►
From YouTube: Deploying Linkerd with Meshery
Description
Presenting Meshery (https://meshery.io) - the universal Service Mesh Management Plane that allows to deploy, configure and benchmark all the leading service mesh implementations.
In this demo I show how to deploy Linkerd (https://linkerd.io) to a #kubernetes cluster using Meshery.
Meshery is a project by Layer5 - https://layer5.io
Presented by Ant Weiss of Otomato Software (https://otomato.io)
Follow me on twitter for more service mesh videos: https://twitter.com/antweiss
A
Hello,
folks,
how
are
you
and
vice?
That's
me,
I'm
here
and
today
we're
going
to
be
talking
about
link,
Rd
and
measure
e,
how
to
use
measure
e
to
deploy
a
linker
D,
how
to
use
linker
D
with
misery.
So
both
logos
of
Mary
on
the
left
and
off
linker
D
on
the
right
are
here
on
the
screen
and
even
from
this
picture,
you
can
see
that
these
two
make
a
perfect
pair,
even
the
color
palettes
match.
But
what
are
they
before
we
talk
about
what
they
are
we'll
have
to
talk
about.
A
What
a
service
mesh
is,
because
this
whole
video
will
be
about
service
measures.
So
what
is
a
service
mesh?
Do
you
know
what
a
service
mesh
is
I'll
tell
you
what
I
think
it
is
so
as
Christian
pasta,
the
evangelists
for
the
so
IO
company
says
the
moment.
You've
broken
down
your
application
into
a
myriad
of
small
services
and
the
moment
you've
deployed
those
services
on
your
cloud
platform,
you
invariably
already
get
a
service
mesh.
A
The
multitude
of
communications
transactions
network
connections
between
your
services
is
what
forms
this
mesh,
but
to
what
do
we
mean
when
we
say
that
service
matches
technology
well
service
mesh?
Is
this
new,
exciting
tag
that
allows
us
to
take
this
multitude
of
communications
between
our
services
and
to
be
able
to
analyze
and
control
this
network
traffic?
So
how
is
this
being
may
made
available?
Well,
the
service
match
consists
of
two
main
parts:
the
data
plane
and
the
control
plate.
The
first
part
is
what
allows
us
to
capture
the
traffic
and
that's
the
data
plane.
A
The
upper
part
of
our
diagram
capturing
the
traffic
is
being
enabled
by
a
mesh
or
a
network
of
small
lightweight
proxies,
whereas
each
proxy
resides
next
to
each
one
of
our
services
in
kubernetes.
This
will
be
a
sidecar
container
inside
each
one
of
our
pods.
Now
what
this
process
does?
Is
it
intercepts
all
of
the
network
traffic
going
into
our
pod
and
going
on
out
of
our
pod
and
by
intercepting
that
traffic
it
makes
it
available,
makes
it
analyzable
and
controllable
by
with
the
help
of
the
lower
part
of
our
diagram,
namely
the
service
mesh?
A
Control/Play
service,
mesh
control
plane
is
the
brain
of
the
mesh.
It
allows
us
centralized
control
of
the
mesh
and
it
usually
will
be
consisting
of
such
components
as
well,
some
kind
of
a
control
interface,
which
would
be
user
interface
and
command
line
interface
or
an
API,
and
we
usually
will
expect
some
competence
dedicated
to
the
observability
OPA
network
traffic
inside
the
mesh.
A
So
we
can
observe
what
happens
with
the
traffic
and
there
will
be
some
kind
of
a
policy
engine
here
that
defines
what
kind
of
traffic
is
allowed
and
what
kind
of
traffic
is
prohibited
in
the
mesh
and
maybe
what
and
how
we
want
to
shape
the
traffic
according
to
which
rules
and
what
policies
do
want
to
shape
the
traffic
that
occurs
among
other
services.
And
finally,
when
there
is
so
much
traffic
and
so
much
data
transactions
between
our
services
in
the
cluster.
We
do
want
all
of
this
traffic
to
be
secure
to
be
encrypted.
A
And
how
do
we
do
that?
That's
not
that
easy
to
allow
our
services
to
communicate
securely.
In
order
to
make
this
easier,
we
do
expect
the
service
mesh
to
provide
us
some
kind
of
a
mutual
TLS
solution
and
that
that's
again
what
most
service
measures
do
now
after
we
define
what
service
meshes,
are,
let's
see
how
and
what
is
linked
early,
but
before
that,
let's
talk
about
measure
ease.
A
So
what
is
measure
measure
is
a
project
developed
by
a
service
mesh
community
named
layer
5
which
I'm
proud
to
be
part
of
what
is
later
fire
well
layer
layer
5
is
a
service
mesh
community,
where
we
study
service
missions.
Where
we
talk
about
service
meshes
where
we
promote
service
meshes,
we
would
have
service
mesh
parties
and
service
mesh
concerts
where
we
all
get
crazy
in
the
service
mesh
pit.
No,
that's
not
actually
what
it
is
about
so
yet
service
mesh
is
a
community
know.
