►
From YouTube: Noobernetes 101: Top 10 Questions We Get From New K8s Users - Neependra Khare & Karthik Gaekwad
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
Noobernetes 101: Top 10 Questions We Get From New K8s Users - Neependra Khare, CloudYuga Technologies & Karthik Gaekwad, Oracle
Neependra and Karthik have Kubernetes 101 courses on edX and lynda.com that are very popular and have had over 200,000 views. Over time, they’ve received many questions from folks who are new to K8s, after taking the courses, or in training. They want to spend some time elaborating on the most common questions we’ve received, and answer them once in for all!
They’ll cover the following faq’s they get over and over again:
- What kind of services should I use for my applications?
- Shouldn’t I be using Istio instead of K8s now? What does K8s buy me when I can use istio?
- How can we do capacity planning in K8s?
- Why there is a high learning curve in K8s? Isn’t K8s too complicated?
- What is the best way to set a development environment with K8s?
- and many more
You’ll leave with a much better understanding of K8s, the ecosystem, and have the knowledge to take on the cloud native world.
About Karthik Gaekwad
Karthik Gaekwad is a veteran engineer who enjoys building software products from scratch using cloud and container technologies. He has worked in both large enterprises and startups with his career spanning National Instruments, Mentor Graphics, was the first hired engineer at Signal Sciences, an early engineer at StackEngine (which was acquired by Oracle). Karthik works at Oracle as a Principal Engineer to build products in the cloud native space.
He has programmed in many languages including Java, C, C#, Python, and Ruby. He first learned Golang in 2012, and it had been his language of choice since. He has written many production-level applications using Go and appreciates its simplicity and flexibility.
Karthik graduated from the University of Arizona in 2007 with an MS in Computer Engineering and currently lives in Austin, Texas with his family. He organizes several conferences including Devopsdays and Container Days and has chaired the devops tracks for the Agile Conference, and All day devops. He is also an accomplished author for LinkedIn Learning and lynda.com. In his free time, Karthik enjoys spending time with family, keeping up with the latest trends in software, dabbling in new product ideas, and blogging on theagileadmin.com
About Neependra Khare
Neependra Khare is the Founder and Principal Consultant at CloudYuga Technologies, which offers training and consulting services around container technologies such as Docker and Kubernetes. He is a Certified Kubernetes Administrator (CKA).
Neependra authored “Introduction to Kubernetes” course on Edx for CNCF. He also author a a book on Docker, “Docker Cookbook” in 2015 and co-authored a course on “Cloud Infrastructure Technologies” on edX for The Linux Foundation. He ran Docker meetup group in Bangalore, India for more than 5 years and very active in other meetup like Kubernetes, AWS, DevOps etc.
Before starting CloudYuga, he was Principal Software Engineer at Red Hat and worked on projects like Gluster, OpenShift. As IT professional he worked as System Administrator, Support Engineer, Kernel and Filesystem Developer and Performance Engineer.
B
Back
everyone:
my
name
is
Karthik
and
I'm
Lee
paid,
so
they
will
talk
about
super
LEDs
101.
We
have
able
to
cover
top
10
questions
we
get
while
teaching
kubernetes
courses.
I
am
founder
of
cloud
yoga
and
you
provide
consulting
training
around
containers,
kubernetes
I'm,
also
a
CNC
ambassador
and
CK
the
course
which
then
talked
about
on
ethics,
which
has
more
than
63,000
people.
A
B
A
Banner
just
said
that
there's
a
quiz
after
this-
it's
not
a
test,
but
it
you
can
consider
it
as
a
test.
If
you
want
to
win
the
cool
stuff,
my
name
is
Karthik
I
work
at
work.
Oh
and
before
everybody
comes
in
asking
about
BBN
questions
will
forewarn
you
that
I
know
very
little
about
databases.
I
know
some
stuff,
but
there's
a
lot
more
people
that
know
about
databases
at
Oracle,
I,
actually
work
on
the
cloud
side
of
things,
so
I
work
on
a
team
called
Cloud,
Reader,
glads,
Cloud,
Reader,
Oracle
comm.
A
We
have
a
ton
of
content,
therefore,
to
help
people
get
started
with
cuba,
Nettie's
and
the
surrounding
projects
in
the
CN
CF
that
have
to
do
with
cloud
native
stuff,
I'm
from
Austin,
Texas
and
I.
