►
From YouTube: Getting Started With Kubestr
Description
Kubestr can assist here in three ways:
· Identify the various storage options present in a cluster.
· Validate if the storage options are configured correctly.
· Evaluate the storage using common benchmarking tools like FIO.
About the Guest
🎓 Michael Cade | Community Technologist KastenHQ Veeam 🎓
🐦: https://twitter.com/MichaelCade1
⚒️ About the Kasten by Veeam
Kubernetes-native Backup, Disaster Recovery, and Application Mobility for all your cloud-native applications.
🌍 : https://www.kasten.io/
🌟 : https://github.com/kastenhq/kubestr
⚒️ About the Veeam Software
Veeam is the global leader in Backup that delivers Modern DataProtection.
🐦: https://twitter.com/Veeam
🌍 : https://www.veeam.com/
👨🎓 About the Host 👨🎓
🐦: https://twitter.com/cloudnativeboy
🤝: https://www.linkedin.com/in/saim-safder/
Join Cloud Native Islamabad community.
🙌 : https://discord.gg/yypcwN6
🧑🤝🧑 About the Community 🧑🤝🧑
🐦: https://twitter.com/CloudIslamabad
🤝: https://www.linkedin.com/company/cloud-native-islamabad/
Here's we host our Cloud Native Webinar's Thanks to CNCF:
📺: https://community.cncf.io/islamabad/
🔔: https://www.youtube.com/channel/UCzeq-M5SLkFyXe71TYUkgjg
😎: https://www.instagram.com/cloudnativeislamabad/
#CloudNativeIslamabad #Kasten #Kubestr
A
A
And
today
is
my
second
day
of
doing
the
live
stream,
live
stream
and
the
cloud
native
islamabad
and
I'm
very
excited
today,
because
we're
talking
about
storage
and
we
haven't
covered
a
lot
about
storage
in
this
cloud
native
slumber,
youtube
channel.
I
hope
this
has
been
been
the
year
of
globe
cloud
native
storage
for
everyone
make
it
more
easy
and
more
accessible
to
everyone.
A
So
thanks
everyone
for
joining
us
today
is
my
guest
from
rheem,
michael
kate,
he's
a
community
technologist
at
the
casting
by
wien
company,
and
they
bring
out
some
of
the
awesome
tooling
in
the
ecosystem
around
the
kubernetes
story.
So
over
to
you,
michael
so
tell
us
about
the
experience,
be
working
on
the
casting
team
and
plus
your
open
source
journey.
So
far,.
B
Yeah
thanks
for
having
me
first
of
all
and
everyone
out
there
on
the
internet
so
yeah,
I
guess
from
a
custom
point
of
view.
Our
focus
is
really
about
data
management,
data
protection
in
kubernetes
or
in
the
cloud
native
space.
But
what
I
think
is
open
is
more
around
the
open
force
arena.
So,
yes,
we
focus
on.
We've
got
a
couple
of
projects
that
focus
on
application
frameworks,
so
we
have
canister
which
allows
us
to
take
some
application
consistency
from
within
our
kubernetes
environments.
B
So
that's
really
been
my
biggest
focus
for
the
last
few
months
and
then,
basically,
nearly
a
year
ago,
we
released
another
project
called
cubester
and
cubester,
and
this
was
me
also
ramping
up
on
in
the
cloud
native
space
right.
So
so
I
was,
I
very
much
come
from
a
storage
background.
I
I
come
from
a
virtualization
and
storage
background,
mostly
on
the
operations
side.
Looking
after
infrastructure
and
then
being
with
veeam
for
the
last
seven
years
has
been
mostly
focused
around
data
management.
B
For
the
virtualization
and
cloud
environment,
then
I've
been
moved
into
casting
to
focus
on
community
and
being
that
technologist
and
go-to
person
there
from
from
that
point
of
view,
and
then
along
came
this
open
source
like
side
swipe.
That's
obviously
everything
to
do
with
cloud
native
and
everything
that
we're
doing
here,
so
my
interest
has
grown
and
grown
over
the
last
12
months,
not
only
grown,
but
also
the
education
around
that.
But
one
thing
I
did
find
is
that
everything
everything
has
a
learning
curve.
Everything
has
a
potentially
steep
learning
curve.
B
Now,
just
because
I
come
from
storage,
that
doesn't
mean
you're
going
to
get
kubernetes
storage
straight
away.
It
doesn't
mean
if
you
come
from
networking
you're
going
to
get
kubernetes
network,
and
I
know
you've
done
some
stuff
around
kubernetes
networking
and
also
the
compute
and
every
everything
on
here.
But
storage
is
just
that.
Another
thing
that
we
have
to
learn
another
thing
that
we
have
to
get
our
heads
around
and
be
able
to
understand
that
and
the
that
was
really
how
cubester
was
born
in
that.
