►
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
Kubernetes is now the de-facto platform to deploy and manage stateless applications. It is also making strides for deploying and managing stateful applications. Managing stateful applications is more than provisioning and relocating containers upon failures. Stateful application deal with data gravity, IP persistence, and almost always mandate procedures to protect data. Traditional methods for protecting data are application-level snapshots or storage snapshots provided by vendors. This method does not suffice for the applications running on Kubernetes. This presentation will shed light on some of the challenges and provide some approaches towards complete data protection for applications such as Oracle RAC, Postgres, Cassandra, MongoDB, etc.
A
Let's
get
started
hi
everyone
thanks
for
joining
us
today,
we're
gonna
welcome
to
today's
CN
CF
webinar
data
protection
for
applications
running
on
kubernetes
I'm,
Jorge
Castro,
community
manager
of
VMware
in
a
cloud
native
ambassador
I'll,
be
moderating.
Today's
webinar
we'd
like
to
welcome
our
presenter
today,
Ravi
a
Lu
Bo,
yena
senior
architect
at
Robbin
dot,
I/o
a
few
housekeeping
items
before
we
get
started
during
the
webinar.
You
are
not
able
to
talk
as
an
attendee.
There
is
a
Q&A
box
at
the
bottom
of
your
screen
in
the
your
screen
in
the
zoom
window.
A
So
please
feel
free
to
drop
your
questions
in
there
and
we'll
get
to
as
many
as
we
can.
At
the
end,
this
is
an
official
webinar,
the
CNC
FNS,
such
as
subject
to
the
CN
CF
code
of
conduct.
Please
do
not
add
anything
to
the
chatter
questions
that
will
be
in
violation
of
that
code
of
conduct.
Basically,
please
be
respectful
to
all
of
your
fellow
participants
and
presenters
and
with
that
I'll
hand
it
over
to
Ravi
to
kick
off
today's
presentation.
A
B
Robin
specializes
in
running
app
data,
AV
workloads
on
kubernetes
platform.
So
with
that,
let's
start
on
our
topic
today,
which
is
data
protection
for
applications
running
on
kubernetes.
So
before
we
talk
about
data
protection,
let's
take
a
quick
look
at
the
lay
of
the
land
as
far
as
kubernetes
is
concerned.
B
Let's
look
at
the
state
of
the
applications
that
are
running
on
cobalt
is
today
all
over
the
Priya,
where
of
the
the
first
category
of
applications,
which
are
stateless
applications
here,
node.js
pythons
and
web
applications
tax
engine
x-h
a
proxy
exit,
so
these
are
qualified
as
stateless
applications,
but
it
is
another
big
big
segment
with
three
big
pillars
in
them,
which
is
your
sequel
databases,
the
traditional
Oracle's
of
the
world.
As
ap
is
my
sequel,
posting
services
they
run.
B
Majority
of
the
data
works,
the
enterprise
and
the
new
players
trying
to
offset
some
of
the
sequel.
Databases
are
that
no
sequel
databases
which
can
be
more
woody
bees,
Cassandra,
Couchbase
influx,
etc.
They
are
distributed
in
nature
and
they
are
trying
to
take
off
some
of
the
workloads
out
of
the
sequel,
databases,
the
third
and
the
biggest
segment,
which
is
loosely
termed
as
big
data.
This
is
where
we
have
Hadoop
stacks
elasticsearch,
ELQ.
B
Kafka
pipes
and
other
database
products,
so
this
is
a
live
plant.
They
have
different
segments
of
applications
and
this
is
in
green
because
kubernetes
solves
it
exceptionally.
Well,
whoever
it
is
is
the
de
facto
standard
for
managing
the
stateless
applications
and
it
is
making
rounds
into
sequel,
no
sequel
and
big
data
applications
that
work
is
in
promise
and
kubernetes
is
doing
that
covering
that
ground
gradually.
So
that
is
the
state
of
the
applications
today
on
kubernetes.
B
So
with
that,
let's
get
an
application
composition
in
kubernetes.
If
you
take
a
simple
application
like
my
sequel
traditionally,
if
you
were
to
deploy
it
on
bare
metal
or
OG
machines,
you'd
be
concerned
about
storage
and
compute
storage
could
be
a
direct
attached
disk
or
it
is,
it
could
be
coming
from
an
array
or
you
compute
and
you
compute
is
a
virtual
machine
or
you
install
my
second
on
a
bare
metal,
so
bottom
line.
