►
From YouTube: Certs magic with Saiyam - Episode3
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
A
A
A
B
Hello,
everyone
and
welcome
to
cloudnative.tv.
We
are
live
on
search
magic
with
sayam,
and
this
is
episode
number
three.
So
before
we
start
so
this
is
an
official
live
stream
of
cncf
and,
as
such
is
subject
to
cncf
code
of
conduct.
Please
do
not
add
anything
to
the
chat
or
questions
that
would
be
in
violation
of
the
of
that
code.
Basically,
please
be.
B
Of
all
your
fellow
participants
and
presenters
so
search
magic,
which
I
am
actually
search
magic,
is
a
show
where
we
talk
all
about
kubernetes
certification,
cka,
cka
dcks,
and
we
kind
of
go
through
some
of
the
concepts.
Then
do
some
of
the
problem
solving
and
see
how
things
work
overall
and
for
the
first
episode,
we
covered
the
introduction
to
the
certification,
so
why
these
certifications
are
important,
why
they
play
a
important
role
and
why
you
should
care
about
it.
B
So
I
think
that's
that's
pretty
important
and
we
covered
all
the
like
all
the
building
blocks
of
the
certification
and
the
preparation
material
like
where
you
should
be.
You
know
focusing
on
where
you
should
be
getting
the
learning
materials
from
and
all
the
stuff.
In
the
next
episode
we
had
the
we
had
temp
from
the
linux.
B
Training
team
itself
and
we
discussed
about
the
communities
architecture,
the
orchestration.
What
do
you
mean
by
orchestration
communities,
architecture
how
it
works?
What
are
the
building
blocks?
The
scheduler
controller
manager
atcd?
What
is
the
control
plane?
What
is
the
node
and
then
on
the
node?
You
have
the
cube
cdlq
proxy
cri
cni
csi.
So
all
these
things
we
covered
and
also
there
was
a
deep
time
on
demo
on
kubernetes
setup
using
cryo.
B
So
if
you
have
not
watched
all
these
are
available
on
youtube
as
well
as
on
twitch,
so
you
can
just
have
a
look
at
that
today.
Let's
continue
that
journey
and
we'll
be
diving
into
more
of
the
concepts.
Most
importantly,
the
keyword
objects
like
the
pods
deployments
replica
sets
how
they
work
some
of
the
scenarios
in
in
the
from
the
certification
point
of
view.
Obviously
we
are
focusing
on
that,
but
I
also
want
you
to
be.
B
You
know,
taking
care
of
like
how
you
can
use
it
in
your
daily
like
day
to
day
work
life
as
well,
so
we'll
cover
the
concepts,
but
we
cover
it
in
such
a
way
that
it
helps
you
both
organically.
In
your
work
environment,
as
well
plus
from
the
certificates
perspective
as
well.
A
B
A
B
In
the
meanwhile,
like
you,
can
you
know
just
tell
where
you
are
joining
in
from,
and
you
know
keep
sharing.
So
yes,
we
also
have
two
giveaways
for
the
session
that
I
do
so
make
the
stream
more
interactive.
Now,
since
I
will
be
presenting,
so
I
have
to
switch
back
back
and
forth
between
the
tab,
and
that
is
the
adjustment
that
I'm
doing
right
now,
so
we
have
to
switch
between
back
and
forth.
B
So
I
I'll
be
looking
at
the
comments
section
as
well
of
for
like
what
all
things
are
there
and
how
things
are
working
awesome.
So
I
have
my
equipment
ready
now
so
make
this
session
interactive
and
the
two
2
folks
who
have
the
most.
You
know
interactive
session.
We
will
be
doing
a
slack
giveaway
of
a
50
discount
voucher
for
your
what
you
call
certification.
A
B
Awesome
so
I
hope
you
are
able
to
see
the
screen
and
let's
cover
some
of
the
concepts,
so
I
see
a
lot
of
people
joining
in
welcome
everyone,
garish
sanskriti,
one
sheikh
yohan,
so
glad
to
have
all
of
you
over
here.
Please
share
this
on
twitter
as
well,
because
we
are
just
getting
started
and
we'll
cover
tons
of
you
know
good
material
today
to
be
honest
and
we'll
do
the
hands-on
as
well.
B
So
now
what
we
have
is
in
kubernetes,
we
have
kubernetes
objects,
so
we
have
your
parts
we
have
deployments,
we
have
replica
sets.
B
B
B
And
basically
tons
of
other
kubernetes
objects
as
well.
We
also
have
the
concept
of
crds
so
when,
when
you
have
the
controller
deployed
for
the
extent
for
extending
kubernetes,
so
you
can
create
the
custom
resource
object
as
well,
which
behaves.
B
Manner
like
you'll,
be
having
the
same
four
things.
Your
api
version
kind
spec
metadata
section,
but
these
are
the
basic
building
blocks,
building
blocks,
which
are
very
much
necessary
for
for
getting
started
at
least
so
the
first
or
the
smallest
unit
of
whatever
you
can
call
it
a
smallest
unit
or
basically,
where
your
application
actually
runs.
Your
application
runs
as
a
container
within
a
port,
so
a
pawn
will
be
the
smallest
unit.
So
if
you
know
from
the
previous
stream,
we
have
this
node,
which
is
joined
to
the
control
plane.
B
A
B
B
So
pod
basically
consists
of
one
or
more
containers:
c1
c2,
c3,
sharing
the
same
network,
the
name
space
and
sharing
that
same
resources
from
within.