B
B
We
know
that
we
want
to
understand
whether
we've
validated
and
understand
whether
it's
working
correctly,
because
we're
relying
on
all
of
that
configuration
we're
relying
on
that
yaml
to
say,
do
the
right
thing
and
that's
fine
everything
I
show
you
with
cubester
today
can
be
done
via
like
you
could
do
this
yourself,
you
could
roll
your
own
people
have
the
beauty
about
cubester.
Is
that
it's
fully
automated
into
a
single
go
binary
that
allows
us
to
spin
that
up
inside
our
kubernetes
context
and
then
run
some
tests
against
that?
B
A
That's
the
I
think
I've
been
going
through
is
the
system
of
the
setup.
I
know
this
story
is
going
to
be
really
very
different,
might
be
complicated
for
the
some
of
the
folks
who
are
really
new
to
gubernators,
but
their
day
to
job
is
actually
on
the
storage.
They
might
not
be
so
much
a
die
hard
fan
of
the
kubernetes,
but
they
want
something
to
have
set
up
and
the
story
is
set
up
and
configured
correctly
and
previously
we
don't
have
those
such
tools,
but
thankfully,
right
now
the
cubester
they
can
validate
they
can
they.
A
Can
you
let
just
a
one
boundary
as
you're
mentioning,
and
it's
really
easier
for
a
lot
of
the
folks
as
well,
and
hopefully
the
devils.
Do
you
prepared-
and
you
we've
seen
during
the
lives
discussion,
really
help
us
actually
advance
our
career,
that
we
can
more
resilient
about
kubernetes
story,
because
a
lot
of
the
company
is
not
resilient
enough
to
going
to
the
storage
like
running
the
state,
full
app
in
kubernetes
and
feel
resilient,
but
right
now
like
tooling,
the
people
got.
A
I
think
they
asked
us
to
go
about
the
stateful
app
and
that's
a
good
good
story
to
hear
about,
and
also
people
are
listening
to
us
or
watching
live.
Follow
michael
k
on
twitter
he's
been
he's
been
and
follow
his
journey
of
the
devops
90
day
of
devops
and
for
one
following
every
day.
I
learned
so
many
cool
things
from
it.
So
before
we
jump
start,
the
cubes
just
tell
us
about,
in
a
few
words
your
journey
so
far
in
the
devops
90
day
of
devops.
B
B
B
For
me,
and
I
really
wanted
to,
I
tend
to
set
goals
at
the
very
start
of
the
year
from
like
january,
the
1st
to
march
the
31st,
because
that's
90
days,
if,
like
this
q1,
it's
a
it's
a
nice
start
kickoff
to
the
year,
whether
it's
a
health
kick
or
whether
it's
a
like
learning
journey
and
really
that's
where
the
90
days
of
devops
came.
But
it's
not
just
as
as
anyone
knows,
and
if
you
don't
that's
fine
as
well,
but
devops
isn't
a
tool.
B
So
what
I
did
was
I
took
those
90
days
and
instead
of
going
90
days
into
one
topic,
which
some
people
can
clearly
do
and
hats
are,
but
I
like
so
good-
that's
really
good.
I
can't
do
that,
so
I
split
it
down
into.
I
think
13
13
weeks,
12
13
weeks
of
seven
days
and
we've
got
topics
that
we're
gonna
hit
for
seven
days.
We're
gonna
cover
the
fundamentals,
the
foundation,
and
then
we
move
on
to
something
else.
So
seven
days,
including
the
weekend,
we're
gonna
spend
an
hour
a
day.
B
It's
been
super
interesting,
like
we're
over
a
third
of
the
way
now
so
good
stuff,
and
then
we
start
to
get
into
things
like
version
control,
and
now
we
get
into
like
containerization
next
and
then
kubernetes,
but
only
seven
days,
so
we're
only
going
to
get
a
foundational
level,
but
having
a
having
a
lot
of
knowledge
about
a
lot
of
things
helps
build
that
picture
better
and
I
feel
like
if
I'd
have
had
that
a
year
ago.
It
would
have
helped
me
massively.
A
Yes,
yeah,
I
think,
I
think
absolutely
and-
and
I
think
we're
really
lucky,
because
right
now
we
have
so
much
content
available
on
the
internet
for
the
devops.
In
the
my
days
like
they're
back
in
five
more
years
later,
we
don't
have
enough
about
the
content
available
on
the
devops,
and
people
are
jumping
into
it
and
being
able
to
not
know
what
actually
my
job
was,
but
thankfully,
now
we're
able
to
understand
what
is
actually
our
role
was
and
people
who
are
following
your
journey.
A
They
can
understand
that
they
need
to
do
the
networking
of
the
networking
setup,
how
the
configuration
management
setup,
how
the
linux
or
windows
or
mac
os.
You
need
to
understand
these
knowledge
and
how
you
compile
different
languages
of
code
and
how
to
make
testing
easier.
So
I
think
you
need
to
learn
so
many
things
about
of
that.
So
now
you
I
can
you
can
share
your
screen
and
jump
on
the
story.