B
You
are
only
concerned
about
compute
and
storage,
but
when
you
come
to
cool
this
platform
cool,
it
is
a
very,
very
opinionated
framework.
So
you
have
to
do
things
in
a
certain
way
and
that
certain
way
is
this:
if
you
were
to
run
my
sequel
on
kubernetes,
it
is
not
just
a
PVC
and
a
pod,
it's
not
just
origin
compute.
You
need
to
have
all
other
elements
to
give
you
a
fully
functional,
Mexican
application.
It
is
there's
a
deployment
and
there
is
a
replica
set.
There
is
a
service
definition,
there's
a
secret
and
config
map.
B
So
if
my
sequel
is
that
complex,
let's
look
at
the
other
applications.
Is
there
my
sequel
where
I
live
is
the
distributed?
Sequels
and
MongoDB
is,
is
a
no
sequel
database
with
lots
of
components
that
are
configures?
There
are
contexts
servers,
there
are
query
doubters
and
there
are
shards
replicas
and,
as
you
move
up
the
application
pillars,
its
stateless
to
the
sequel
to
was
equal
to
Big
Data.
The
complexity
only
increases.
B
If
you
look
at
I'll
stack,
there
are
several
components
in
understand
right,
not
as
large
time
so
this
Cubana
elasticsearch
itself
is
a
distributed
application.
So
applications
on
kubernetes
are
complex,
craymore,
lots
of
components.
So
with
that,
let's
jump
onto
data
production,
what
is
data
protection
mean
and
why
is
it
different
in
kubernetes
world
when.
A
B
B
Layers
to
running
an
application-
it's
it's
it's
completely
different!
Then
you
want
to
run
it
on
a
bare
metal.
You
just
take
an
RPM
and
install
it,
but
there
is
an
kubernetes.
It
is
hidden
by
multiple
layers.
You
start
with
kubernetes
and
then
darker
and
a
CRI,
CNI
and
CSI
contribute
to
that
dog
running
of
the
dollar,
and
these
are
typically
large
scale
applications,
and
we
are
also
talking
about
multiple
data
centers.
B
What
cooperate
is
allows
us
to
stretch
your
clusters
across
geo
that
across
Amazon
zones
or
regions,
if
we're
talking
about
multiple
datacenters
here
and
the
application
spectrum
also
is
gradually
changing
from
sequel
to
no
sequel
database
and
pts,
so
application
stencils
are
very
complex.
So,
given
this
environment,
what
does
production
mean?
It
means
it
means
several
things.
B
It
could
mean
encryption
or
replication,
but
we're
going
to
talk
about
three
big
pillars
here
now
we
got
to
look
at
it
from
three
main
dish
box,
which
is
how
do
you
protect
the
application
deployment
from
poor
planning
resource
planning?
And
why
is
it
important
we
will
get
into
why?
Why
do
you
have
to
do
it
so
starting
and
why
certificate
in
commodities?
The
second
thing
is,
it
is
user
errors
cannot
be
avoided
in
it.
So
how
will
you
protect
the
application
from
user
errors
and
the
third
and
the
biggest
bucket
is?
B
How
do
you
fail
use
a
bound
to
happen?
This
can
fade,
nodes
can
freely
run,
sites
can
fail,
and
how
do
you
protect
the
applications
from
all
these
videos?
Only
three
like
I
said
there
could
be
many
more,
but
on
in
this
presentation
will
focus
on
these
three.
So
what
is
resource
planning
to
explain
this?
Let's
talk
about
a
distributed
application
like
the
saga,
the.
B
Is
designed,
is
you
need
Cassandra
a
distributed
application?
You
pick
up
three
nodes
and
we
don't
Cassandra
on
these
three
nodes
and
any
data
that
goes
to
one
node
is
replicated
to
the
other
terms.
It
is
configurable,
but
typically
it
is
a
three
been
implicated
system,
so,
which
means
the
computer
is
residing
on
three
physical
servers
and
the
storage
is
directly
attached
to
the
physical
terrace,
which
means
cassandra
is
writing
to
three
different
disks,
so
data
set
up
copying
model.