B
Isolation
of
the
pod,
they
can
talk
to
each
other
on
localhost.
B
All
these
things
are
there
and
the
pod
is
obviously
the
part
gets
its
own
ip,
whichever
the
cider
range
that
you
have
given
by
setting
up
the
cluster
and
then
the
due
to
the
all
the
ip
table,
rules
that
set
up
the
support
to
part
note
communication,
those
words,
but
those
are
part
of
the
what
you
call
the
networking
section
so
today,
just
to
give
you
the
glimpse
we
will
be
covering.
B
It's
loading
yeah,
so
we
did
cover
the
architecture.
Installation
some
of
the
things
of
that
edc,
backup
restore,
will
cover
some
some
other
day,
but
at
least
we
know
how
to
set
up
a
basic.
You
know,
cube
adm
cluster
and
all
that
today,
we'll
focus
on
more
on
the
workload
and
the
scheduling
part
and
see
like
how
the
how
the
deployments
and
all
things
work,
how
the
scheduling
takes
place
and
all
those
things.
B
So
with
that,
let's,
let's
keep
continuing
and
we'll
go
to
the
section
of
parts.
B
B
So
why
I'm
telling
this
is
because
the
three
building
blocks,
the
four
building
blocks
to
be
honest,
are
in
any
of
the
yaml
file,
would
be
your
api
version,
basically
to
which
the
the
kind
the
object
belongs
to
so
there
there
is
v1
beta
1
v1
alpha.
So
there
are
alpha
features
beta
feature,
then
there's
two
stable
ones.
So
that's
how
the
api
version
would
go.
So
you
need
to
see
that
and
then
you
have
the
kind.
What
actually
you
want
to
create
or
what?
B
Actually
you
want
to
tell
the
cube
api
server
to
create
and
manage
on
its
own.
So
if
you
create
a
deployment-
and
you
say
like
I
and
you
submit
a
request
to
create
a
deployment
using
the
yaml
file
and
then
it
will
convert
that
into
json
object
and
then
pass
it
on
to
the
api
server
and
then
do
the
magic
that
it
does.
So,
even
if
you
are
doing
the
same
by
cube
cdl
it,
you
know
converts
that
to
json
and
then
passes
to
the
api.
B
So
and
then
it
does
all
the
magic
where
the
scheduler,
which
will
do
the
node,
based
on
the
request
and
the
limit,
we'll
also
talk
about
the
request,
source
request
and
resource
limit,
which
is
actually
very
critical
and
important
piece,
and
I
will
tell
you
exactly
from
where
you
have
to
learn
this,
because
the
documentation
for
that
is
really
really
solid.
B
And
after
that
you
have
the
metadata
where
you
define
the
name,
the
labels
and
all
those
things.
The
annotations,
the
specs
section,
is
really
the
actual
piece
where
you
define.
What
exactly
is
the
correct
characteristics
like?
What
exactly
is
the
image
that
you
need?
It
can
be
obviously
in
genex
or
it
can
be
your
custom
image.
So
this
is
the
application
that
you
would
want
to
run
so
it
can
be.
B
You
know
any
of
your
images
that
that
you
have
dockerized
or
containerized,
or
it's
just
the
oci
compliant
image
that
you
can
run
from
here.
A
B
Be
on
gcr,
it
can
be
on
github,
it
can
be
on
whatever
the
docker
ones.
So
you
have
specific
container
registry
that
you
have
set
up
for
that
and
then
the
name
and
all
those
components.
Obviously
there
are
a
lot
of
other
things
like
you
can
specify
a
node
not
like
this.
You
can
specify
the
node
name.
You
can
specify
the
node
selector.
You
can
specify
the
resource
request
and
the
resource
limit
for
this
particular
container,
so
there
are
tons
of
other
things
that
you
would
be
able
to
specify
as
well.
B
So
this
was
about
the
pod.
So
let's
go
into
the
next
version
because
usually
in
certificate
from
now,
let's
think
from
the
certification
perspective
now
from.
B
So
all
these
things
can
be
asked
when
you
are
in
the
exam,
a
very,
very
basic
kind
of
questions,
so.
A
B
Cool
now
you
are
able
to
see
the
terminal
window,
so
we
have
cube
serial,
get
nodes.
B
Awesome,
so
this
is
a
1.21
cluster
with
one
master,
two
nodes
and
cube
cdl
get
those
hyphen
o
wide.
So
basically,
hyphen
o
wide
is
the
command
that
you
can
get
additional
set
of
information
on
any
of
the
resources.
You
can
do
it
on
pods.
You
can
do
it
on
services.
It
gives
you
additional
fields
that
that
you
will
be
getting
we'll.
Also
talk
about
how
to
get
the
specific
values,
because
those
are
very
important
from
the
certification
point
of
view
as
well.
A
You
can
see
it
has
a
lot.
B
Of
things
like
the
container
runtimes-
and
it
has
container
d,
it
has
the
os
the
internal
ips,
the
kernel,
that
is,
the
the
os
image,
and
you
know
the
status
of
that.
So
all
these
things
are
there
when
you
are
defining
the
node
now,
when
you
are
asked
to
create
a
pod,
so
in
the
certification
exam,
the
easiest
way
and
the
fastest
way
is
to
create
via
the
qctl
command
line
itself,
because
it
saves
you
time
now.