B
B
We
should
be
good,
we
are
definitely
live
so,
first
of
all,
if
you
want
to
learn
anything
more
about
cubestar,
this
is
probably
the
first
place
to
go
cubester.io
bit
of
a
simple
introduction
to
what
it
is
I'll
touch
on
what
that
is
as
well.
Obviously
as
we
go
through
some
resources,
some
that
I've
done
some
that
the
community
have
done
yeah,
so
some
are
on
there
and
then
how
do
we
use
it?
How
do
we
download
the
binary?
How
do
we
get
to
use
it?
B
B
B
And
if
you
don't,
if
you
know
you
do
or
don't
know
when
you
roll
like,
let's
I'll
use
eks
as
the
example
eks
or
any
of
them,
they're
made
up
of
virtual
machines.
The
nodes
within
your
kubernetes
cluster
are
made
out
of,
for
the
most
part,
they're
made
out
of
ec2
instances
or
virtual
machines
within
within
the
cloud
platform
and
a
diff.
Each
node
has
a
different
storage
capability
so
and
if,
for
those
that
have
looked
at
that
long
list
of
different
compute,
you
have
a
very
different
like
you
have
a
lot
of.
B
You
have
an
overwhelming
amount
of
options
when
it
comes
to
storage
and
compute,
and
actually
the
compute
will
affect
the
storage
and
the
storage
will
affect
the
compute.
You've
got
things
like
ultra
ssd,
you've
got
ssd,
you've
got
spinning
discs,
you've
got
so
many,
and
if
you
look
at
like
microsoft,
are
the
same
with
a
with
azure,
you've
got
aws,
you've
got
gke
and
everything
else
in
between.
B
So
one
of
the
things
that
the
first
easy
button
for
cubester
is
about
identifying
that
various
storage
options
that
you
have
in
your
in
your
kubernetes
cluster.
Now
we
work
on
a
the
current
context.
So,
if
your
cube,
if
you
have
many
different
content
cubers,
if
you
have
many
different
kubernetes
contexts,
then
we
can
zip
between
them
and
we
can
go
and
have
a
look
at
what
h1r
is
and
what
what
the
it's.
B
Basically
there
to
identify,
what
storage
is
being
used
or
not
being
used,
because
that's
another
thing:
if
you've
run
these
ec2
instances
up
and
you're,
not
using
any
of
that
storage,
then
why
have
you
got
it
like?
You
probably
don't
need
it,
especially
if
you're
using
something
like
ebs
or
you're,
using
something
else,
external.
B
I
want
before
I
put
a
load
of
the
actual
real-life
data
on
that
I
want
to
make
sure
that
it's
configured
correctly,
so
I
want
to
validate
that
the
storage
options
are
configured
as
they
should
be.
I
want
to
make
sure
that
it
does
take
a
snapshot
because
I've
got
a
volume
snapshot
class.
I
want
it
to
automate
that
and
that's
what
cubester
will
do
and
again
I'll
run
through
that
shortly
as
well,
and
then,
thirdly,
is
benchmarking.
B
B
Well,
the
same
thing
happens
from
a
from
a
kubernetes
point
of
view
or
any
platform,
point
of
view,
so
being
able
to
evaluate
the
storage
using
something
like
fio,
which
is
flexible
io,
which
has
been
around
for
years,
but
using
that
same
trusted
tried
solution
gives
us
the
ability
to
understand
okay.
If
I've
chosen
this
compute
node
and
this
storage
tier
in
this
situation,
I'm
gonna
get
this
performance
and
vice
versa,
and
we
start
to
get
that
and
the
very
beginning
of
the
project.
B
The
idea
was:
is
that
we'd
entice
storage
vendors
cloud
providers
to
actually
provide
their
own
benchmarking
for
their
op
for
their
options
that
they
have
so
that
the
community
could
just
come
to
cubester,
whether
it
be
cubestrio
or
the
github
and
be
able
to
understand.
Well,
if
I
choose
node
a
type
1,
I'm
going
to
get
this
amount
of
iops
in
my
cluster.
B
If
I
choose
node
b
type
3,
I'm
gonna
get
this
or
if
I'm
using
a
enterprise
storage
solution,
I'm
gonna
get
this
all
good
stuff
and
then
more
recently
we
released
the
ability
to
actually
browse
or
view
the
content
of
a
persistent
volume
claim.
Now.
The
important
thing
here
is
that
we're
not
going
into
the
live.
Persistent
volume
claim
we're
actually
cloning,
the
persistent
volume
claim
using
the
csi
technologies,
the
the
volume
snapshot
and
being
able
to
then
mount
that
to
a.
B
You'll
see
where
I
go
here
when
I,
when
I
talk
about
this,
is
that
not
very
many
people
put
unstructured
data
into
persistent
volume
claims,
it's
mostly
databases
or
some
sort
of
messaging
queue
that
really
in
a
flat
flat
file
file
system
it
it's
not
worth
it
you'll,
see
that,
because
I'm
going
to
use
my
sql
as
a
not
as
an
option
and
you're
going
to
just
see.
Well,
that's
the
database
file
great.
I
don't
get
anything
from
that,
so
this
is
another
ask
for
the
community.