B
Now,
if
you
take
that
to
Sandra
and
deployed
on
kubernetes
how
it
looks
like
it
is
very
well
possible
that
all
the
three
pods
could
be
placed
on
a
single
server.
That
is
the
first
problem,
but
kubernetes
provides
a
primitive
it's
using,
which
you
can
spread
these
three
pods
on
to
three
different
physical
server,
and
that
is
by
using
more.
B
So
what
happens
in
this
case?
Three
pods
are
running
on
three
different
servers,
but
they
are
writing
to
the
same
disk,
because
these
are
virtual
volumes
that
were
given
on
the
commodities
and
what's
the
problem
here,
it
is
not
tolerant
to
display
use
in
a
way
we
have
negated
the
assumptions
of
the
application.
Cassandra
Cassandra
assume
that
I'm
right
into
three
different
disks,
but
because
of
the
virtualized
software-defined
storage
layer,
those
assumptions
are
violated,
and
how
do
we
fix
this?
So
here,
let's
protect
Cassandra
by
doing
a
proper
planning.
B
B
Now
that
is
Cassandra.
Let's
talk
about
Hadoop,
let's
move
to
the
extreme
right
side
of
our
application.
Spectrum.
Hadoop
is
not
an
application.
It
is
a
platform.
It
is
a
stack.
There
are
several
components
in
Hadoop.
Let's
start
from
Cloud
Data
Manager,
which
is
just
an
uber
UI
layer
which
will
orchestrate
the
entire
Hadoop
cluster
for
the
cloud
or
a
distribution
of
course,
so
cloud
our
manager
is
not
natively
protected.
By
that
I
mean
it
will
use
a
my
sequel
database
to
write
to
storage
the
configuration
metadata.
B
What
that
database
is
not
protected,
it's
not
replicated
or
anything
so,
which
means
the
whoever
it
is
has
to
provide.
A
storage
block,
which
is
internal,
implicit
cloud
or
manager,
requires
the
replicated
storage
and
there
are
data.
Nodes
and
data
nodes
are
similar
to
the
Cassandra
nodes,
because
data
node
does
three-way
replication,
and
but
we
have
to
make
sure
that
all
the
three-volume
that
are
supplied
to
data
nodes
are
coming
from
three
different
disk
levels.
B
There
is
named
node,
which
is
active
and
passive
and
binding,
nor
on
the
active
me
Nimue,
and
there
is
zookeeper
which
is
latency-sensitive.
It
has
to
make
an
IO
in
a
very,
very
short
duration.
Otherwise,
it
will
evict
the
node
so,
which
means
it
has
to
be
writing
to
a
local
distance
or
it
has
to
see
latest
it
as
close
to
the
local
disk
as
possible.
And
there
are.
B
B
B
First,
the
summary
of
what
we
have
discussed
is
data.
Have
the
applications
deal
with
multiple
volumes
like
we
have
seen
in
Hadoop
or
Cassandra,
every
volume
will
have
different
I/o
characteristics,
zookeeper
relating
to
sensitive,
whereas
the
data
volumes
for
data
logger,
throughput,
intensive
consolidation
of
multiple
applications
onto
a
single
kubernetes
cluster
makes
the
problem
even
harder,
and
some
of
these
applications
have
built
in
replication
like
Modi,
Vista,
Sandra's
and
majority
of
the
no
sequel
applications
do
replication
by
themselves,
so
that
makes
the
allocation
tricky.
So
what
are
we
looking
for?
B
How
can
we
plant
the
computer
on
the
storage?
So
all
the
applications
are
protected,
so
we
are
looking
for
applications
where
storage
provisioning
stack.
When
cassandra
is
asking
for
three
volumes,
the
storage
stack
should
understand.
The
three
volumes
are
related
that
cassandra
treats
these
three
volumes
as
special.
The
story
stack
needs
to
know
about
it,
but
in
the
current
state
of
kubernetes,
the
CSI
doesn't
know
that
these
two
volumes
are
related.
The
calls
that
we
get
are
create
a
volume
amount
of
volume.
B
A
B
A
B
B
B
B
Now
we
take
a
snapshot
again,
so
we
capture
all
the
data
changes
in
the
snapshot.
