.
Here
is
the
picture
with
my
main
cluster.
Here
is
the
picture
within
only
the
v
cluster,
so
I'm
gonna
keep
this
open
just
to
you
to
have
it
as
a
reference.
I
in
a
normal
scenario,
you
are
not
you
don't
care
about
the
main
cluster
anymore.
D
A
D
Yeah
so,
as
you
can
see
here
when
when
I
deploy,
then
I
will
show
you
when
I'm
deploying
a
v
cluster
when
I'm
deploying
a
v
cluster,
it's
unstable
set
and
I
stay
full
set
with
their
persistent
volume
right,
and
this
is
actually
the
the
the
components
to
make
the
effect
about.
I
have
another
cluster.
D
C
Right,
yeah
one
one
other
thing
too,
when
so
what
v
cluster
does
is
when
those
pods
get
scheduled
on
the
host
cluster
on
the
underlying
cluster?
They
get
renamed,
and
so
it
depends
both
the
name
of
the
v
cluster
and
the
name
of
the
name
space
within
the
v
cluster
to
the
pod
name.
So
that
way,
you're
never
going
to
get
collisions
because
you're
never
going
to
have
more.
D
Exactly
so,
as
you
can
see
here,
my
hello
world
me
as
a
user.
I've
got
my
hello
world,
but
the
real,
the
real
flash
it's
here.
This
is
in
the
main
cluster.
What
is
the
real
part
happening
in
an
isolation?
So
well,
I'm
working
and
suddenly
my
colleague
comes
up
and
say,
like
hey
antonio,
these
guys,
the
this
user
wants
a
proof
of
concept
so
quickly.
We
need
to
deliver
this.
Oh
jesus,
christ,
stress
multitasking,
so
I
mean
it's
funny
because
yeah
I
mean
those
mono
tasks.
I
don't
know
who
works
like
that?
D
No,
let's
sorry,
I
need
to
get
back
to
the
main
one
and
I'm
gonna
pause.
This
is
one
one
thing
I
really
love
by
big
cluster.
It's
like
I'm
gonna
pause.
What
means
is
like
I'm
gonna
scale
down
that
that
stayful
set
and
also
automatically
all
the
resources
that
they
are
related
to,
that
stateful
set
that
v
class
that
are
going
to
be
turned
down
so
not
consumption
anymore.
I'm
working
in
local
and
I'm
very
concerned
very
concerned
about
the
resources
that
I'm
utilizing
yeah.
A
D
C
Let
me
just
add
one
thing:
antonio,
the
pause
thing
is
great
and
that
actually
wasn't
in
the
initial
version
of
the
cluster
that
got
added
afterwards,
and
one
thing
that's
been
really
great.
Is
the
community
around
v-cluster
has
been
really
active
and
people
have
come
up
with
a
lot
of
great
ideas,
and
so
a
lot
of
the
things
that
we've
added
since
the
initial
release
have
actually
come
from.
You
know
user
feedback
from
the
community,
so
it's
been
really
fun
to
see
it
grow.
D
All
right
so
next
step,
I'm
gonna
create
I'm
gonna
spin
up
a
v
cluster
right.
Let's
do
it,
let's
see
how
fast
it
is.
We're
gonna
see
here
that
sooner
or
later
we
will
see
a
new
pot.
We'll
saw
it
there.
Yes,
my
hood
is
your
test.
This
is
the
task.
I
have
to
run
right
now,
as
you
can
see
already
created
15
seconds
to
have
already.
D
C
Oh
sorry,
antonio,
I
just
want
to
mention
one
thing
when
when
v
cluster
is
creating
the
cluster
inside
that
name
space
on
the
host
cluster,
that's
just
a
helm,
command,
and
so
the
v
cluster
cli.
Does
that
for
you?
If
you
do
v
cluster
create,
but
you
could
also
just
use
helm.
You
know
if,
like
a
sort
of
part
of
some
git,
ops,
workflow
or
something
like
that,.
D
So
we
connect
actually
I'm
going
to
pull
the
new
cube
config
right
to
connect
to
that
v
cluster.
This
can
be
done
in
one
one
step,
I'm
just
doing
it
in
two
just
to
show
the
capabilities.
So
now
I
have
my
own
hood
with
environment
is
to
test,
and
if
I
do
get
an
s,
I
see
that
I
have
like
a
fresh
new
cluster.
D
A
B
That's
a
good
question,
yes,
and
no.
The
problem
so,
what's
what's
certainly
possible,
is
essentially
since
the
state
of
the
virtual
cluster
is
saved
and
by
default
at
least
as
an
sqlite
database
on
a
persistent
volume,
you
can
just
use
like
a
backing
or
a
backup
solution
like
valero
back
up
that
namespace
and
like
essentially,
yeah
restore
it
in
another
kubernetes
cluster,
which
is
also
why
this
usually
I
mean
this.
B
This
works
most
of
the
time,
however,
there's
some
smaller
limitations
to
this,
for
example,
that
naming
references
could
be
wrong
inside
the
namespace
and
like,
for
example,
the
the
persistent
volumes
also
need
to
be
backed
up
by
valero.