Organize
a
couple
of
conferences
in
Austin
over
there
and
I've
been
in
industry
for
for
a
while.
So
enough
about
us,
quick
question
for
the
audience
show
of
hands
of
people
like
how
many
of
y'all
are
developers
most
folks
about
operations.
A
Ok,
good
everybody
on
the
top
seems
like
they're,
either
DevOps
or
developers,
which
is
good
because
that's
who
this
talk
is
for
and
with
respect
to
Cuba
Nettie's.
How
many
folks
are
new
to
kubernetes
show
of
hands
a
few
new
folks,
and
you
guys
you
do
have
experience
with
docker
containers
and
show
of
hands
as
well
good
and
then
would
you
classify
yourself
as
like
a
new.
B
A
So,
first
time
kubernetes
I,
don't
know
much
okay,
good
good
number
of
people.
So
that's
great
because,
honestly,
it's
a
big
platform.
So
there's
a
lot
of
stuff
to
kind
of
cover.
So
today's
talk
we're
gonna.
Basically
talk
about.
You
know
all
of
the
most
common
questions
that
any
pandro
and
I
have
gotten
over
time,
so
I
actually
authored
the
LinkedIn
on
lynda.com
courses
for
learning
communities
and
a
bunch
of
other
cnc
of
stuff.
B
So
but
we
will
start
talking
about
containers,
we
generally
say
that
containers
are
lightweight
px.
It
is
good
for
analogy,
but
it
does
not
true
at
all.
So
I
want
to
get
with
VMs.
Is
you
have
a
hypervisor
in
between
the
hypervisor?
You
would
deploy
a
guest
OS
and
then
then
you
put
a
application
on
top
of
that,
but
remember
the
end
goal
is
to
run
the
application
not
to
run
the
wheels
so
as
I
can
run,
the
applications
on
my
environment
and
I
can
save
cost.
B
That's
going
to
be
a
benefit
for
myself
in
my
company
right.
So
then,
so
in
the
container.
What
you
kind
of
run,
everything
on
one
forced
to
us,
which
means
are
sharing
your
host
OS
there
with
all
the
containers
which
gives
you
more
density,
so
you
can
kind
of
save
coughs
there.
You
can
have
more
density
of
the
containers
of
the
applications
there.
B
Once
you
made
a
Penta
nice
application,
you
can
deploy
anywhere
on
environment
feet
on
the
ana
workstation
or
on
the
cloud
it
works
as
it
is,
and
it
becomes
very
easier
to
manage
because
you
have
memorize
it.
You
might
have
many
fest
files
of
kubernetes
or
docker,
and
it
just
one
one
come
on
in
that
beginning
get
deployed
same
everywhere,
and
so
let's
talk
about
another
kind
of
thing.
B
What
we
are
seeing
right
now
is
light
weight
Williams,
where
you
could
now
have
a
VM
one
VM
part
container,
so
this
light
weight
causes
it's
coming
right
now,
like
hyper-v
containers,
contract
containers,
AWS
firecracker,
where
you
deploy
a
one
VM
and
one
container
together.
That
way,
you
get
a
security
off
the
view,
but
the
benefits
of
containers.
B
B
It
would
be
bit
hard
to
expose
the
applications
excellent
work,
but
you
can
definitely
do
it.
Similarly,
for
that
storage
there
are
software
using
which
you
can
skim
the
software,
even
if
you're
on
trend,
but
of
course,
on
the
cloud
you
would
you
don't
have
to
care
about
the
control
plane
of
the
state.
Are
there
the
class
would
have
managed
for
you.
Then
there
are
things
like
load,
balancer
ingress
or
to
the
scaling
that
would
all
be
managed
by
the
cloud
provider
for
you.
B
A
So
the
next
part
that
we
kind
of
come
to
is
you
know,
self-install
cuban
NES
and
managed
to
admit
these
defender
touched
a
little
bit
on
this.
But
let's
take
a
look
at
this
is
a
little
bit
more
detail
when
Cuban
Indies
first
kind
of
came
about
you
know
you
had
to
actually
spend
a
lot
of
time
to
install
communities,
get
it
up
and
running,
etc.