How
can
I
mount
that?
B
How
can
I
take
the
same
methodology
but
then
use
something
like
a
mysql
in
a
web
browser
to
be
able
to
mount
that
so
that
we
could
then
see
the
data
and
that
clone
data
and
then
think
of
the
possibilities
there
about
being
able
to
understand
what
that
data
is.
What
do
we
do
with
that
data?
At
that
point,
should
we
be
doing
that
data
anything
with
that
data?
Should
we
be
masking
it
before
it
happen?
There's
a
lot
of
different
areas
that
this
could
go.
B
Okay,
so
they're
the
four
things
I'm
going
to
show
in
the
in
the
demo.
Then
we've
got
a
small
little
demo
here,
but
obviously
I'm
not
going
to
show
you
that,
because
I
want
to
show
you
it
live
and
then
we've
got
the
resources.
So
just
so,
you
are
also
aware.
So
there
is
a
link
down
the
bottom
of
this
page
that
will
take
you
to
the
github
page
and
in
here
again,
very
similar.
B
It's
going
to
give
you
some
short
instructions,
pretty
much
like
most
readmes
within
github,
and
it's
going
to
give
you
the
the
idea
of
everything.
The
other
thing
I'll
point
out
is
that
it's
pretty
active
in
terms
of
issues.
If
you
find
something
that
you
want,
that's
missing.
Let's
add
it
as
an
issue.
B
We've
got
engineers
that
are
dedicated
to
this
from
a
maintaining
point
of
view,
but
we'd
also
like
your
contributions,
your
prs,
but
find
something
that
you
want
and
put
it
in
there,
like
we've
seen
a
couple
of
couple
of
pr's
that
are
that
are
in
there
as
well.
So
I
think
we
could
probably
jump
into.
B
B
B
I
can
have
it
running
in
like
five
minutes
and
then
go
and
deploy
k10
and
I've
got
some
other
in
that
same
in
in
my
own,
repo
and
I'll
show
that
at
the
end,
if
I
remember
where
the
90
days
of
devops
is
there's
also
the
project
pace
idea
as
well,
so,
okay,
so
we've
got
our
nodes.
We've
got
our
storage
class,
one
more
thing
to
check
volume,
snapshot
class
we've
got
our
host
path,
snap
class,
and
this
is
what
that
add-on.
B
B
So
arcade
is
a
application
marketplace
for
kubernetes.
If
I
do
arcade
and
I
think
it's
info
or
okay
get
so
this
is
again
another
open
source
tool
and
if
you
want
another
way
of
being
able
to,
you
can
obviously
download
the
binary
from
github.
But
if
you're
already
using
arcade,
then
you
could
do
arcade
get
and
then
arcade
get
or
arcade
install
cubesta.
I
can't
remember
which
one
it
is
and
that
will
go
and
download
the
latest
binary
from
the
github
site
anyway
it.
B
But
you
can
also
see
in
here
that
we've
got
lots
of
various
different
other
tools
that
are
available
in
there.
So
again,
let's
clear
that
so
cubester
on
its
own.
If
we
take
a
look
at
cubes
to
help
you're
going
to
basically
see
the
options
very
short
sweet
in
what
it
does,
it's
going
to
it's
a
cubester
as
a
tool
that
will
scan
your
kubernetes
clusters
and
validate
the
storage
systems
in
place,
as
well
as
run
some
performance
tests.
B
So
here
are
the
three
things
that
I
mentioned,
and
then
we
have
cubes
that
that
can
be
run
standalone.
That
will
give
you
that
identification
of
the
storage
that
you
have
and
I'll
run
that
shortly
so
pretty
simple
stuff
and
I'm
going
to
run
through
each
one
of
these
options
as
we
as
we
move
through.
B
B
B
Okay,
so
I
have
got
it
in
there
perfect.
Basically,
all
I
did
was
in
the
path
of
my
cube
ctl.
If
you
move
the
cubes
to
binary
and
you
put
cube
ctl
hyphen
cubesta,
then
that
becomes
a
a
plug-in
that
you
can
use
on
your
system
and
we're
working
on
whether
we
actually
want
to
go
down
the
route
of
putting
that
into
the
package
managers
or
the
plug-in
managers
around
that
as
well.
B
B
Obviously
no
one
should
be
running
this
in
production,
but
if
we
were
to
run
another
one
and
if
we
were
to
run
a
validated
one,
it
would
come
up
in
that
list
and
say
this
is
a
validated
csi
driver.
You
can
see
that
has
a
storage
class
associated
to
it
and
it
tells
you
you
can
run
an
fio
test
against
this
by
running
this,
and
then
we
have
our
mini
cube
host
path,
which
is
an
unknown
driver
type.
B
A
storage
class
is
standard,
it
is
not
csi,
but
you
can
still
run
fio
against
it,
because
there's
still
many
companies
out
there
that
don't
run
csi.
Yet
it's
only
relatively
new.
So
with
that
we
can
now.
Okay,
we've
done
that.