Now
this
is
where
things
get
tricky.
Let's
see
if
you
were
to
change
the
password
in
Coober,
it
is
the
password
changes,
the
two
places,
depending
on
how
the
entry
points
are
written,
it
is
possible
that
there
secret,
let's
change,
two
places
now,
let's
take
capture
the
snapshot
again,
we.
A
B
B
B
B
B
Virtual
machine
world,
this
is
called
convict
rift.
The
configuration
of
the
system
has
drifted,
I
think
that's
not
captured
in
the
snapshot,
and
that
is
the
problem
that
existed
for
a
long
time
in
bare
metal
world
in
virtual
machines,
and
there
are
tools
designed
to
solve
these
problems.
Chef
puppet.
There
are
several
other
tools
which
do
configuration
management.
B
Is
there
an
alternative
to
this?
Can
we
protect
the
application
pattern
in
the
Google
DS
world?
Of
course
we
can,
let's
replay
the
whole
scenario.
They
created
an
application
after
the
creation.
I'll
take
a
snapshot
here.
This
is
the
initial
snapshot.
Let's
make
some
data
changes.
Add
some
cables
take
a
snapshot
again,
so
we
capture
the
entire
application
again
make
password
changes,
capture.
B
On
fig
changes,
a
snapshot
capture,
the
config
changes,
so
if
you've
notice
what's
happening
here
is
we
are
not
just
capturing
the
data,
what
we
are
capturing
the
configuration
which
resides
in
config
maps
and
secrets,
we
are
capturing
the
topology,
which
is
how
many
replicas
and
what
is
the
topology
of
the
application,
and
we
are
also
capturing
the
service.
So
we
capture
the
application
in
its
entirety.
All
the
building
blocks
of
the
applications
are
captured
in
these
snapshots.
B
B
B
B
You
will
have
a
web
server,
a
middleware
and
a
database,
so
in
that
all
the
layers
may
not
provide
snapshotting
capabilities
or
checkpointing
capabilities,
so
in
Cooper
it
is
if
you
were
to
deploy
the
entire
stack
and
if
the
platform
allows
you
to
take
a
snapshot
of
the
entire
stack
that'll
be
much
much
better.
Peace
of
mind
and
state
of
snapshots
are
prone
to
contractors,
issues
that
we
have
seen
there.
B
Quantic
map
secret
and
volume
on
a
different
tag.
Point
in
time
and
consistency
group
is
very,
very
critical,
and
this
is
a
core
of
some
multi-volume
snapshotting
philosophy.
When
we
snapshot
Cassandra,
which
could
be
using
20
or
30
volumes,
we
have
the
snapshot
of
all
the
volumes
in
one
shot.
Otherwise
you
will
have
consistency.
Issues
think
about
Oracle
RAC
Oracle
RAC
has
is
a
sin
which
is
not
file
system
based
you'd
have
multiple
tree.
B
B
B
So
what
are
we
looking
for
again
to
protect
the
applications
on
kubernetes?
We
will
need
application
level
snapshots,
not
just
the
data
snapshots,
so
that
is
the
takeaway
from
from
this
exercise.
So
if
you
look
at
to
two
big
aspects,
one
is
the
application
of
air
storage
provisioning
to
protect
the
applications,
and
you
also
need
application
level
snapshots
again
to
protect
the
data.
Heavy
applications
on
cobalt.
B
Let's
see
the
work,
so
that
is
built
into
Robin
here,
so
you
can
take
any
application.
You
can
create
a
Helen
based
application
on
cobalt
ease
and
with
one
command.
You
can
take
a
snapshot
of
that
application.
Again.
My
snapshot
I
mean
capturing
everything,
all
the
aspects
of
the
application
or
just
volume,
data
config
and
metadata.
B
B
We
can
also
push
the
snapshot
to
an
external
target,
to
our
cloud,
to
protect
the
the
application
again
and
we
can
restore
the
entire
snapshot
back.
So
these
are
the
flows
that
that
are
built
into
Robin,
but
the
concepts
in
general
will
apply
so
Robin
backups
are
full
fully
self-contained.
Like
I
said
we
capture
everything
and
the
entire
application
can
be
restored.