So
you
know
the
canonical
examples
of
this.
A
A
Cluster
goes
down,
that's
something
that
you
know
you
are
kind
of
taken
care
of
and
like
some
of
the
typical
use
cases
for
this
are
you
know
when
folks
are
installing
this
on
premise,
you
know
typically,
you'll
have
a
team
that
installs
your
cluster
does
the
operation
stuff
for
the
cluster
and
make
sure
everything
is
up
and
running
or,
and
then
the
other
use
cases.
You
know
if
you
have
something
very
specific,
that
you're
trying
to
do.
A
You
have
like
custom
cluster
requirements,
that's
where
we
see
a
lot
of
folks
using
the
the
self-install
cuban
Nettie's
version,
I
like
to
so
I
grew
up
in
the
80s
and
I
give
this
analogy
of
you
are
like
he-man
or
you
are
like
Shira,
and
you
have
all
the
power
to
do
whatever
you
want
with
the
cluster.
Whether
it's
know,
if
you
want
to
change
things,
you
have
the
ability
to
do
that.
A
So
with
the
manage
side,
if
you
take,
if
we
take
a
look
at
that,
you
know,
I
have
this
little
diagram
here.
That
says
you
know
customer
manage,
which
is
what
we
did
initially,
on
the
manage
cluster
side,
you
are
kind
of
responsible
for
the
applications.
So
me
Pedro
mentioned
a
little
while
earlier,
where
you
know
we
have
this
idea
of
a
control
plane
and
the
cloud
provider
kind
of
man
is
the
control
plane,
which
kind
of
helps
you
with
the
scaling
higher
ability.
You
know
backup,
recovery
and
all
of
those
kind
of
things.
A
So
you
kind
of
offloading
some
of
that
work
to
to
your
cloud
provider.
Whoever
that
is
the
benefits
of
this.
Is
you
know
you
don't
have
to
spend
a
lot
of
time.
Creating
your
cluster
can
just
you
know,
click
a
button
or
maybe
run
run
a
script
or
something
like
that
to
get
your
cluster
up
and
running.
It's
lower
risks
because
you
don't
have
to
like
tweet
a
lot
of
stuff
up
front
and
you're.
You
know
it's
business
value
because
you're
trying
to
innovate
a
lot
faster.
A
So
my
analogy
for
this
is
you
know
it's
like
Voltron,
how
many
people?
Well
maybe
this
shows
my
age
because
I
grew
up
in
the
80s,
but
I
kind
of
see
this
as
Voltron,
where
it's
like.
You
know,
everybody
plays
their
plays
their
little
part.
But
then,
when
you
try
and
do
everything
together,
you
form
you
form
the
you
form
Voltron.
So
it's
all
about,
like
teamwork,
between
different
teams
in
your
organization
and
your
cloud
provider
to
get
your
platforms
up
and
running.
A
So
one
question
that
we
get
a
lot
is
like:
okay,
there's
many
ways:
do
this
solve
the
same
problem
like
help
me
choose
one
way
or
another.
So
if
you're
looking
at
the
managed
services
like
managed
cubed
many
service
providers,
some
of
the
pros
are,
you
know,
you're
offloading
a
lot
of
the
work
to
the
cloud
provider.
A
So
a
couple
of
things
to
watch
out
for
and
so
I
kind
of
wrote
this
recipe
where,
if
you're,
basically,
if
you're
new
to
communities
and
if
you're
using
cloud
you
know,
start
off
using
a
managed
Cooper
Nettie's
provider,
whether
it's
like
on
Amazon
or
if
it's
not
either
or
cloud
whatever,
like
everybody,
has
a
managed
service.
So
start
there.
A
So,
whenever
I
ask
people
like
what
do
they
mean
by
cloud
native
and
everybody
kind
of
has
like
a
different
definition
of
this
overtime,
so
some
people
say
like
oh
yeah,
you
know
you
deploy
like
microservices
with
kubernetes
continuously
using
a
CI
CD
pipeline,
and
almost
every
word
here,
except
the
words
using
an
is
a
buzz
word.
So
it's
a
little
confusing.
So
the
question
that
we
get
is
do
I
need
to
break
up
all
these
things.
A
Do
all
these
things
at
the
same
time
or
can
I
kind
of
break
yourself,
and
my
response
is
well.