We've
identified
the
storage
that
we
have
available
within
our
system.
Now
we
could
just
do
cube,
ctl
get
storage
class,
and
that
would
give
you
some
information,
but
it
won't
necessarily
give
you
all
of
the
information
that
you
that
you
want.
B
B
This
is
where
we
start
to
see
a
little
bit
more
option
or
a
few
more
options,
so
we
could
just
run
it
against
the
storage
class
and
the
volume
snapshot
class,
and
that
will
go
away
and
I'll
show
you
what
that
does
in
a
second,
but
we
also
can
define
what
else?
What
do
you?
What
else
do
you
want
to
do
so?
Do
you
want
to
use
a
specific
name
space
for
it,
or
is
it
just
going
to
go
into
the
default?
B
Do
you
want
to
skip
something?
Do
you
want
to
do
like
you
can
see
which
ones
are
required,
but
you
can
see
here
that,
like
what
do
you
actually
want
to
check-
and
these
are
always
open
to
to
community
feedback
on
what
else
could
we
do
well?
I
think
that
would
be
pretty
interesting
to
find
out
other
areas
that
we're
that
we're
missing
or
that
are
missing.
So
if
we
run
cubester
csi
check
and
we
do
a
against
the
storage
class,
csi
host
path
and
storage
class,
sorry,
the
volume
snapshot
class
host
path.
B
B
It's
now
creating
that
that
clone
and
now
it's
getting
rid
of
it.
Because
what's
going
to
happen,
is
it's
going
to
come
back
here
and
it's
going
to
say
we
created
a
pod
yep.
We
just
saw
that,
although,
albeit
I
was
late
and
then
we
created
a
snapshot
and
we
restored
that
pod-
that
pop
that
cloned
pod
and
the
pvc
and
then
we
went
away
and
we
cleaned
up
so
another
way
in
which
we
could
look
at
that.
If
I
run
it
again
is
I
could
go
to
cube
ctl
get
pvc.
B
B
B
I
can't
see
the
chat.
So
if
there's
any
questions,
then
please
interrupt
and
I'll
I'll
cover
some
of
them.
So
then
the
next
one
is
about
performance
benchmarking.
So
if
we
run
cubes
cube
seats,
cubester
fio
slash
help.
This
is
where
again,
we've
got
those
options
around,
or
what
do
you
want
to
run
like?
B
We
could
just
run
the
standard
fio
test
that
comes
out
of
the
box,
which
I
think
is
a
4k
test
and
a
8k
test
read
them
right,
which
might
be
sufficient
for
what
you're
trying
to
achieve,
but
it
might
be
that
you've
also
got
a
specific
fio
file
that
you've,
that
is
natured
to
your
application
and
the
data
set
that
you're
pushing
into
that
system.
Volume
claim
and
you
want
to
test
to
make
sure
that
you've
picked
the
right,
node
and
the
right
storage.
B
So
there
we
can
start
to
do
other
things
and
I'll
show
you
what
that
looks
like.
But
these
are
the
options
you
can
see
here.
We
can
change
the
size
as
well,
because
we
know
that
if
we
create
a
default,
100
gig
pvc,
then
that
might
not
be
relative
if
you're
looking
at
terabytes
of
of
of
data,
so
it
might
be
that
you
actually
want
to
go
and
create
a
pvc,
that's
much
much
bigger!
B
B
B
You
see
here
that
we
have
100
gig,
because
that's
the
default
setting
on
our
system
and
we're
going
to
run
against
that.
You
can
see
here
we're
still
running
if
we
go
back
here,
you
can
see
that
we're
running
the
fio
test
default,
fio
I'll,
tell
you,
which
one
I'll
show
you
which
what
the
default
tests
are
so
here-
and
this
is
another
community-
ask
like.
Obviously
we
can.
We
can
output
this
to
json
and
we
can
start
putting
that
into
different
engines
to
better
visualize
this.
B
But
as
it
comes
through
the
terminal,
you
can
probably
see
that
it's
you
can
get
what
you
need,
but
I
think,
if
you're
doing
this
at
scale
and
automating
this
and
wanting
that
information,
you
probably
want
some
sort
of
and
if
you're
doing
it
consistently
like
this,
isn't
a
job.
This
isn't
a
oh
we're
going
to
use
cubesta
once
on
day
zero
and
then
we're
never
gonna
use
it
again
because
we
know
especially
on
premises
like
storage,
can
get
degraded
in
performance.
So
let's
make
sure
that
we're
getting
the
right
slas
to
our
storage.
B
Okay,
so
we
can
see
the
version
that
we're
running.
We
can
see
that
from
a
read
point
of
view,
we're
doing
something
with
4k
block
size,
2
gig.
I
o
depth
random,
read
and
the
read
and
we're
getting
7
524
we've
got
a
read
of
128.
So
sorry,
it's
4k
and
128k
reads
and
128k
right,
but
there
you
go.
You
can
see
what
we've
got
in
terms
of
iops
very
quickly
on
that
on
that
system.