It's
not
just
the
data
or
one
or
two
tables
the
entire
application,
including
the
topology,
is
rolled
back.
B
So
let's
look
at
it.
Third
and
the
biggest
aspect
of
the
the
production,
which
is
hardware
and
side
phases.
You
need
to
protect
against
this
value,
node
values
in
size
failures,
so
we've
already
seen
some
form
protection
offered
by
the
the
proper
planning
and
snapshots,
but
this
is
the
third
phase
of
it.
So
why
do
we
need
backups?
B
For
the
biggest
reason
is
I
think
protecting
against
upgrades
and
when
the
application
is
running
and
also
for
compliance
and
auditing
reasons,
what
backup
allows
you
to
do
is
potentially
this
form
of
backup
allows
you
to
do.
Is
it
will
make
the
process
of
hardware
refresh
easy
data
center
migration
you're
not
locked
into
a
vendor,
because
as
long
as
you
have
access
to
backup,
you
can
recreate
it
somewhere
else.
B
So
how
do
backups
look?
Look
like
the
backups
are
based
off
of
snapshots.
Snapshots
are
local
snapshots
will
give
you
a
consistent
point
in
time
image
of
the
application.
We
have
seen
the
picture
on
the
on
the
left
side.
Here,
initial
snapshots
data
changes
past
no
change
the
context
changes.
These
are
snapshots.
These
are
four
snapshots
that
we
have
taken.
B
Backups
are
off-site,
they
don't
reside
on
the
same
media
as
the
original
data
or
these
snapshots.
So
we
can
register
a
cloud
repository
and
move
the
snapshots
to
the
cloud.
So
when
you
move
it
to
cloud,
we
have
to
worry
about
two
aspects:
time
and
cost
of
course,
so
we
want
to
move
them
to
a
cheaper
solution,
because
the
backups
could
be
just
a
compliance
or
audit.
B
B
You
always
go
for
incremental
backups
and
when
you
are
resurrecting
an
application,
which
means
you
take
an
application
in
the
you
take
a
backup
in
the
cloud
and
you
want
to
create
that
application
in
the
cloud
for
several
reasons,
one
you
want
to
migrate,
an
application
which
means
you
you
take
incremental
backups
put
it
in
the
cloud
when
it
is
time,
cut
the
power
and
create
the
application
in
cloud.
So
you
have
move
your
application
from
on-premise
into
cloud
or
you
are
an
application
running
on
Prem.
B
What
you
want
additional
compute
capacity
to
run
completed
this
joint
job.
Let's
see
every
quarter,
you
want
to
run
the
books
so
which
means
you
have
the
data
set,
but
you
want
additional
compute
capacity
to
crunch
the
numbers
at
the
end
of
the
quarter,
so
that
is
when
you
will
never
at
your
backup,
create
an
application
in
cloud
and
run
use
out
there.
So
this
is
one
of
the
benefit
that
the
backs
provide
in
this
form,
which
is
applicational
back.
So
how
do
we
save
time
during
this
process
of
cloning?
B
A
B
A
B
B
B
A
B
So
how
did
we
do
it?
This
is
the
Robin
architecture
overview,
so
we
take
kubernetes
as
the
centerpiece
for
orchestrating
the
the
pods,
which
is
compute,
and
we
quickly
realized
that
our
castrating
compute
is
not
enough
for
hosting
data
heavy
applications.
It
may
be
enough
for
stateless
applications
for
stateful
applications.
We
need
complete
control
over
storage
computer
network,
so
we
build
a
distributed
storage
stack.
There
is
a
volume
stack
with
all
the
enterprise-grade
features
that
you
expect
from
big
storage,
these
natural
clones
quality
of
service
thin
provisioning
compression
encryption.
So
we
had.
B
We
have
a
complete
storage
tab.
On
the
other
side,
we
leveraged
multiple
networking
stacks
any
way
we
build
a
programmable
network
stack,
so
that
gives
us
control
over
storage
networking
compute.
On
top
of
that,
the
biggest
P
that
we
have,
these
application,
workflow
manager-
or
you
can
call
it
like
section
manager
using
that
we
were
able
to
provide
one-click
department
of
the
application.