I
have
told
it
multiple
times
in
the
certification
areas
as
well.
B
You
can
you
have
to
create.
I
mean
you
want
to
take
care
of
the
time,
because
the
time
runs
out
very
fast.
Sometimes
there
are.
There
are
few
questions,
because
all
questions
are
scenario
based.
So
there
can
be
a
long
kind
of
question,
but
it
will
be
step.
I
mean
you,
don't
have
to
be
stuck
in
a
particular
question
and
you
don't
look
at
the
clock
because
that
I
mean
it
happens
all
the
time,
so
it
has
happened
to
me
as
well
like
in
a
particular
question.
B
I
can
spend
a
big
chunk
of
my
time.
I
can
actually
waste
a
big
chunk
of
my
time,
which
I
could
have
used
in
solving
other
problems
and
come
back
later,
this
particular
question
and
solve
it.
So
don't
do
that
so
take
care
of
the
time,
and
if,
if
you
are
not
able
to,
you
know
see
like
what
is
happening,
then
there
are
different
tricks
that
that
you
know
you
can
keep
in
mind
like.
If
I
don't
know
anything
then
I
can
use,
I
can
do
step
a
step
b.
B
Step
c
have
a
plan
that
is
important.
Alliances
are
completely
your
choice.
I
have
never
used
aliases
in
any
of
my
certification
exams
and
I
have
cleared
all
of
them,
but
a
lot
of
people
uses
them
and
they
help
you
a
lot.
I'm
not
saying
it
doesn't
help
you,
but
they
are
again
totally
optional
and
completely
based
on
your
choice.
So
if
you
are
like,
if
you
think
you're
like
you
know,
kgp
is
way
faster
than
cuba.
City
you'll
get
parts
then
definitely
set
up,
realises
first
and
use
them.
B
It's
pretty
simple
from
that
perspective,
so
let's
do
cube
ctl
run.
So
let's
say
you
have
to
run
a
pod,
so
we'll
run
the
engine
export
and
we'll
give
a
image
so
hyphen
hyphen
image.
B
You
can
define
other
attributes
like
port
and
all
those
things
I'll
show
you
that
as
well.
So
you
can
see
the
pod
is
created.
Now,
if
you
do
not
define
anything
very
simple,
it
will
create
that
in
the
default
name
space
you
can
see
cube
ctl
get
ns.
You
have
different
set
of
name
spaces
already
present
over.
A
B
And
you
have
a
cube
serial
get
parts
would
have
been
created
in
the
default
name
space,
so
you
can
see
it's
already
in
container.
Creating
means
it
is
just
pulling
the
image
and
getting
ready
for
that.
B
B
As
part
of
the
deployments
that
I
would
have
already
deployed
onto
the
cluster
now
other
thing,
if
there
are
too
many
things
that
you
have
to
apply
on
to
I
mean
you
have
to
customize
like
a
multi-container
board
very
good
example.
B
Two
images
so
we'll
be
using
engine
x
and
then
hyphen
oeml,
then
hyphen
hyphen,
try
run.
B
B
So
basically,
you
immediately
get
the
sample
of
the
yahoo
file.
So
some
of
the
things
that
you
can
immediately
you
know
do
like
if
you
want
to
run
it
with
a
specific
label.
Definitely
do
that,
like
color
red,
something
like
that.
B
And
maybe
the
pod,
maybe
the
container
name
has
to
be
important
and
then
resources.
Obviously
you
have
to
set
some
resource
limits
and
maybe
you
need
to
specify
a
node
name
or
maybe
you
need
to
schedule
it
on
a
specific
node.
Actually
I
don't
know
any
name
of
the
specific
node,
so
qct
will
get
nodes.
So
maybe
you
want
to
schedule
that
on
node
3.,
so
let's
give
the
node
name
oops.
B
Let's
give
the
node
name
node
three.
So
basically
I'm
combining
a
lot
of
questions,
so
I'm
scheduling
it
on
a
specific
node.
So
that
comes
as
a
separate
that
can
come
as
a
separate
problem
like
scheduler
node,
on
a
separate
on
this
particular
node.
Now
in
practical,
this
can
be
very
helpful
where
you
have
your.
B
Which
are
targeted
for
specific
nodes,
and
you
want
to
you
know,
specify
that
on
you
want
to
run
that
on
a
particular
node.
It
can
happen
all
the
time
it
can
happen
to
different
types
of
workload.
It
can
happen
to
different
node
sizes.
So
basically
I
can
have
different
node
pools
in
in
my
kubernetes
cluster.
So
let's
say
I
have
different
sizes
of
virtual
machines
connected
or
if
I
take
example
of
c
or
cubics,
which
is
us-based,
I
create
a
cluster
with
different
load,
pools.
B
Node
pools
so
I
add
a
small
load
pool
first
and
of
of
three
nodes,
then
I
add
another
load
pool
of
maybe
larger
sizes
of
three
nodes,
so
I
can
actually
do
based
on
two
things.
One
can
be
the
node
selector.
So
let
me
show
you
that
I
think
I
have
to
share
my
complete
screen
else.
I
have
to
switch
back
and
forth
so
anyways.
What
I'll
do
is
I'll
open
that
road,
selector
stuff.
B
And
first
I'll
finish
this
and
then
I'll
move
to
the
theory
section
again
also
we
were
talking
about
a
multi-container
board.
So
let's
have
that
over
here
image,
maybe
I'll
meet
whatever.
