►
Description
OpenShift Commons Briefing
June 17th, 2020
In this OpenShift Commons Briefing, we get preview of a new cloud-native data protection platform for Kubernetes. In this webinar, Prashanto Kochavara (Trilio) gave an overview of the cloud-native market and the challenges with traditional data protection technology. He then introduced a new application-centric backup and recovery product that was designed from the ground-up to support the scale, performance and mobility requirements of Kubernetes container environments across any public or hybrid cloud environment.
Speaker Bio:
Prashanto Kochavara is Director of Product at Trilio, a leader in cloud-native data protection. Kochavara is responsible for the product strategy, roadmap and execution of the company’s backup and recovery platform for Kubernetes.
A
Our
guest
today
is
Prashant
Oh,
a
cup
of
evaru,
who
is
going
to
talk
about
Trilly
of
all
four
kubernetes
and
Trillium
is
a
new
open
ship,
Commons
member
and
the
way
this
whole
works
is
Prashant.
O
is
going
to
talk
and
give
a
presentation
he's
going
to
do
a
bit
of
a
demo
as
well,
and
then
we'll
have
live
Q&A
at
the
end
of
this.
So
without
any
further
ado
Prashant.
Oh,
please
introduce
yourself,
tell
us
all
about
what
Trilly
o
is
doing
and
what
you're
doing
with
truly
evolve.
B
Great,
thank
you.
Dan
Mosley,
were
you
excited
to
be
here
at
the
open
chef,
Commons
briefing,
my
name
is
Prashant
a
kookaburra
and
the
director
of
product
for
kubernetes
at
Tranio,
and
today
I'm
going
to
be
presenting
trillion
world
for
kubernetes
I'm,
going
to
be
talking
about
the
product.
The
architectural
functionality
use
cases
compatibility
and
be
following
up
with
a
recorded
demo,
to
give
you
guys
an
idea
as
to
how
the
product
works.
B
So
the
agenda
for
today
is
I'm.
Gonna
talk
a
bit
about
the
company
rally
up
who
we
are
and
what
we
have
been
doing
and
then
we'll
dive
into
the
customer
problem.
You
know
talking
about
legacy
data
protection,
solutions
versus
cloud
native
data
collections
and
why
one
is
needed
more
today.
Next,
we
will
give
an
overview
of
the
product
that
we
have
built,
which
is
truly
a
wall
for
kubernetes,
and
then
we
will
follow
up
about
the
technical
details
of
the
product.
B
Jumping
right
in
into
the
company
overview
who
we
are
so
Trillium
was
founded
in
2013.
We
are
the
leading
data,
protection
solution
for
OpenStack
and
Rab
environments.
We
have
enterprise
customers
located
all
over
the
globe,
along
with
our
offices
as
well.
We
are
back
I
leading
tech,
luminaries
and
venture
capitalists.
We
have
been
in
the
business
for
the
past
seven
years,
so
we
have
been
doing
data
protection
for
a
while.
So
we
have
come
up
with
a
lot
of
different,
innovative
technologies
that
we
have
attended
as
well
and
then
from
a
partner
ecosystem.
B
We
we
are
that
had
an
IBM
partner,
we've
been
partners
for
a
very
long
time
and
I've
been
doing
business
also
for
a
very
long
time,
and
as
mentioned
over
here
on
the
slide,
we
are
and
it
had
certified
father
as
well
and
from
a
platform
perspective.
We
have
products
to
do
data
protection
of
OpenStack.
As
I
mentioned,
we
have
products
to
do
data
protection
of
that
add
virtualization,
and
now
we
are
pivoting
to
address
the
data
protection
needs
of
a
cloud
data
environment.
B
Now,
let's
talk
about
the
customer
challenge
and
to
do
that
I'm
going
to
illustrate
by
applications
and
how
they
have
evolved
over
time.
So
back
in
the
day,
we
had
bare
metal
servers
with
one
operating
system
and
applications
running
on
those
operating
systems.
These
applications
were
tightly
integrated
when
I
say
tightly
integrated.