One-Click
protection,
which
is
snapshots,
one
click
load
scale
upgrade
and
back
up
all
the
workflows
are
handled
by
devotional
manager.
B
It
will
give
you
an
experience
of
our
application
is
in
service
for,
like
maybe
redshift
any
other
theater
this
office.
You
can
do
it
in
in
the
private
cloud
using
kubernetes
and
leveraging
containers
and
Robin
is
a
software
platform
only
and
you
can
deploy
it
on
any
of
the
native
kubernetes
platforms,
including
all
the
clouds.
It
appears,
as
you
Google
platform.
B
A
B
Of
events,
11
billion
security
events
are
ingested
and
online
today,
and
we
have
multiple
petabyte
scale.
Hadoop
clusters
running
in
a
single
server
farm
and
the
uniqueness
of
Robin
is,
we
can
run
oracle
RAC
on
containers
and
we
are
running
it
at
scale.
We
run
400,
rack
databases
and
one
of
our
customer
sites
on
Google
is.
B
So
bottom
line,
so
we
need
application
of
external
provisioning
application
level
snapshot,
application,
backups
and,
of
course,
you
need
a
storage
tag
that
provides
replication,
features,
encryption
and
compression
features
to
give
an
end-to-end
protection
for
the
application
applications
running
on
commodities,
and
that
is
Robin.
Are
you
and
thank
you.
A
All
right,
we're
gonna
open
it
up
for
questions.
Michael
Ryan
asked
a
question
about
15
minutes
and
it's
a
stateful
pet
sets
question
mark,
but
Michael.
If
you
could
resubmit
your
question,
maybe
with
more
context
that
would
be
really
good.
I'm
gonna
move
on
to
Roberto's
question
as
well.
He
asks
does
Robin
preserve
the
DB
consistency?
A
B
It
does
yes,
she
does
that's
unique
to
Robin
because
probably
has
a
concept
of
application
bundle.
You
can
think
of
it
as
helpless
Plus
application
bundle
is
the
construct
that
will
allow
you
to
plug
in
applications
like
a
box,
and
you
could
think
of
it
as
I'll
chat
with
application
lifecycle
hooks
after
the
application
is
created.
If
you
want
to
write
some
custom
logic,
let's
say
seeding
the
initial
tables
that
can
be
put
in
post
create
hook.
So,
after
all,
after
all
the
pod
creations,
we
go
out
on
the
force
play
at
home.
A
B
Maybe
very
expensive
station,
it
sure
so,
who
says:
does
it
is
the
first
initiative
taken
by
kubernetes
to
solve
the
data
heavy
applications?
But
that
is
not
a
complete
solution,
because
what
stateful
said
did
was
it
offloaded?
The
storage
question
to
somebody
else,
which
is
the
CSI
attack
in
handle
very
cleanly
handled
the
compute
part
with
additional
primitives
like
no
definitely
no
Dante
affinity
and
all
that?
But
when
it
comes
to
storage,
stable
state
hooks,
it
has
limited
control
on
it.
B
It
cannot
dictate
where
the
block
should
come
from
and
there
were
additions
made
in
kubernetes
like
local
persistent
volumes,
but
that
also
didn't
solve
the
entire
problem,
because
local
persistent
volume
will
take
over
the
entire
time.
But
that
goes
against
the
concept
of
consolidation,
so
that
also
didn't
solve
the
problem.
B
B
A
B
I
have
already
completion
of
the
question,
but
il
electric
booth
Raman
can
run
on
vias
and
it
can
use
any
storage
that
is
plugged
into
virtual
machines.
So
it
is
a
Street,
the
virtual
machine
as
as
a
host,
and
it
can
on
top
of
that,
like
it
can
run.
It
appears
there
is
a
unique
value
prop
on
AWS
and
and
any
sanitized
storage,
because
Robin
can
actually
orchestrate
the
block
devices
attached
to
the
hosts.
For
example,
it
can
orchestrate
EB
as
one
it's
attached
to
a
double
ace
instances.
B
B
So
you
can
ideally
I
think
in
order
to
avoid
the
conflict.
Rick
drift
problems,
you
you
want
to
revert,
revert
the
entire
application,
which
is
using
a
single
come
on
Robin
to
restore
or
Robin
rollback.