A
A
B
B
This
can
happen.
So
that's
why,
if
you
want
to
use
the
shortcuts
use
them,
you'll
see
really
good
parts.
So
you
can
see
the
example
border
is
getting
created,
and
now
there
is
a
difference
in
numbers
over
here.
So
now
you
you
get
to
know
another
concepts
like
if
there
are
different,
multiple
containers
inside
a
pod.
It
will
be
showing,
as
here
so
one
out
of
one
means
one
out
of
one
container
from
that
pod
is
ready
or
running
now.
B
Zero
out
of
two
means:
zero
out
of
the
two
containers
defined
in
the
pod
specification
is
not
ready
or
not
running.
So
that's
why
it
is
in
container
creating
state.
So
there
are
obviously
different
port
states
container,
creating
then
you're
running.
Then
your
crash
loop
back
off
error
and
all
those
states
which
are
there
so
completed.
B
B
Pulled
it
created
the
container
that
it
is
and
started
that
and
above
also
you
can
see
both
of
the
containers
are
there.
So
you
have
one
container
with
container
id
this
another
one.
Engine
x
is
one
with
container
id
this.
So
you
can
see
you
have
a
multi-container
port
running
on
a
specific
node
which
is
demo
three,
so
we
actually
specified
the
node
name,
which
is
demo
three,
so
it
actually
ran
on
that
particular
node
itself.
B
B
All
things
we
did
on
the
recap
section,
so
we
created
a
pod,
we
created
upon
default
namespace.
We
created
a
pod
with
two
containers,
so
it's
a
multicontainer
pod.
We
assigned
the
port
to
a
specific
node.
B
We
created
a
pod
with
a
specific
label
and
what
else
yeah
that's
pretty
much
it
that's
pretty
much
it
that
we
did
till
now.
But
that
covers
a
lot
of
questions.
Actually,
if
you
are
doing
so,
you
can
create
your
own
cheat
sheet
and-
and
you
can,
you
know
practice
that
very
fast
like
how
to
create
a
multi-click
cut
in
a
part.
Do
a
dryer
and
get
the
yaml
file
edit.
That
and
do
it.
B
B
There
are
questions
on
cri
and
I
think
I
don't
know
whether
it's
then
cka.
I
can't
remember
you
call
that,
but
they
can
be
in
cks.
I
believe
so.
We
have
a
checkpoint,
it's
okay.
In
the
end,
the
the
cris
part
will
be
you
know,
or
to
do
for
pulling
the
image
and
running
all
the
stuff.
But
even
if
you
have
container
d,
the
the
commands
for
cube
cdl
creating
the
deployment
parts
replica
set
services
would
remain
the
same.
B
So
I
don't
think
there
will
be
any
change
and
what
else
are
there
yeah?
There
are
resource
limits
which
we'll
be
talking
about.
You
can
set
resource
quota.
Yes,
on
the
on
the
pause
and
the
name
name
spaces
as
well
yeah.
There
are
different
ways
like
you
can
set
the
aliases
yourself
and
then
you
know
you
can
do
all
sorts
of
stuff.
So
that's
completely
on
you.
B
I
have
never
said
that,
so
it's
it's
completely
optional
up
to
you.
What
is
the
difference
between
node
name
and
node
selector,
so
node
selector
will
basically
be
based
on
the
labels,
and
node
name
would
be
exactly
specifying
which
node
it
is
going
to
you
can
have
like,
like.
I.
B
Have
a
multiple
you
can
have
an
overpool
of
large
size
with
specific
labels,
so
you
can
specify
like
your
workload,
can
go
to
any
of
the
large
size
nodes.
Then
you
can
use
no
selector
easily,
and
what
else
is
there?
Can
you
explain
the
port
lifecycle?
I
think
I
explained
some
part
of
that.
You
have
different
like
how
how
the
pod
is
actually
running
and
can
two
containers
in
a
pod
xp
expose
on
the
same
port
port
obviously,
would
be
not
the
same
for
both
the
containers
because
they
are
like
local.
B
B
Where
we
are
stop
screen,
you
have
to
see
my
happy
face
again:
share
screen
window
assigning
parts
share
so,
as
you
can
see,
assigning
part
two
nodes,
this
is
basically
I
can
share
the
link
as
well.
If
you
like.
B
So
assigning
font
to
notes
very
simple:
you
can
use
either
node
selector
and
attach
the
label
to
the
node.
So
basically,
node
selectors
will
be
having
like
this
node
selector
disk
type
ssd
and
this
particular
label
would
be
on
the
node,
so
any
node
having
the
disk
type
ssd
would
be
the
one
that
will
be
that
the
scheduler
will
show
you.
This
particular
part
too,
and
then
you
have
like
affinity
and
t
affinity
is
something
which
is
not
something.
I
don't
remember.
Yeah.
A
B
Think
it's
part
of
the
certification
criteria,
but
yeah
theory
wise.
It's
very
you
know
it
can
be
a
useful
in
in
some
ways,
so
it's
basically
to
provide
a
node
selector.
Very
simply
you
can
do
so.
Anti-Affinity
and
affinity
is,
is
a
you
know,
expansion
of
the
type
of
constraints
that
you
can
have,
so
you
can
have
logical
operations
on
on
the
nodes
of
it
with
the
load
selector.