But
if
you
do
this-
and
yes,
it's
possible
to
also
kind
of
create
a
snapshot
from
a
virtual
cluster.
However,
it's
not
yet
super
easy
to
do
that,
but
yeah.
We
are
working
on
this
to
make.
Eventually
this
hopefully
happen
to
create
your
own
vcluster
snapshots
and
easily
switch
those
between
different
kubernetes
clusters.
A
D
D
C
Yeah-
and
I
think
I
think
that
that's
a
that's-
a
really
big
advantage
for
that-
that
local
development
kind
of
scenario
you
know,
even
when,
when
you
were
talking
about
using
k3s,
I
think
it's
going
to
be
slower
to
like
create
and
destroy
those
those
k3s
clusters
than
it
is
to
just
have
one
cluster
and
create
a
bunch
of
virtual
clusters.
Inside
of
it.
A
D
D
C
D
As
you
can
see,
the
pots
were
created
just
now.
Now,
I'm
back
to
the
hello
world
application
that
I
was
working
on
previously
so
yeah.
That
is
the
demo
I
mean
it
took
me
five
minutes
to
really
have
the
full
power
of
a
kubernetes
cluster,
deploy
in
istio
testing
out
things
about
istio
and
and
then
get
back
to
my
previous
life.
D
D
C
Yeah
you
can,
you
can
have
an
ingress
when
there's
we
have
documentation
in
the
the
cluster
docs
that
walk
people
through
how
to
how
to
set
up
that
sort
of
stuff.
So
the
port
forwarding,
I
think,
is
really
handy.
You
know
just
to
get
started
really
quickly,
but
if
it's
you
know
more
of
a
long-lived
cluster,
that's
going
to
be
around
you.
You
definitely
can
can
set
it
up
with
ingress
or
something.
A
Very
cool
okay,
so
I
guess,
if
the
audience
and
me
are
convinced
you
know,
virtual
cluster
is
really
useful
for
single
cluster
right,
as
in
tony,
you
showed
us
your
spin
upper.
It's
your
cluster.
You
also
spin
up
another
test
cluster.
How
does
this
work
with
multiple
clusters?
The
reason
why
I'm
asking
is
in
istio
in
service
mesh
multi
cluster
is
a
very
common
topic
where
people
want
to
connect
their
services
across
multiple
clusters.
D
D
This
is
in
my
opinion.
This
is
quite
accurate
because,
usually
in
a
production
environment
you
have
different
networks.
So
the
way
you're
going
to
have
like
multi-cluster
multi-mess
in
different
networks.
This
is
going
to
be
the
deployment
that
you're
going
to
use
and
what
happened
in
my
two
v
clusters.
It's
in
one
I'm
deploying
this
and
in
the
other
one
and
deploying
that
maybe
it's
gonna
take
a
bit
longer.
That
is
better
to
have
questions
and
things
like
that,
but
I
just
want
to
say
that
we
proved
it
before
that.
D
We're
gonna
use
the
the
external
endpoint,
the
external
ip
in
aws
is
the
hostname
and
in
google
is
going
to
be
an
ap
and
your
local
is
going
to
be
a
plain
ap
so
to
make
the
connection
in
a
multi-cluster
you're
going
to
use
that
ip
or
that
hostname,
so
that
this.
This
service
is
gonna
connect
to
the
this
gateway
to
perform
the
east-west
traffic
across
the
both
clusters.
D
D
D
So
I
got
my
contacts
one
and
two
and
let's
see
if
it's
already
up
and
running
we're
gonna
connect
to
cluster
one
and
we're
gonna
connect
to
cluster
two,
and
that
is
the
way
that
we
are
able
easily.
There
is
no.
There
is
no
brainer
here
because,
as
I
said,
the
communication
is
gonna
go
through
the
through
the
external
ip
from
the
cluster,
so
the
behavior
is
exactly
like
if
they
were
totally
different
clusters
and
works
smoothly.
I'm
personally
using
this
to
test
the
multi-cluster
scenarios
with
istio
yeah.
A
Yeah
totally
so
just
to
help
our
audience,
I
think
antonia
in
your
cluster,
one
of
your
cluster
was
running
hello
version.
One
and
the
other
cluster
was
running
on
version
two,
so
the
command
you
showed
it
shows
it
was
able
to
hit
a
hello
version,
one
in
one
of
the
cluster
and
also
hello
version
two
in
the
other
cluster.
D
C
D
It's
a
when
I'm,
as
you
can
see
here,
I'm
when
you
talk
about
that.
These
are
the
you
mentioned
that
you
could
install
it
through
helm,
so
the
cli
is
embedding
helm
so
yeah.
This
is
my
configuration
as
you
can
see,
I'm
using
a
key
3s
every
single
v
class
is
a
k3s.
I'm
faking
the
no
I'm
not
faking
the
notes,
I'm
taking
the
real
notes.
D
I
could
fake
it
easily.
So
even
with
one.
My
for
example,
the
classic
standard
deployment.