Layer,
5
is
a
service
myth.
A
Community,
that's
dedicated
to
three
main
goals.
First
of
all
is
creating
the
service
mesh
landscape
so
that
any
organization
or
any
individual
interested
in
evaluating
in
deploying
service
mesh
can
go
to
that
landscape.
Studied
landscape,
understand
what
it's
all
about.
What
different
meshes
are
about
what
different
solutions
on
the
market
are
about
what
each
one
of
them
provides
now
in
order
to
study
that
and
in
order
to
provide
actual
tools
for
playing
with
service,
meshes
we're
building
software,
and
that
would
be
Mary.
A
The
project
that
we'll
be
looking
at
today,
Mary
is
a
unified
management
plane
for
all
of
the
leading
service
meshes
on
the
market.
Today,
and
by
being
that
management
plan,
it
actually
becomes
a
playground
for
different
service
meshes
and
it
allows
us
to
not
only
install
and
configure
service
meshes.
A
And
finally,
what
is
linker
D
well
Wingra
D
is
one
of
those
service
meshes
that
we
study
and
developed
at
layer
five.
So
what
is
linker
D?
Actually
we're
talking
about
the
project
called
linker
d2o
and
it
comes
from
the
service
pioneers
its
elder
brother,
the
linker
D,
the
original
one
can
be
was
developed
by
the
guys
at
a
company
before
that
the
company
named
linker
D
and
they
were
actually
the
ones
who
coined
the
term
service
mesh.
Nobody
was
talking
about
service
measures
before
them
now,
specifically
Winkler
g2o
is
a
service
misdirected.
A
All
that
Cuban
air
is,
unlike
other
service,
meshes
that
talk
about
being
able
to
deploy
the
mesh,
not
only
in
kubernetes
but
also
outside
of
the
cluster
linker
D
is
for
Cubans
only
and
that's
why
it's
probably
the
best
service
match
for
kubernetes.
It
uses
these
specially
developed
cycle
proxy
called
linker
D,
which
is
written
in
rust,
one
of
the
most
exciting
modern
programming
languages
also
very
performant.
It
employs
the
approach
of
minimum
configuration.
That
means,
once
you
have
link
ready.
One
is
you
have
it
deployed,
you
already
start
getting
value
out
of
it.
A
It
has
telemetry
built
in
it's
known
for
its
high
performance
again
compared
to
other
service
measures
on
the
markets,
especially
steel,
and
it's
known
for
its
operational
simplicity.
Now
that
we've
talked
about
all
of
this
without
further
ado,
let's
go
to
the
practical
part,
let's
install
memory,
let's
install
linker,
do
with
measuring
and
see
how
they
all
play
together
nicely.
B
A
I
am
you
can,
of
course,
always
install
massery
with
the
help
of
the
wonderful
package
manager
for
Mac
OS,
so
you
can
just
go
brew,
install
measure
e,
but
we
won't
be
doing
that
today.
Instead,
we'll
use
the
Linux
approach
which
works
on
Linux
and
on
Mac
as
well,
because
I'm
aware
of
the
fact
not
all
of
you
out
there
are
using
Mac's
for
your
work,
so
we'll
use
curl
pipe
to
bash,
which
is
of
course
not
a
very
secure
option.
A
But
again,
if
you're
writing
this
on
your
own
computer,
if
you're
only
using
this
for
evaluation
on
your
testing
environment,
that's
perfectly
okay,
nothing
to
be
afraid
of,
and
so
that's,
where
we'll
be,
bringing
our
measure
installation
script
from.
Let's
see
how
this
works
you,
so
this
command
downloads
everything
needed
for
mastery
of
the
metric
components
and
also
the
wonderful
mastery
command-line
utility,
the
mastery
CTL
that
we'll
see
in
a
moment
now,
once
all
the
components
are
downloaded,
the
script
also
runs.
A
Massery
starts
all
of
the
containers
that
grandma
Sheree
by
the
way,
make
sure
it
does
require
docker
to
be
installed
on
your
laptop
on
your
desktop,
wherever
you
run
it
on,
if
you're,
not
installing
it
straight
to
your
cluster
and
that's
not
what
we're
doing
here
and
once
mesh
restarts
it
already
opens.
You
know
browser
you
can
see
that
it's
already
configured
to
work
with
the
kubernetes
cluster
that
taken
from
my
original
cube
serial
context.
Yeah.
You
can
see
it
right
here,
that's
my
cluster
on
the
IBM
cloud,
and
this
is
Mishra.
A
So
what
we
have
on
Mishra
here
would
first
of
all
see
the
connection
status.
We
can
see
all
of
the
active
adapters
running.
We
can
see
the
adapters
on
the
left-hand
side
too.
You
can
see
that
we
currently
have
an
adapter
for
consult
from
Hoshi
Corp
for
Sto.
That's
the
probably
the
market
leader
in
service
measures
today
for
network
service
mesh,
which
is
a
totally
different
kind
of
a
mesh
and
for
octarine
and
citrix.
But
today
we
will
be
talking
about
linker
this.