It's
like
blind
men
and
the
elephant
right,
depending
on
what
kind
of
part
you're
touching
in
the
elephant.
It's
like!
Oh
yeah,
you
know,
like
somebody
might
say.
Oh,
this
is
cloud
native,
somebody
might
say
no!
No!
This
is
just
like
CI
CD,
somebody
might
say.
Oh,
this
is
just
a
container
so
like.
How
do
you
kind
of
take
the
next
step?
A
So
basic
principle
here
is
that
cuban
at
ease
is
a
platform,
and
humanities
plays
really
nicely
with
a
lot
of
different
things.
So
you
can
kind
of
orchestrate
it
to
do
a
lot
of
different
things.
Your
job
as
developer
operations-
or
you
know,
platform,
engineer
or
whatever
is
to
decide
them
the
priority
of
what
you
want
to
do.
A
Think
about
having
a
CI
to
build
your
container
images,
because
a
lot
of
the
stuff
is
based
on
containers.
If
you
have
a
pipeline
to
build
your
images
when
with
containers
like,
do
that,
also
use
the
same
CI
pipeline
to
be
able
to
push
those
images
to
a
registry,
and
then
you
know
once
you're
once
you've
kind
of
done
with
that
have
a
deployment
strategy
set
up
for
cuban
edits,
so
I'll
turn
it
back
to
the
painter.
B
So
there
are
different
ways
where
you
can.
You
can
use
the
different
applications
and
communities
cluster
when
as
simple
you
can
just
have
your
Yamada
JSON
file.
You
can
put
them
in
your
source
in
a
source
code
among
the
source
code
and
then
deploy,
and
there
are
tools
like
draft
and
scaffold
they
can
handle
what
the
changes
in
your
source
code
and
apply
the
applications
for
you
for
that.
Your
dev
cycle
becomes
very
easy.
So
that's
where
drafting
scaffold
can
hold
a
kind
of
benefit.
B
You,
where
you
have
a
source
code
and
then
just
point
those
tools
to
that
source
code
repository
and
it
will
deploy
the
application
for
your
development
and
debugging.
Then
there
are
tools
like
help
case
on
it,
customize
which
can
kind
of
convert
event
into
a
Amala
file,
but
they
were
telling
you
to
do
it
in
a
way
where
you
parameterize
your
parameters.
For
example,
you
can
say
I
want
to
parameterize
my
image
name,
my
replica
count
and
so
on.
So
you
get
powerful,
parameterize
them
and
then
those
tool
would
eventually
deploy
the
applications.
B
B
It'll
come
on
on
the
production
I
want
to
do.
Is
you
kind
of
have
a
git
repository
where
your
ops
team
do
like
check-in
off
the
food,
which
is
the
come
on
example?
If
I
want
to
change
that
account
of
the
image
of
my
application,
I
would
go
and
make
a
check
in
to
my
source
code,
and
there
will
be
rules
which
supports.
B
Going
to
watch
those
guitars
get
changes
and
deploy
the
applications
for
you
in
production.
So
with
the
great
office
you
can't
forget
the
history
or
who
made
what
changes
in
my
deployment
or
in
my
production
and
so
on.
So
it's
a
really
good
thing
to
do
so.
Once
you
have
deploy
the
applications
in
your
luster,
you
would
want
to
expose
them
with
excellent
work.
I
got
different
ways
which
you
can
use
to
expose
those
applications
which
you
have
deployed.
B
So
here's
a
quick
intro
about
kubernetes
architecture
here,
so
we
have
the
Navy
server,
which
can
there's
other
controls
up
there
with
other
components,
and
then
we
have
a
Q
proxy
that
you
proxy
once
on
each
node,
everything
in
the
master
as
well.
The
Q
proxy
allows
the
external
acts
of
the
application.
So
when
you
hit
any
URL
which
is
getting
served
by
the
kubernetes
cluster,
you
are
eventually
going
to
hit
the
Q
proxy
on
any
other
node
and
which
would
kind
of
take
you
to
the
inside
cluster
and
expose
the
applications.
B
B
Then
what
we
can
do
next
one
is,
you
can
expose
services
in
the
type
would
put,
though
I
think
is
not
usable
there,
but
very
clear
defined.