Bearing
in
mind
this
is
a
desktop
machine,
so
it's
not
gonna.
B
It's
not
gonna
blow
the
it's
not
gonna,
be
incredible
but,
like
I
say
we
can
bring
our
own
fio
and
maybe
it's
something
like.
I
want
to
see
how
quick
it
is
to
read
backwards.
The
whole
file
system
and
if
we
just
go
here,
you're
gonna
see
that's
all
gone
now
terminated
and
if
I
do
it,
the
pvc
has
also
been
been
dealt
with.
Let's
I
guess
we
could
just
run
that
again
and
then,
if
we
go-
and
this
might
work,
if
that's
still
there,
that
file's-
not
there.
B
B
B
B
If
we
wanted
to
do
a
particular,
let's
say
we
wanted
to
do
a
50
gig
just
for
time,
more
than
anything
and
space
on
my
on
my
machine
is,
we
could
just
add
a
minus
zed
and
clear,
and
now,
if
we
go
and
look
at
this
again,
we
can
now
see
that
that's
a
50
gig,
so
you
can
cater
this.
So
the
idea
here
is
that
you
can
bring
bring
the
size
of
what
your
workload
needs
to
look
like,
so
that
you
can
match
it
to
what
your
application
your
data
service
is
going
to
be.
B
Look
and
feel
as
your
application,
I
guess,
is
a
good
way
of
putting
that.
Okay,
so
that
that's
all
the
stuff
that
was
in
that's
been
in
cubesta
since
really
the
beginning,
there's
been
some
enhancements,
but
really
that's
the
that's
the
that's!
That's
what
we've
had
now.
The
new
feature
that
came
out
just
before
christmas
was
the
ability
to
browse
I'm
actually
going
to
pull
this
up
here,
because
I
think
in
fact
it
won't
matter.
B
B
So
if
we
were
to
go
and
run
cubester
browse
make
sure
that
that's
right
we're
going
to
nominate
the
persistent
volume
claim
that
we
want
to
clone
we're
going
to
choose
our
volume
snapshot
class
again
and
we're
going
to
choose
the
name
space
in
which
the
persistent
volume
claim
lives
and
then
what
we're
going
to
do
in
fact
kill
that
this
is
the
magic,
so
cube.
Ctl
get
pods
minus
w.
Okay,
my
sequel's
running
all
good,
then.
B
Now
I'm
gonna
get
lost,
qctl
get
pvc,
my
production,
app,
okay,
we've
got
one,
it's
one,
gig
all
good
so
on
here.
Let's
go
ahead
and
hit
that
so,
first
of
all,
we're
going
to
take
that
snapshot
again.
Okay,
we
see
that
we're
cloning
that
okay
makes
sense.
Now,
let's
go
and
have
a
look
at
our
persistent
volume
frames.
Okay,
we
have
a
clone
good
stuff
and
we're
creating
a
container.
B
File
browser
so
that
we
can
now
see
inside
of
that
pvc
now,
if
I
drill
into
here,
you're
gonna
see
some
very
database
centric
things
my
sequel
and
you
could
go
and
grab
your
database
if
there
is
anything
actually
on
this
system
and
you
can
go
and
pull
that
out
and
you
can
go
and
do
whatever
you
need
to
do
to
it,
but
where
we
think
that
we're
going
to
go
next
is
well.
How
do
I
make,
instead
of
it
being
the
file
browser
container
image
that
we're
using
here?
B
What
about
a
web
version
of
my
sequel
or
a
web
version
of,
or
a
web
version
of
postgres
to
be
able
to
mount
that?
And
whether
there
is
I've
not
had
chance
to
look
into
that
yet?
But
what
you're?
What
we're
doing
here
and
how
we're
interacting
is
not
actually
with
our
live
system
like
so
we
could
come
in
here.
We
can
see
our
data.
We
can
manipulate
that
data.
B
So
that's
another
another
area
and
if
I
hit
hit
close
on
that
or
I
could
just
come
in
here,
you
see
that
we've
put
forward
in.
If
I
now
hit
control
c,
it
then
goes
away
and
cleans
up
all
of
those.
You
can
see
that
pretty
much
straight
away.
It
was
terminating
so
that
you're
not
you're
not
using
that
bandwidth
you're,
not
using
that
that
extra
little
tiny
bit
of
compute
so
they're
the
four
key
areas
to
show
of
of
cubester
yeah.
I
think
the
biggest
ask
is:
is
that
useful?
B
A
So,
michael,
I
am
going
through
that
going
to
the
docs
chat
as
well.
Just
just
go
going
through
this,
some
of
the
use
cases
the
fyu
like
one
of
the
thing
you
did,
you
explained
about
like
the
fio,
what
it
is
actually
and
like
one
of
the
uses
you've
done
like
you
have
a
file
storage.
We
have
a
container
story
that
is
really
huge
and
you
need
to
have
a
that
available.
Persistent
available
for
you
like
in
the
past.