B
So
if
we
go
over
here
so
you
can
see
you
can
see
you
can
have
affinity,
then
you
can
have
node
affinity
now
there
are
different
things
like
required
during
scheduling
ignored
during
execution.
So
all
these
things
are
very
deeply
explained
in
in
this
particular.
You
know
talk
which
you
can
go
through.
We
will
not
cover
that
so,
inter
affinity
and
anti-affinity,
so
that
the
part
do
not
go
to
specific
nodes.
All
these
things
are
there
match
expressions,
so
you
can
do
that.
B
So
the
these
are
called
the
constraints,
and
so
these
are
the
practical
use
cases
that
some
of
them
have
given
the
documentation
states
very,
very
well.
B
There
are
courses
on
on
kubernetes,
which
I
have
already
told
you,
which
are
there,
that
you
can
study
from
the
paid
or
the
free
ones,
but
the
kubernetes
documentation
is
the
single
source
for
your
single
sort
of
truth
for
you,
because
the
kubernetes
documentation
is
superb,
it
explains
the
concept
in
depth
and
it
has
a
lot
of
great
examples.
Like
you
know,
you
have
a
complete
deployment
with
affinity,
pod
affinity
and
no
dfinity,
so
you
can
see
it's
support,
affinity
above.
Firstly,
I
think
no,
no
definitely
but
spot
affinity.
Only.
B
A
B
Yeah
this
one
should
be
the
node
affinity,
so
you
will
be
having
those
then
on
the
down.
You
will
be
having
another
thing
which
I
just
showed
the
demo
for
which
is
node
name,
so,
basically
the
simplest
form
of
the
node
selection.
Where
you,
you
are
not
setting
some
of
the
constraints
you're,
not
setting
labels,
you
exactly
know.
Okay,
there
is
a
node
xyz,
which
I
want
my
this
particular
workflow
to
schedule
to
then
you
can,
you
know,
should
you
do
that
particular
node
and
yeah?
B
Obviously
there
are
some
limitations
that
which
you
have
to
take
here.
Yep,
that's
pretty
much
it
with
respect
to
you
know
the
kind
of
scheduling
that
you
can
work.
So
yes,
this
comes
under
scheduling.
B
B
And
we
will
explain
it
in
the
form
of
you
know
kind
of
working
writing
as
well.
So
in
the
what
you
call
managing
resource
for
the
containers,
what
we
have
is.
B
A
A
A
B
I
should
be
visible
now
yep,
so
that
screen
should
be
visible
now,
so
basically
we
have
the
managing
resources
for
the
container.
So
when
you
specify
the
the
pod
the
pod
spec
in
that
you
can
optionally
specify
the
the
resources
that
are
required
for
the
container,
obviously
the
cpu
and
the
memory.
B
B
So
how
much
the
container
has
to
request.
So,
let's
say,
like
you,
specify
some
some
memory
in
in
the
request
that
a
container
needs.
Let's
say
one
gig
of
ram
or
one
gig
of
memory,
so
it
will
request
for
that.
B
But
if
you
do
not
specify
the
limit
now
you
have
to
understand
it:
resource
request
and
resource
limit
resource
request
and
resource
limit,
so
the
resource
request
and
the
resource
limit.
So
you
have
one
gb
one
gig
of
ram
that
you
have
specified
for
the
memory
now.
The
first
of
all
the
the
scheduler
will
schedule
it
on
the
node,
which
has
enough
capacity
so
obviously,
if
it
will
not
schedule
on
a
resource
where
the
memory
lift
is
less
than
one
gig,
so
it
won't
schedule
on
that.
B
So
it
will
schedule
on
the
on
the
node,
which
has
which
can
cater
this
particular
use
case.
So
that
will
be
there
now,
if
you
do
not
enforce
this
particular
part,
which
is
called
the
limit.
What
this
can
do
is
obviously
it
minimum.
It
will
require
1
gigs
and
if
this
particular
node,
where
it
is
getting
scheduled
to
is
say,
8
gigs,
then
it
can
consume
more
because
it
has
available
memory,
so
it
can
consume
more.
So
if,
if
something
is
not
there,
then
it
can
definitely
consume
more
so
in
this.
A
B
Over
here,
so
you
can
see
we
have
pod,
we
have
the
pod
spec
section
and
in
the
power
spec
section
we
had
containers-
and
I
showed
you
like.
There
was
a
section
called
inside
the
containers,
resources
and
the
resources
were
like
this.
When
we
created
the
yaml
file
now
in
this,
it
will
be
expanded
where
you
will
be
specifying
like
the
request
for
cpu
and
memory
and
we'll
be
specifying
the
limit
for
cpu
and
memory.
Now
we
can
force
enforce
the
limit
as
well
like
it
should
not
go
beyond
this
and
request.
B
It
can
request
for
this,
so
you
can
specify
in
the
spec
section
of
the
container.
So
that's
that's
how
you
will
be
able
to
specify
now,
let's
go
back
to
actual
resources
and
let's
see
one
example
over
here
yeah
now
here
you
can
see
a
pod
is
there
and
it
is
also
multi
container
pod,
and
you
can
see
that
this
particular
container
from
the
pod
is
in
the
resource,
then
requests
and
then
we
have
resource
limits,
so
resource
request,
resource
limit.
You
have
to
specify
both
and
yes,
you
have
to.
B
You-
can
specify
the
memory
and
cpu.
So
minimum
64
memory
and
the
cpu
to
50
and
the
memory
will
be
128
the
the
limits
and
the
cpu
is
500.