That
means,
like
all
the
components
of
that
application,
were
running
within
the
same
domain.
B
Then
we
moved
into
the
virtualization
world
when
we
moved
into
virtualization,
we
did
not
change
the
architecture
of
an
application.
Application
still
work.
The
same
way
were
tightly
integrated.
You
know
all
the
components
were
tightly
integrated
running
within
the
operating
system
boundary.
But
what
changed
was
the
efficiency
of
the
underlying
hardware?
So
virtualization
did
not
change.
The
architecture
brought
an
efficiency
of
the
overall
resources
within
your
datacenter.
B
Now,
when
we
get
into
a
cloud
native
world,
what
we
see
is
that
an
application
is
broken
down
into
multiple
micro
services
right
all
these
micro
services
are
independent
components
that
run
within
an
environment
which
have
their
own
identity
and
in
order
to
protect
an
application,
all
these
individual
components
need
to
be
protected
together.
They
need
to
be
backed
up,
not
only
the
not
only
the
data
per
portion
of
it,
but
also
the
metadata
portion
of
it.
B
B
You
need
a
cloud
native
application
to
provide
data
protection
for
cloud
native
applications
running
in
your
environment,
for
example,
you
cannot
have-
or
it
would
be-
really
costly
to
run
a
VM
based
data
protection
technology
to
be
protecting
a
complete
cloud
native
environment.
So
you
would
have
to
have
a
virtualization
or
VMs
running
to
protect
your
cloud
media
when
it
couldn't
be
a
right
fit
with
traditional
interpretation
technologies.
You
also
have
an
application
or
as
dependents
that
we
spoke
about
there
siloed.
They
are
focused
on
monolithic
applications.
B
These
applications
do
change
footprint,
but
it
is
very
rare.
They
are
generally
focused
on
the
storage
or
the
data
volumes
and
they
do
not
focus
on
the
metadata
lay
out
the
topologies
and
so
on
and
from
a
rural
perspective
or
a
persona
perspective.
These
traditional
solutions
are
focused
more
towards
you
know:
infrastructure,
admins
versus
developers
or
their
box
admins.
And
when
we
look
at
the
cloud
native
applications,
as
we
mentioned,
they're
modular
micro-services
oriented
each
has
its
own
identity
they're
highly
available.
They
are
highly
scalable,
they
change
their
footprint
on
demand.
B
B
The
first
we
believe
in
the
Ling
our
product
as
an
agentless
product.
We
do
not
insert
any
agent
within
a
virtual
machine
if
you're
protecting
a
virtual
machine
by
our
OpenStack
products,
or
we
do
not
have
any
site
cards
to
protect
your
pods
and
your
workloads
and
then
open
chef,
Donna,
kubernetes
environment.
We
are
completely
multi
cluster
or
multi
tenant.
We
adhere
to
the
principles
of
our
back
of
whatever
platform
we
work
with,
we
believe
in
the
concept
of
self-service,
UI
integration.
B
We
are
non-disruptive,
so
there
is
no
destruction
of
your
existing
workflows
or
existing
applications.
When
you
install
kilovolt
or
when
you
create
really
want
in
your
environment,
then
we
follow
an
open,
universal
backup
scheme.
Of
what
this
open
universal
backup
schemer
does
is
not
only
it
avoids
vendor
lock-in
but
you're
free
to
use
your
data
after
it's
backed
up
to
do
additional
work
flows
and
the
opening
was
the
backup
schema
that
we
use
be
talking
more
about
it
and
for
the
slides,
provides
lot
of
data
efficiency
features
as
well.
B
So
you,
you
are
able
to
eliminate
a
hardware
deduplication
appliance
because
of
the
innovative
style
of
our
data
is
copied
into
the
target
because
of
the
underlying
format.
Apart
from
that,
there
are
additional
features
around
granular
file
system
recovery
that
can
be
achieved
with
this
open,
Universal
backup
schema
as
well.
B
So
going
clockwise.
Firstly,
we
are
application-centric,
so
we
focus
on
the
application
layer.