B
B
A
B
Obviously
it's
going
to
depend
on
where
that
is
whether
it's
on-prem,
whether
it's
in
the
cloud
and
whether
you
have
any
control
over
that
as
well.
Like
I've
known,
I
know
a
lot
of
us
on
the
call
will
will
not
have
any
idea
about
what
this
underpinning
storage
is.
I
speak
to
a
lot
of
developers
on
a
daily
basis
that
believe
that
the
the
database
is
the
storage
layer,
but
from
an
operations
background.
A
So
I've
learned
this
today,
like
I've
heard
about
have
heard
how
about
fio
before
and
the
the
use
cases
to
draw.
It's
really
really
really
applicable
like
and
the
way
like,
it's
hiding
the
complexity
from
that.
So
that's
the
wonder.
Wonderful
thing
is
and
also
another
thing,
one
of
the
things
I
see
like
in
the
kubernetes
previous
version,
the
new
version
came
up.
There
are
a
lot
of
the
things
happening
like
so
many
apis
or
the
storage
scenes
behind
the
scene.
A
Kubernetes
new
updates
come
up
and
the
stuff
are
not
working
and
I
think
you
can
validate
validate
those
things
up
using
the
cubester
as
well.
So
can
you
draw
some
points
on
here
like
in
the
kubernetes
1.23?
I
see
some
of
the
some
of
the
deprecation
of
the
csi
things
and
some
of
the
new
capabilities
come
up
and
cubesat
does
just
validate
you
like.
They
can
confirm
you,
that's
and
this
about
the
news
kubernetes
api
course
to
the
csi.
So
talk
us
about
this.
These
wonderful
use
cases
of
using
the
cubes
too.
B
Yeah
yeah
absolutely
so
obviously,
there's
a
lot
of
there's
a
major
shift.
In
fact,
I
think
I've
got
a
nice
slide
that
kind
of
touches
about
that
without
making
it
a
slide,
because
everyone's
had
enough
slides
right,
maybe
I
haven't
got
it.
Maybe
I
haven't
there's
so
there
was
something
in
in
one
of
the
sessions.
B
B
We
don't
care
as
long
as
they
stay
here
then
we're
always
going
to
work,
whereas
the
intrigue
provisioner
and
that
storage
and
then
what
we've
seen
over
the
last
12
to
18
months
is
a
lot
of
migrations
from
intrigue
provisioners
or
someone
correct
me
on
the
chat
as
well.
If
anyone's
there
is
it
the
flexible
volume,
so
entry
provision
of
flexible
volumes,
csi
just
another
way
of
being
able
to
add
storage
to
your
kubernetes
cluster.
B
So
the
depreciation
is
my.
My
understanding
is
that
it's
more
around
the
entry
provisioner
and
also
that
shift
of
people
moving
or
like
migrating
to
csi
store
back
storage
or
their
role
in
new
kubernetes
cluster,
which,
by
the
most
part,
is
de
facto
csi,
which
obviously
gives
you
that.
But
you
want
to
make
sure
that
it's
still
configured
correctly
we're
in
that
transitional
period,
where
you
still
especially
with
the
public
clouds,
you
might
still
be
deploying
the
csi
volume
snapshot,
pod
and
and
configuration
for
that.
B
So
one
is
the
volume
of
clusters
that
some
customers-
not
customer
people,
are
running
smaller
clusters,
but
with
storage
requirements.
So
having
that
output,
you
saw
that
output.
We
can
output
to
json
as
well.
I
didn't
run
through
that,
but
you
can-
or
you
can
have
it
just
in
the
terminal
again.
That
might
be
fine
if
you're
just
running
it
against
one
cluster.
B
But
if
you
want
to
automate
that
task
to
go
through
20
different
clusters,
which
is
not
unheard
of
then
that
that's
not
going
to
be
readable,
like
your
terminal,
is
going
to
fill
up.
It's
going
to
not
be
very,
it's
not
going
to
you're
going
to
make
more
work
for
yourself.
So
one
of
the
biggest
answers
around.
How
do
we
visualize
it?
How
do
we
better
visualize?
How
do
we
export
out?
B
Maybe
it's
j
and
I
think
it
is
json-
is
the
right
right
export
for
that
data
and
then
how
do
we
import
that
json
into
something
else,
and
it
might
be
that
we
do
cubes
to
visualize
and
it
spins
up
another
pod,
but
the
and
the
visualize
aging
pushes
into
that,
and
then
we
get
a
web
page
with
those
results.
Something
like
that.
I
don't
know
what
that
looks
like
I'm,
not
a
developer.
B
I
bet
most
of
us
on
the
call
we
can.
We
can
probably
span
all
over
the
operating
system
ecosystem,
also
extending
of
that
csi
driver.
Like
again,
I
said
about
this
being
an
evolution,
storage
companies
developing
against
that
csi,
and
we
obviously
want
to
be
able
to
be
able
to
support
that.
So
we
add
that
this
is
the
screenshot
I
was
looking
for,
so
where
the
mini
cube.
One
is
not
supported,
it's
not
on
a
public
list.