So
do
not
go
beyond
that.
So
that's
how
it
is.
You
know
getting
that,
and
also
it
is
like
saying
how
the
part
is
scheduled.
Obviously,
each
node
has
maximum
capacity
and
it
it
should
ensure
each
resource
sum
of
resource
request
for
the
scheduled
container
is
less
than
the
capacity
of
the
nodes.
B
B
B
So,
that's
why
the
documentation
is
your
single
source
of
truth.
It
is
explaining
very
clearly,
with
the
help
of
the
examples
as
well
how
you
can
use
that,
so
that
was
the
part
of
node
selector
and
all
that
one
last
thing
that
we'll
cover
today
is
is
the
deployments
part
which
is
very
important
and
we'll
also
cover
one
very
cool
again
like
we
covered
five
six
scenarios
in
single
example.
B
B
In
in
your
exam
certificate
exam,
you
have
what
you
call
kubernetes
documentation
allowed
so
make
sure
you
use
it
very
wisely
because,
sometimes
like
especially
for
the
cases
like
persistent
volume,
precision
volume
claims
where
you
cannot
create
them
directly
by
the
cube,
ctl
imperative
command
line.
We
you
can
directly
search
for
some
of
the
things
you
can
search
for
actually
a
lot
of
things
which
you
will
directly
get
a
pod
spec
from
here.
B
So
basically
deployments
is
a
declarative
way
for
ports
and
replica
assets.
You
can
describe
the
desired
state
of
the
deployment
and
in
in
previous
session,
we
talked
about
the
deployment
controller
that
takes
care
of
the
deployments
and
it
creates
the
replica
sets
and
make
sure
that
you
have
all
the
time
specified
number
of
replicas
running,
so
that
is
taken
care
by
the
deployment,
so
that
is
kind
of
the
power
of
kubernetes
that
you
are
trying
to
use.
You
are
trying
to
use
the
deployment
controller
that
ensuring
that
hey
kubernetes.
B
I
want
minimum
three
replicas
for
my
application
so
that,
if
any
of
my
like,
if
my
traffic
is
more,
I
can
pre-handle
that.
So
I
know
like
traffic
to
my
pod.
Traffic
to
my
application
will
be
more
so
I
can
have
more
number
of
replicas.
I
can
have
10
number
of
replicas
and
there
are
different
ways
of
you
know
there.
There
are
auto
scalers,
horizontal
part,
auto
scalers,
hpa
vpas
that
can
based
on
the
certain
metrics,
can
also
auto
scale
your
deployments.
B
B
So
so
you
have
the
deployment
you
will
create
a
replica
set
and
you
will
be
having
minimum
number
of
deployments
sorry
minimum
number
of
ports
as
defined
in
the
replicas,
which
will
be
running.
So
if
you
define
three,
it
will
be
running
three.
You
will
be
so
now
the
questions,
so
you
can
be
asked,
create
xyz
deployment
in
abc
namespace,
then,
with
a
certain
image.
B
B
You
see
the
status
of
the
deployment
you
can
see
the
status
you
can
record
kind
of
the
deployment
so
that
you
can
do
a
roll
back,
so
you
can
do
a
roll
back
as
well.
So
if,
if
the
image
that
you
have
specified
is
not
right,
so
you
can
immediately
roll
back
the
image
as
well.
That's
also
one
of
the
cases
that
I
mean
one
of
the
scenarios
that
can
be
asked
in
with
respect
to
the
exam,
and
these.
B
Of
the
cases
that
that
you
can
tie
with
your
day-to-day
life
as
well-
obviously
you
you
you
see
like
if
you
update
a
particular
image,
and
that
is
wrong,
or
that
is
not
working,
so
you
can
immediately
roll
back
to
the
previous
one
for
the
deployment,
so
that
can
be
there.
There
are
different
deployment
strategy,
rolling,
rolling,
update.
B
B
Okay,
let's
do
that
so
of
all
the
commands
are
actually
there
updating
a
deployment.
So
you
can,
you
know,
set
the
image
and
you
can
do
that.
We
will
be
using
the
same
all
the
same
things
we'll
be
using
exactly
same
scenarios
for
sending
the
image
and
doing
all
those
stuff.
So
let's
go
back
to
the.
A
B
B
B
B
So
we
already
have
cube
cdl
get
deployed
I'll,
just
delete
that
you'd
see
ds,
delete,
deploy,
okay,.
B
B
So
you
can
see
usage
cube,
cdl,
create
deployment,
give
the
deployment
name,
give
the
image
name
and
some
of
the
commands.
Now
these
some
of
the
commands
and
the
options
can
be
like
the
port,
the
replicas
and
the
output
like.
If
you
want
the
output
as
a
you
know,
on
the
whatever
file
it
is
and
then
you
can
have
the
try
run.
Obviously
you
can
have
the
dryer
and
client
and
get
the
the
copy
of
the
yaml
like
we
did.
So,
let's
do
that.
First
locate,
ctl
run
index.
B
B
Edit,
the
yaml
file,
as
per
the
question
instructions,
as
per
the
scenario
instruction,
you
can
quickly
get
this
yaml
file
store
that
you
know
just
do
this
in
a
dp
dot
yaml
and
we
can
do
the
dp
dot
yaml
edit.
So
if
we
can
do
the
bim
vi
of
that,
we
can
maybe
set
the
number
of
replicas
to
two
and
sorry
number
of
replicas
to
do.