But
you
see
there,
the
the
blue
one
on
the
top
of
each
load
is
called
mode
would
so
we
are
exposing
our
services
by
the
type
word
put
so
that
note
word
is
like
we
are
blocking
a
port
on
each
node
and
whenever
you
hit
the
node
port
on
any
of
the
node,
the
request
goes
to
the
service
and
from
service
with
the
Fox.
B
So
if
you
look
at
here,
we
don't
have
any
part
running
for
the
blue
application
on
the
node
tree,
but
as
we
hit
the
note
3
our
traffic
routes
from
the
node
for
the
service
and
then
to
the
pots,
but
you
can't
release
the
node,
both
legs
or
the
public
services.
I
mean
you
can
use
it,
but
you
not
prefer
it
because
if
you
give
a
customer,
an
IP
in
the
port
number
combination
is
going
to
fail.
I
mean
customer
would
say
why.
Why
are
you
giving
me
this
thing?
B
B
Weight
is
being
built,
DNS
cannot
take
the
combination
of
IP
and
the
port,
it
can
just
say
the
combination
of
name,
and
it
can
happen
to
a
complete
IP
address
there.
So
for
that
purpose
we
have
a
service
type
where
we
expose
services
with
load,
balancer
and
undermine
cloud
provider
could
help
you
to
create
a
load
balancer
IP
or
a
DNS
name.
There
is
a.
A
B
With
the
type
of
balancer
I'm
going
to
do
create
a
node
balances,
which
is
not
would
be
very
scalable
for
me
and
has
been
expensive.
So
what
I'll
do
next
is
I'm
going
to
deploy
some
the
colors
in
this
controller
and
I'll?
Definitely
in
this
controller,
as
one
type
word
balancer.
So
instead
of
having
ten
word
puzzle,
tech
services,
I
would
have
just
one
load
balancer
for
my
ingress
and
that
interests
with
then
map
my
name's
to
the
services.
B
Example,
you
can
see
here
in
two
examples
for
this
channel
so
where
we
see
an
example,
dot-com
class,
who
are
the
green
from
they
are
able
to
do
all
the
green
service.
Other
one
I
have
the
worst
worst
example
where
I
say
blue
run,
example.com
or
the
other
whatever
the
color
is
there
with
that,
you
can
kind
of
come
to
the
right
service
there.
So.
B
Thing
what
we
talked
about
is
like
in
this,
how
we
deploy
it.
So
you
have
an
assembler
part,
which
kind
of
takes
care
of
the
backend
stuff.
So
I'll
just
leave
it
here.
You
can
get
a
slice
later
for
kind
of
getting.
What
is
there
so
the
other
ways
also
by
which
it
can
access
the
applications
like
alb
from
Amazon,
then
we
have
Google
load,
balancer
and
so
on.
They
can
also
help
you
do
kind
of
allow
the
exposing
of
reification
from
the
clusters
of
external
world.
A
So
just
kind
of
move
on
you
know
we
talked
about
cuteness
itself,
but
one
of
the
interesting
sub
segments
is,
you
know
it's
not
just
communities
today,
there's
also
concepts
of
cyrillus
and
concepts
of
service
mesh.
So
let's
take
a
look
at
this
a
little
bit
more.
One
of
the
questions
we
get
is
like,
should
I
be
using
kubernetes
or
should
I
use
the
server
list
or
should
I
be
using
Castillo
instead
of
kubernetes
instead,
so
some
general,
like
knowledge
on
this,
is
speaking.
A
Is
you
know
everybody
in
software
if
you've
been
around
for
a
while,
they
say
well,
it
depends
on
your
use
case,
and
so
we're
always
kind
of
you
know,
keep
saying
that.
But
the
real
answer
is,
you
know,
take
a
look
at
the
problem
that
you're
trying
to
solve
and
then
see,
if
humanity's,
actually
solves
that
problem
for
your
team
or
for
your
organization,
and
so
you
know
and
the
other.
The
other
thing
that
kind
of
happened
is
this:
Cuban
net
is
actually
a
platform.
A
So
there's
a
you
know,
you
know
you
need
you
need
kind
of
a
team
to
be
able
to
manage
that
a
platform
either
if
you're
doing
an
unmanaged
amends
to
Vietnamese
I,
your
self
excetera.