B
Obviously
the
ebs
one
from
from
aws
is
so
it
gives
you
that
it
gives
you
that
option
there
as
well,
and
you
can
see
that
it
gives
actually
a
little
bit
more
information.
It
gives
you
your
storage
class,
your
volume
snapshot
class
and
then
it
says
about
your
fio
and
to
test
csi
snapshot
and
restore
functionality.
B
I
think
what
else
was
that
so
and
then,
more
recently
the
biggest
ask
is
about
how
do
we
like?
They
think
that
browse
feature
is
great,
but
how
do
we
expose
that
to
a
relevant
tool
that
can
mount
that
database
or
mount
the
databases
within
the
pacifi
within
the
cloned
pvc
and
then
you've
gotta?
Ask
yourself
whether
that's
the
right
thing
to
do
as
well
and
that's,
I
think,
where
we've
potentially
slowed
down,
because
it's
great
for
a
demo
like
I
can
show
you
on
my
sequel.
B
Maybe,
but
I
think
it
has
to
be
governed
and
it
has
to
be
treated
correctly
from
that
point
of
view.
But
that's
my
data
management
head-on.
My
the
possible
hat
comes
off
at
that
point
and
it
goes
oh,
let's
how
do
we
do
it?
How
do
we
visualize
that
data?
Because
then
we
can
start
to
do
some
interesting
things
on
top
of
that
data,
without
affecting
the
production
data
and,
generally
speaking,
that
snapshot
that
we're
taking
that
clone.
A
Yes,
absolutely
absolutely
and
see
the
visualize
like
the
things
you're
talking
about.
I
think
I'm
think
I'm
waiting
for
those
things
that
come
up
and
hopefully
that's
it's
going
to
be.
It
makes
the
rest
of
the
people
life
really
easy.
So
before
I
let
you
go
just
tell
us
about
the
people
who
are
really
new
to
kubernetes
storage,
where
they
should
go
and
start
learning
about
it
and
some
tips
around
what
you
learned
over
the
year.
Like
some
of
the
tips,
you
have
learned
about
configuring,
the
kubernetes
story.
B
So
you
can't
go
wrong
with
kubernetes.io
like
the
the
documentation
is
amazing
and
then.
Secondly,
I
don't
know
about
everyone
on
the
call
or
on
the
on
the
stream,
but
I'm
a
little
bit
addicted
to
youtube,
like
I
literally
spend
hours
hours
a
day
on
youtube
and
it's
generally
tech
related
youtube
and
there's
some
great
content.
Creators
out
there,
like
myself,
who
focus
on
that
learning
in
public
and
sharing
that
that
their
their
overview
of
community
storage
or
other
areas,
and
I'm
sure,
like
you,
don't
need
to
look
far
just
like
so
annae's.
B
I
don't
I
don't.
I
don't
think
personally,
you
need
to
pay
for
this
content.
I
think
there's
enough
out
there
and
it's
a
growing
content,
like
you
said
at
the
start,
is
that
we've
got
we've
got
endless
amounts
of
content
around
devops
around
kubernetes
yeah.
I
I,
I
think,
there's
a
there's
a
lot
to
lot
to
be
going
up.
A
B
A
B
A
Possible,
I
think
he
has
every
possible
youtube
content.
You,
if
you
browse
that's
his
channel,
is
pop
up,
because
that's
is
he's
going
to
tell
everyone
very.
I
really
like
about
it
and
also
the
night
to
day
of
devops.
That's
I'm
really
excited
about
and
sharing
the
links
over
here
grab
the
link
from
here
and
just
go,
follow
the
journey
of
michael
gate
on
the
get
off
github.com
and
tell
us
about
what
he's
talking
about,
and
I
think
you
hope
I'll
learn
done.
A
A
lot
is
if
you're,
following
his
journey
as
well,
and
also
on
the
story
side,
if
you
need
more
knowledge
around
that
michael
k
is
always
visible
on
the
twitter
side
and
or
they
can
ask
if
you
have
other
questions
related
to
storage,
join
him.
Follow
him
on
the
twitter,
he's
been
a
wonderful
person
to
engage
with
and
I'm
learning
every
day.
This
is
very
learning
disorder
for
me
on
the
kubernetes
story
and
the
devops
side.
So
thanks,
michael
and
really
wonderful,
chats
coming
up
like
england,
there
is
saying
it's
amazing,
like
hipster
and
king.
A
A
good
good
some
comments
and
that's
making
me
more
smile
out,
because
I
have
tested
out
cubester
and
makes
more
sensible
sense
for
me
to
learn
more
about
it.
So
thank
you,
michael.
It's
been
a
wonderful
session
with
you
today,
we'll
be
continuously
collaborating
on
the
future
and
makes
people
life
more
easier
on
the
story
side
as
well.
So
thanks
for
joining
me
on
the
today's
live
stream,
hope
you
enjoyed
the
time
with
me.
Thank
you.
Everyone
who
showed
up
for
this
live
stream
hope
to
catch
up
to
you
again
in
the
future.