We
can
change
the
you
know,
respect
the
container
name
to
infinix
2
or
something
like
that.
So
anything.
B
B
B
A
B
Is
there
so,
basically,
we
can
set
the
image
for
that
a
very
easy
command.
Is
there
a
cube
cdl
set
image
for
the
deployment
in
linux
and
from
engine
x2
nginx
1.18?
So
you
can
see.
B
We
can
describe
yeah,
you
can
see
the
deployment
controller,
you
know
scaled
up
and
scaled
down
because
it
is
now
you
know
doing
the
changing
the
image,
so
the
latest
image
is
getting
changed
to
what
you
call
1.18.
B
B
B
A
B
See
so
this
is
to
check
whether
the
deployment
is
successfully
rolled
out
becomes
very
handy
in
some
some
automation,
stuff.
B
We
keep
checking
the
roll
out
and
then
we
can
move
on
to
the
next
step.
Something
like
that.
So
if
I
go
back
and
do
get
caught
I'll
be
seeing
all
the
parts
are
in
running
state,
so
we
created
the
deployment
we
scaled,
that
we
change
the
image
yeah.
We
can
record
that
so,
let's
change
the
image
actually
to
1.19
and
hyphen
iphone
record,
and
now
I
describe
we
can
see
the
image
is
changed
to
1.19,
and
this
is
how
it
does
it.
B
It
is
being
done
and
now
we
can
actually
see
cube,
ctl
roll
out
his
free
deployment
in
the
next.
So
you
can
see
it
is
this
and
now,
let's
do.
B
B
B
As
well,
so
you
can
have
all
these
things
so
scaling
up
scaling
down
is,
is,
very
you
know,
very
common
and
when
implemented
with
personal,
auto
scaler,
you
can
have
that
and
obviously
yeah.
You
can
also
do
one
thing
which
is
called
qcdl
get
parts,
so
you
can
actually
cube
ctl
expose
and
if
I
just
click,
sorry
cubesale
exposed
pod
and
hyphen
hyphen
help,
so
you
can
actually
provide
hyphenation
help
and
it
will
give
you
some
of
the
commands,
which
are
very
handy.
B
So
you
can
see
this
create
a
service
for
a
valid
pod,
so
cube
cdl
exposed
pod
with
a
valid
name,
and
then
you
give
the
what
you
call
a
port
number
and
you
can
also
give
like,
with
the
name
front
end
or
just
like
without
any
name,
and
you
can
also
specify
the
type
you
can
specify
the
protocol.
So
let's
do
like
what
parts
we
have.
Let's
do
one
quickly
get
pause.
We
have
five
minutes
q
across
come
on.
B
B
Now
I
can
take
any
of
the
ips.
Obviously,
external,
like
is
right,
now
not
displayed,
because
I
have
not
done
that
particular
portion,
but.
B
Of
the
things
that
we
have
done
by
setting
up
the
cluster,
which
I'll
just
show
you
how
to
do
that
very
very,
very,
very
quickly
come
on
clear,
so
we
have
done
today,
parts
we
have
done
today.
Sorry,
we
have
done
today
parts.
B
B
Yep,
let
me
just
hear
back
quickly,
doctor
happy
face
share
back
and
we
share
this
particular
screen.
B
Yeah,
so
we
have
done
kubernetes
objects,
we
have
done
parts,
we
have
done
deployments
and
we
have
seen
like
creating
them,
editing
them
multi
container
boards,
the
rolling
them
out
and
scaling
them
status,
record
roll
back,
changing
the
image
and
editing
the
yaml
file.
What
are
the
resource
and
the
request
that
you
can
specify
in
the
resource
section
of
the
containers?
B
What
are
the
node
names
scheduling
them
on
a
specific
node
based
on
node
name,
node
selector.
I
told
you
the
theory
wise.
So
all
these
things
we
have
done
very
quickly
just
have
a
look
at
this
particular
gist,
so
I
created
this
particular
gist
to
set
up
kubernetes
on.
Basically,
these
are
the
three
nodes
which
are
there,
so
I
created
three
instances:
compute
instances
on
sigo,
so
I
asked
platform,
and
then
I
have
just
done
this
where
I
have
set.
B
Obviously
the
container
d,
because
I
told
you
it
is
container
d
based
and
then
the
swap
off
fsdb
fstab,
and
then
we
have
all
the
fancy
stuff
for
setting
up
cube,
let
q
adm
and
cube
cdl
holding
them
and
then
yeah.
This
is
not
required.
So
you
do
this.
After
that,
we
obviously
do
cube
adm
in
it
and
specify
the
cidr
range
and
boom.
After
that
we
do
all
the
steps
or
that
are
defined
in
the
output
of
the
cube
adm
command,
and
then
we
join
the
particular
join
the
respective
nodes.
B
So
that's
pretty
much
it.
Let
me
go
to
the
chat
and
I
think
we
have.
B
I
don't
think
it
was
echoing
yeah.
Let
me
go
back
to
this
in
the
exam
question
for
particular
node.
Should
we
go
for
node,
name
or
nt
affinity?
Exactly
depends
really
what
the
question
is,
and
obviously
the
weightage
as
well.
I
I
would
personally
like,
if
it's
very
straightforward,
that
this
has
to
be
scheduled
or
not
this.
Why
not
go
for
node
name,
it's
it's
very
simple
and
yeah
labels
can
be
there
to.