If
you
are
doing
it
from
backup,
you
do
a
Robin
Esther,
if
you're
doing
from
a
snapshot
to
do
a
Robin
roll
back.
But
if
you
want
to
restore
a
particular
object
there.
A
B
B
You
can
create
a
thin
cloud,
take
the
components
that
you
need
and
delete
the
clone,
and
if
you're
talking
about,
let's
say
hypothetical
case
you
just
you
want
to
create
or
restore
a
particular
namespace
in
in
Cassandra
or
you
want
to
create
a
clone
of
a
particular
namespace
or
a
key
key
space
in
Cassandra
again
in
the
lifecycle
hook
comes
into
place
here
you
want
to
clone
a
Cassano
database.
What
do
you
want
a
particular
portion
of
row?
You
want
to
do
data
masking,
then
you
will
say
you
will
write
a
post.
B
Lo,
no
Robin
will
create
a
clone
of
the
application
and
it
will
trigger
the
post
loan
hook.
So
we
should
go
and
do
data
masking.
So
many
of
the
things
will
be
enabled
by
the
drawing
application
bundles
and
it
is
a
vvl
writing.
L
chat
or,
in
fact,
more
easier
easier
than
writing
am
chatting
politely.
A
B
We
do
yes,
we
do,
and
that
is
one
of
the
interesting
aspects
robin
has.
Is
it's
a
very,
very
fine-grained
policy
engine?
It
supports
our
application,
you
can
set
the
fall
domain
for
replication.
You
can
say
my
replicas
need
to
be
on
two
different
drives
or
loads
or
racks
or
data
centers.
That's
one
level,
and
the
other
aspect
of
that
is
I
want
my
compute
and
storage
to
be
coming
from
the
same
node,
which
is
what
we
call
compute
and
storage
affinity.
B
But
so
you
can
always
have
so
in
a
consolidated
storage
place,
the
software-defined
storage
plane
there
were
blocks
could
be
coming
from
anywhere.
Robyn
provides
a
policy
which
says
if
my
data
load
is
running
on
a
physical
server.
One
make
sure
you
get
the
data
blocks
from
that
same
server,
so
that
privileges
you
can
set
storage
and
computer
affinity,
anti
affinity
which
will
file
storage
and
computer.
You
can
set
fault
domains
for
application
and
you
can
set
computer
and
affinity,
policies
and
round-robin
policies.
B
B
A
B
B
A
A
B
A
B
Good
question,
so
that
is
a
secret
sauce
in
a
vein,
but
I'll
give
you
some
other
way,
so
we
use
Lee
since
electives
as
as
one,
and
there
are
many
components,
because
if
you
look
at
Cobras
objects,
it's
very
they're
very
loosely
coupled
entities,
you
cannot
tie
all
the
objects
together.
You
need
to
look
at
labels
and
selectors.
You
need
to
get
some
annotations
and
you
need
to
look
at
object
mappings
in
some
case,
so
we
do
all
of
that
and
we
constitute
an
application.
B
So
you
should
use
a
big
help
out
to
create
multiple
state,
two-thirds
replica
assets
and
deployments,
and
you
give
Robin
a
directive
which
says
Robin
app
register
and
you
can
give
labels
and
selectors
or
you
can
you
say
this
is
my
Helen
chart
probably
will
go
crawl
and
figure
out
the
connections
between
all
the
objects
and
create
an
application.
Boundary
view.
A
B
B
A
Okay-
and
that
is,
we
are
caught
up
on
questions
now.
So
let's
give
people
another
few
minutes
here
we
go.
Eric
wants
to
know
up.
That's
just
a
reminder
that
you
can
keep
on
asking
questions
everyone.
We
still
have
plenty
of
time
here.
Eric
wants
to
know.
How
do
you
support
rwx
access
mode
for
the
PV.
B
A
B
A
A
All
right
great
well,
thank
you,
Ravi
for
a
great
presentation
and
for
answering
all
the
user
questions.
Thank
you
for
joining
us
today.
This
webinar
and
the
recording
and
the
slides
will
be
online
later
today,
at
the
usual
places,
we
are
looking
forward
to
seeing
you
in
a
future
CN
CF
webinar,
and
with
that
thanks
everyone
and
have
a
nice
day.
Thank
you.