So
as
you
solve
a
question,
if
you're
like
just
one
developer
and
if
you're
in
a
start-up
or
something
like
that
and
don't
have
like
a
large
team
like,
are
you
ready
to
manage
a
brand
new
platform
because
there's
some
work
associated
with
that
and
also
you
know,
concern
considered
kind
of
your
use
case
behind
it
with
the
respective
service
match.
A
So
if
you
don't
know
how
to
like
manage
super
Nettie's
in
a
production
setting,
you
might
be
setting
yourself
up
failure,
it's
also
new
and
evolving,
which
is
great
because
there's
a
lot
of
new
things
coming
into
the
ecosystem.
From
that
perspective,
so
my
advice
would
be
don't
over
engineer,
because
chances
are
know
right
now.
Some
of
the
features
you
might
have
you
might
want
might
not
either
but
they'll
probably
be
coming.
You
know
somebody
winter
bragging
a
pull
request
for
that.
It's
the
don't
over
engineer
with
respect
to
serverless.
A
B
A
A
So
next,
next
up
so
we
talked
about
you,
know,
Liz
had
a
great
talk
on
all
the
different
things
that
we
do
between
our
back
and
authorization
etc.
So
questions
that
come
up.
It's
like
how
do
you
have
internet
ease?
How
do
they
authenticate
and
like?
How
do
you
kind
of
scale
kubernetes
routines?
I
won't
spend
too
much
time
on
here
on
authentication
other
than
there's
a
bunch
of
different
ways
that
you
can
authenticate.
A
Do
don't
not
do
anything,
it's
like
double
negative,
but
I,
don't
know
if
we're
better
way
to
say
it,
because
if
you
don't
have
an
authentication
strategy,
you
cannot
do
anything
with
restricted
authorization
and
that's
a
bad
thing,
because
you
really
a
lot
of
applications
today,
depending
on
Cooper
names
authorization.
So
you
want
to
have
like
an
authorization
strategy.
One
tip
on
the
authorization
part
is
everybody
today
in
the
community,
for
the
most
part,
users
are
back
for
authorization,
so
that's
kind
of
an
account
there.
A
A
B
A
To
this,
one
is
name
space
and
the
other
separate
clusters,
and
so
with
name
spaces.
You
know
you
kind
of
you.
It's
easy
to
use,
name
places
because
you're
still
running
in
a
single
cluster
and
then
name
space
helps
you
kind
of
separate
out
your.
You
know
your
work
for
your
team
versus
another
two
et
cetera
you.
If
you
decide
to
use
separate
clusters
for
let's
see
or
dab
and
integration
production
or
even
separate
your
clusters
by
Kings.
A
It
kind
of
helps
you
with
separation
of
concerns
so
like,
if
you're
trying
to
do
like
the
finance
team
can
have
their
own
cluster
versus
you
know,
the
e-commerce
team
can
end
up
having
their
their
own
clusters
and
a
lot
more
people
are
creating
this
as
cattle
and
not
pets.
So
especially,
we've
seen
this
a
lot
of
the
manage
side
where
people
will
just
kind
of
shoot
a
cluster
and
they'll
be
like
it
might
be.
A
A
A
Important
things
we
covered,
but
this
is
probably
like
my
most
important
thing
in
this
in
this
talk-
is
there's
credentials
and
secrets
and
how
to
manage
disease
and
I
feel
like
more
people
get
this
wrong
than
right.
If
you're
new
to
communities,
please
take
a
look
at
too
many
secrets,
because
Kuban
any
secrets
is
the
way
to
store
all
your
sensitive
data.
So
if
you
have
database
passwords
yeah
like
something
very
sensitive,
you
want
to
use
a
juvinity
secret
once
you're
kind
of
like
already
falling
in
that
best
practice.
A
You
can
take
a
look
at
it.
You
know
this
their
secrets
and
commissions
providers
secrets
are
actually
stored
inside
of
communities
with
a
dirty
coded
would
be
64.
So
it's
not
really
encrypted
it's
more
encoded,
but
you
can
actually
encrypt
it
using
like
an
encryption
provider.