You
know
just
change
all
sorts
of
stuff.
Then
node
affinity
has
report
life
cycle
before
scheduling.
B
Okay,
there
are
two
rolling
something
in
place:
deployments:
recreate
rolling,
update,
okay,
there's
a
good
conversation
that
is
going
on
between
us,
muff
tab.
I
don't
know
the
name
and
girish
both
of
them
so
which
is
pretty
good
yeah.
We
okay,
I'll,
cover
max
search
and
max
unavailability
sometime
for
sure,
but
yeah.
I
think
it's
already
answer
that
and
you
are
asking
the
link
for
the
gist.
Let
me
first
see
if
it
is
public
one.
B
A
B
Where
I
work
so
you
can
sign
up
or
you
can
log
in
and
you'll
be
getting
the
free
credits
as
well.
Yep,
that's
pretty
much
it
if
you
want.
Actually
you
know
more
on
the
deep
dive
stuff
for
some
of
the
you
know
cloud
native
technologies,
then
obviously
I
have
my
own
youtube
channel
as
well,
so
you
can
go
to
cyanproduct.com
youtube
where
I
stream
about
the
cognitive
technologies,
which
obviously
relate
the
community
stuff.
A
lot
of
things
are
over
there.
B
If
you
go
to
deep
live
sessions,
this
particular
there
was
a
question
last
time
like
I,
I
should
cover
more
other
of
the
cloud
native
things,
but
this
particular
streaming
is
only
focused
to
search
magic.
Only
the
certification
stuff,
so
we'll
be
doing
cloud
native.
I.
B
On
my
channel,
you
can
subscribe
to
that
and
you
know
a
lot
of
that
stuff,
but
you
should
definitely
follow.
Cloudnative.Tv
awesome
shows.
Are
there
an
awesome
cloud
native
folks,
the
ambassadors,
the
founder
community?
They
have
come
together
to
put
down
a
lot
of
great
schedule
for
that
and
a
lot
of
great
shows
are
coming
up
and
there
are
also
sticker
packs
available
on
store.cncf.io
that
you
can
check
that
so
yeah
very,
very
good
question.
B
Yep,
so
this
is
the
webinar
and
it's
on
12th
of
july
22nd
of
july.
Sorry,
where
I'll
be
talking
all
about
cps
preparation
and
I'll,
be
sharing
a
lot
of
things
that
as
well.
I
can
share
the
link
of
that.
It's
okay.
B
Actually
follow
me
on
twitter
because
that's
where
I'll
tell
you
I'll
be
keep
on,
you
know
telling
about
all
the
certification
things
I'll
do
at
different
places,
obviously,
at
some
point
we'll
be
covering
cks
on
sirs
magic
as
well,
but
not
very
immediately,
because
we
are
trying
to
go
step
by
step
so
that
people
who
want.
I
want
people,
you
know
to
get
the
knowledge
for
for
kind
of
free
and
get
them.
You
know
the
right
set
of
track.
B
So
if
you
just
want
that
sort
of
thing,
you
can
just
watch
episode.
One
episode
two.
A
B
Maybe
services
may
be
networking,
I
don't
know,
I
haven't
decided
that,
but
we
will
choose
some
of
the
topics
and
we'll
discuss
on
that.
So
that's
how
it
goes,
we'll
try
to
keep
them
independent.
So
previous
topic
was
pre,
absolutely
independent.
If
you,
if
you
have
the
community
set
up,
you
can
skip
that.
If
you
know
about
parts
and
deployments
you
can
skip
this
one
so
I'll
try.
I
try
to
make
those
independent
and
we'll
continue
like
that
and
for
the
winners.
B
Let's
see
who
were
the
most
active
in
the
chats,
so
I
think
girish
has
been
active
since
very
beginning,
so
kirish
is
one
of
the
winners
for
the
swag,
so
kirish,
please
do
reach
out
to
me
on
github.
Oh
sorry,
on
twitter,
so
that
I
can
give
you
the
coupon
and
let
me
know
the
let
me
see
the
second
one.
B
I
think
us.
B
1995
has
answered
the
questions
really
well
and
kept
it
more
interactive.
So,
first
of
all,
thank
you
for
answering
and
yeah
usma.
So
thank
you
for
answering
and
you
are
the
second
winner.
So
congratulations
to
kirish
and
usma,
please
reach
out
to
me
on
twitter.
My
dms
are
open
I'll.
Let
you
I'll
give
you
the
coupon
code,
50
off
google
code
on
certifications
and
hope
that
you
know
you
get
certified
soon
and
hope
you
like
this
session.
Please
share
that,
so
that
people
can
join
again.
B
So
this
is
a
bi-weekly
on
thursdays.
8
30,
pm
ist
or
8
am
pt.
That
will
be
continuing.
Thank
you
so
much
for
tuning
in
folks.
I
really
liked
interacting
with
you
all
and
explaining
some
of
the
community
stuff
that
I
like
in
simple
terms.
I
hope
you
liked
it.
Please
do.
Let
me
know
how
things
are
going
and
what
else
you
would
like
to
see.
You
can
just
tag
me
on
twitter
and
tell
me
and
yeah.
B
Please
let
me
know
you
know
the
good
things
as
well
like
what
you
really
liked
about
the
session.
Thank
you
all
and
please
follow
cloud
native
tv
and
enjoy
the
awesome.
Other
shows.