A
So
if
we
go
to
linker
B,
we
will
see
that
there
are
three
boxes
here
and
we
want
the
leftmost
box
that
allows
us
to
deploy
a
service
match
on
our
cluster.
Now
before
deploying
the
service
mesh
will
notice
that
up
here
it
says
namespace
default,
so
by
default,
Mary
will
deploy
anything
that
you
tell
it
to
deploy
into
the
default
namespace.
A
We
don't
want
this
today
because,
if
you've
ever
tried
playing
with
linker,
do
you
understand
that
linker
expects
its
components
to
be
deployed
to
the
linker
dining
space
and
that's
exactly
what
we
want
Mary
to
do
here
by
the
way,
if
you
think
we
should,
when
installing
linker
D,
ignore
whatever
the
namespace
definition
here
says
and
still
deploy
it
to
the
linker
D
namespace.
Please
comment
under
this
video
or
maybe
you
want
to
join
layer
5
the
community.
A
The
links
will
be
under
the
video
and
then
you
can
actually
open
issues
and
maybe
provide
for
requests
and
help
with
the
community
effort.
That
would
be
very,
very
nice
and
very
great
for
us,
and
for
you
too,
so,
let's
finally
deploy
link
early
to
our
cluster.
We're
deploying
latest
version
of
linker,
dear
master,
is
working
by
default
with
the
edge
release
of
linker
D,
which
is
the
newest,
the
most
risky
and
maybe
not
production-oriented
version
of
Winker
dear
once
we
hit
the
deployment
button.
A
A
Now
the
way
mystery
deploys,
link
or
D
in
the
background
is
actually
by
first
of
all,
downloading
the
edge
release
of
blinker
de
binary
and
then,
with
the
help
of
this
binary,
it
creates
all
the
deployment
llamo
files
that
it
needs
to
deploy
a
link
ready
into
our
cluster.
This
can
take.
Some
time
depends
on
the
bandwidth
of
your
internet
connection.
A
So
here
we
can
see
that
massery
is
already
creating
the
resources
that
need
to
be
created
in
order
for
a
link
day
to
start
working,
and
once
we
stopped
seeing
these
long
messages
coming
from
the
measure
alinker
the
adapter
and
that's
probably
now
we
can
go
back
to
measure
e
and
see
their
identification
say
here.
Saying
linker
D
has
been
deployed
successfully.
A
Is
a
new
namespace
called
linker
D
in
our
cluster
yeah
right
here,
and
we
will
see
what
the
faults
are
in
that
namespace
and
here
they
are
and
they're
already
running,
which
means
literally,
has
been
deployed
successfully.
Now
that
we're
having
to
be
running,
we
want
to
have
an
application
that
proxy
the
traffic
for
and
that
linker
they
can
help
us
analyze
and
control.
So
we'll
go
back
to
Mary.
Mary
allows
us
to
install
sample
applications.
It's
as
we
said,
a
service
match
playground,
so
we
can
go
and
deploy
the
modular
application.
A
One
of
the
sample
applications
that
is
created
by
the
linker,
the
team
now
you'll
notice
that
what
I
did
here
is
wrong
because
they
didn't
change
the
namespace,
so
the
application
is
now
deployed
to
the
linker
D
namespace.
Now
you
know
what
the
problem
with
that
is
well,
the
problem
is
that
whatever
is
deployed
to
linker,
the
namespace
doesn't
get
the
linker
D
proxy
injected.
Inside
of
the
paths
that
are
deployed,
let's
see
how
this
looks
you
see
all
the
mod
you
got.
The
pods
here
have
only
one
container
inside
them.
A
A
Here
are
all
our
pods
and
we'll
also
see
see
what
services
have
been
created
here.
Okay-
and
we
can
see
there-
our
Web
Services
right
here-
it
expects
to
have
a
load
balancer.
This
cluster
doesn't
have
load
balancers
enabled
so
instead,
we'll
just
port
forward
to
it
to
see
the
application
in
action.
So
we'll
do
cube
serial
port
forward
and.
B
A
A
A
There
is,
and
we
can
see
the
modular
namespace
and
here
are
all
the
services
of
the
modular
application,
and
we
have
all
of
these
stats
right
here
and
here
are
all
the
pods
of
the
modular
application,
and
we
can
see
this
link
really
comes
pre-installed
with
Ravana
and
we
can
easily
open
that
by
clicking
on
the
graph
on
an
icon.
So
everything
is
working.
Lincoln
is
working,
it's
monitoring
traffic.
A
A
Today,
anyway,
I
hope
this
helped
you
learn
about
measuring
and
about
literally
here's
graph
on
ax,
and
we
can
see
that
linker
D
is
sending
all
of
the
data
to
grow
fauna
and
we
have
graphs
for
all
of
the
things
that
has
been
happening
inside
the
application
and
that's
without
any
configuration.
As
we
said,
linker
D
works
with
minimum
configuration.
Thank
you
for
watching
and
I'm
looking
forward
to
your
questions,
bye-bye.