So
you
know
those
are
a
couple
of
couple
ways
to
you
know
having
that,
but
for
what
it's
worth
like
make
sure
you,
you
actually
use
that
you've
been
in
it
secrets.
B
So
we
talked
about
the
publication,
exposing
it
and
then
putting
it
our
secrets
there
I
will
talk
about
the
last
file.
Yet
how
do
we
store
data
into
opportunities
because
eventually
want
to
put
some
data
in
your
application?
Man
is
that
the
data
from
there
right,
so
we
have
a
concept
of
pod
and
commodities
and
a
pod
can
have
multiple
of
containers.
We
assign
a
volume
or
to
a
pot
and
then
each
container
in
the
pot
in
decide
where
they
would
like
to
mount
that
specific
volume.
B
There
are
different
drive
modes
which
you
can
use
for
export
or
that
amount
of
the
volume
like
empty
like
a
scratch
track,
so
amputees
like
a
scratch
instead
of
those
as
a
part
time,
so
actually
no
real.
What
use
kiss,
but
it's
good
for
doing
the
test
in
there
and
we
are
host
part
where
you
can
share
the
host
directly
to
the
container
of
the
part
there
and
NFS
and
other
providers
volumes
like
EBS
and
so
on,
which
you
can
assign
to
your
pots
and
that
amount
it.
B
But
typically,
what
we
use
is
something
goddess
persistent
won't
claim.
So
what
we
need
is
a
way
where
developer
can
say:
I
want
a
volume
of
excise.
Now
the
developer
would
have
to
care
about
whether
there
is
suffering
from
AWS
as
well
or
whatever
it
is
so.
The
developer
says:
I
want
a
war,
we
work
of
a
specific
size
and
for
that
the
neuropathy
is
from
the
Connors
persistent
one
who
claimed
that
finishes
for
a
PVC.
So
you
would
say:
I
want
a
size
of
5
GB.
B
Maybe
some
Hartford
gone
and
it
come
from
two
ways:
one
is
from
the
static
one
who
provisioning
in
which
the
cluster
manager
or
the
admin
would
create
the
volume
annually
for
you
and
then
which
that
ever
go
to
consume.
That's
one
way
and
other
ways
dynamic.
One
is
provisioning
where
your
concept
of
storage
class
and
the
storage
class
concept,
the.
A
B
B
Behind
the
cement
can
come
from
any
of
the
driver,
what
has
been
configured
by
the
cluster
manager
there
so
with
that,
we
can
kind
of
manage
both.
What
is
the
dynamic
and
now
one
second
will
provisioning.
Mostly
people
use
that
when
you
put
a
wasting,
because
you
want
the
kind
of
the
world
event
of
assigned
the
one,
you
had
a
PI,
you
want
when
you.
A
B
It
and
there
are
some
drivers
which
can
also
suppose,
if
you
turn
the
volume
of
Panjim
in
the
beginning,
I
want
this
in
10
to
15
GB,
and
there
are
some
drivers
which
can
help
you
to
scale
from
10
to
15
as
well
as
at
one
time.
So
that's
all
ask
question
what
we
had
for
what
we
can
have
time
to
discuss
today.
So
where
do
we
go
from
here?
B
First
of
all,
as
we
kind
of
know
that
there
is
a
free
course
available
on
X,
which
kind
of
control
return
of
powers
all
of
the
basic
questions
we
asked,
and
you
can
take
that,
of
course,
for
free
and
then,
of
course,
as
a
company
cloud
is
of
you
provide
trainings
and
consulting.
So
you
can
push
out
our
science
there
and,
as
you
mentioned
about
there
going
to
be
a
quiz,
so
here
is
the
slide
for
the
quiz.
A
We
miss
the
slide
or
Playa
Vista
flight
for
a
lockpick.
Well,
I
was
just
gonna
say
on
that,
so
the
the
other,
like
my
recommendations,
would
be.
You
know,
take
a
look
at
the
attics
courts
that
independent
are
talked
about,
taking
a
look
at
cloud
or
Oracle
calm.
That's
where
we
are
actually
pushing
out
a
bunch
of
things
that
you
know
our
customers
are
kind
of
asking
us,
for
you
know
how
do
I
do
Nettie's,
what
or
not
and
yeah
I
think
I
think
that's
about
it.
Thank
you,
Mario.
Thank
you.