Three
nodes
could
be
a
in
the
virtual
cluster,
simulating
that
I've
got
endless
amount
of
nodes
and,
as
you
mentioned,
the
isolation
feature,
I
don't
use
it
here
and
I
don't
use
the
quota
just
because
I'm
previous,
as
you
can
see,
that's
because
well,
I'm
the
owner
of
these
so
yeah
and.
C
B
A
B
A
Yeah
at
the
minimum
you
would
have
to
allow
like
the
gateways
on
to
is
your
cluster
to
be
able
to
communicate
because
all
the
paths
within
both
of
the
cluster,
the
communication
like
across
cluster,
has
to
go
through
the
gateway.
So
at
the
minimum
you
have
to
make
sure
that's
not
blocked
by
the
network
policy,
so
probably
with
more
work.
I
guess
it's
possible.
A
Yeah,
this
is
great
yeah
I
mean
the
the
the
stuff
you
mentioned
about.
You
know
set
up
the
secrets
across
different
kubernetes
cluster
for
istio.
That's
what
we
do
in
glue
mesh
to
save
users
from
setting
that
up,
because
we
didn't
believe
it
was
the
most
secure
way.
So
glumash
has
a
relay
to
help
user
to
even
simplify
that
so
with.
D
That
I
would
like
to
maybe
just
have
to
mention:
maybe
there
is
there-
is
the
in.
Thank
you
for
the
a
lot
guys
because
they
allow
us
to
present
a
post
to
write
a
post
into
their
website,
and
there
is
the
example
that
you
just
mentioned
about
gloomes
gloomes
multi-cluster,
with
v-cluster
in
your
own
local
environment.
So
it's
just
you
can
just
go.
Do.
D
So
well,
basically,
you
can
just
follow
copy
and
paste
commands
and
it's
going
to
be
working.
There
is
one
tricky
thing
in
your
own
local
embed
when
you
deploy
the
clusters
in
the
in
the
in
the
cloud.
Everything
is
easy
because
the
the
ips
are
really
attached
to
the
services
kubernetes
services
when
you
are
working
in
local.
D
Probably
it's
meant
it's
good
good
to
mention
you
can
find
here
in
fulfill
ctl.
You
can
find
the
resources
where
to
do
run
this
in
local,
because
what
you're
going
to
need
is
something
called
metal
b.
This
is
going
to
be
that
component
in
chart
of
a
chat
attaching
a
local
ip
in
order
to
be
able
to
have
the
local
experience,
because
if
you
deploy
this
with
another
simple
container,
you
need
to
always
play
with
the
network
in
order
to
make
the
connection
between
the
host
and
the
container
and
back
and
forth
both
directions.
C
I
I
want
to
mention
one
more
thing
to
lynn.
You
had
asked
earlier
about
how
people
could
learn
more,
and
I
mentioned
that
at
vcluster.com.
There's
those
links
to
like
the
docs
and
the
github,
but
there's
also
a
link
there
to
a
community
slack
that
we
have,
that
has
a
v
cluster
channel
inside
of
it
and
the
maintainers
are
in
there
there's
other
users
who
are
using
vcluster.
C
So
if
you
have
questions
you
want
to
talk
to
somebody,
you
know,
after
the
call
about
it,
that
would
be
a
great
place
to
go.
There.
Also,
now
is
a
v
cluster
channel
in
the
kubernetes
slack,
the
main
kubernetes
community
slack,
but
it's
very
new.
I
don't
think
there's
a
lot
of
people
in
there,
so
you'll
probably
get
more
help.
If
you
go
to
our
community
slack.
A
A
A
C
A
Okay,
that
makes
sense
totally
yeah,
okay,
cool.
I
think
this
is
it
for
this.
I
want
to
take
a
minute
to
thank
you.
Thank
you,
antonio
for
the
wonderful
demo.
Thank
you,
rich
and
fabian
to
bring
your
insights
of
we
cluster.
I
feel
like
I've
been
educated
about
we
cluster.
You
know
I'm
able
to
try
for
my
next
development,
environment
or
demo
environment
that
I'm
trying
to
set
up
now.
So
thank
you
so
much
for
being
willing
to
spend
time
to
educate
me
and
our
audience
about
this
topic.
C
D
C
A
So
folks,
if
you
guys
think
this
is
interesting,
we
would
really
appreciate
a
thumbs
up
to
our
youtube
video.
We
would
also
appreciate
subscribe
so
that
you
don't
miss
any
of
our
future
episodes
and
if
you
have
any
topics,
you
would
love
to
see
related
to
envoy,
istio,
kubernetes,
evpf
and
graphql.
Please
let
us
know-
and
we
will
try
to
find
the
right
expert
to
talk
about
that.
A
To
see
that
yeah
awesome
well,
we
got
a
feedback
from
a
couple
of
our
audience.
Thank
you.
So
much
jorge
awesome
demo.
We
also
got
a
feedback
from
kim.
Thank
you
for
the
enriching
insights
and
demo.
Thank
you
guys
so
much
for
the
feedback
really
really
appreciate
that.
That's
really
what
got
us
going
with
these
live
streams,
so
really
appreciate
that
thanks
all
for
joining
us
today,
bye
now
thank.