We
protect
not
only
the
data
volume
but
also
the
metadata
of
all
the
kubernetes
objects
that
an
application
comprises
of.
We
are
native
to
kubernetes
and
OpenShift,
so
we
are
built
on
the
cube
api
server,
so
you
do
not
need
any
other
CLI
or
api
to
manage
the
kubernetes
environment
or
the
trillionth
environment
within
kubernetes.
B
We
are
deployed
as
customer
resource
definitions
that
you
can
use
to
manage
the
overall
product
now
from
an
application
deployment
and
ecosystem
tooling
perspective.
What
radio
world
can
protect
are
your
applications,
whether
they
are
deployed
by
a
health,
whether
they
are
deployed
by
our
operators
or
whether
they
are
deployed
by
labels
or
just
have
you
know,
custom
label
tags
to
it.
We
also
integrate
into
Prometheus
power
monitoring.
We
have
dashboards
available
in
the
fauna
and
our
logs
are
integrated
into
flu
Indy
as
well.
B
Jumping
onto
the
left
side,
infrastructure
compatibility,
we
leverage
the
CSI
mechanism
to
talk
to
storage.
So
as
long
as
the
storage
has
a
CSI
tribal
available,
we
will
be
automatically
supporting
that
storage
platform
from
a
target
location.
As
to
where
your
backups
are
gonna
be
stored,
we
support
NFS
and
s3,
as
the
underlying
protocol
s3
can
be
on
Prem,
s3
based
storage
or
can
even
be
Amazon
s3
as
the
target
and,
more
importantly,
we
are
a
certified
technology.
We
are
certified.
We
are
certified
operator
with
an
open
chest.
B
We
have
cloud
pack
for
data
and
MCM
or
multi
cloud
management
as
well.
Now,
all
this
we
want
to
make
it
easy
for
our
customers
to
you,
know,
understand
the
product
and
drive
it
and
then
work
it
in
the
add
an
environment
as
well.
So
what
we
have
done
is
we
have
provided
live
environments
directly
through
our
website
that
you
can
come
and
play
with
30
of
all
four
kubernetes
and
also
from
a
licensing
perspective.
We
have
free
and
basic
license
available.
The
free
license
gives
you
unlimited
number
of
nodes
for
30
days.
B
Talk
about
telework
for
kubernetes
being
that
had
operators
certified,
so
we
are
found
directly
within
operator
hub.
So
if
you
go
into
operator
have
today
and
you
search
for
Trillium,
we
will
be
listed
as
an
application
or
an
operator
based
application
that
you
can
install
within
our
OpenShift
environment
and
after
you
install.
This
is
how
they
would
look
as
part
of
the
demo.
We
will
be
touching
a
bit
more
upon
the
cosmetic
of
the
product
next
I'm,
going
to
talk
about
the
packaging
and
accessibility.
So
we
have.
B
Next
I'm
going
to
quickly
talk
about
the
overall
architecture
of
the
product
and
how
trader
world
for
kubernetes
has
been
constructed,
but
at
the
very
first
layer
it
is
the
user
interaction.
This
is
where
the
user
creates
his
custom
resources
for
backup
plans,
targets
backups
and
restores
within
the
next
layer,
which
is
the
control
plane
layer.
B
We
have
the
control
install
for
all
our
custom
resource
definitions
which
monitor
the
custom
resource
definitions,
and
if
there
are
any
changes
to
them,
they
make
or
apply
those
changes
within
the
cube,
API
server
and
what
the
control
plane
does
is
that
it
helps
to
capture
the
metadata
from
the
application
and
it
transfers
that
metadata
into
the
target
location,
which
could
be
a
NFS
or
an
s3
repository
I
mentioned,
and
then
finally,
we
have
the
data
plane
within
the
data
plane
is
where
the
actual
copy
transfer
happens.
Of
the
persistent
volumes.
B
B
After
that
you
have
the
policy,
a
policy
can
be
scheduling,
policy
or
a
retention
policy
which
says
how
often
to
take
the
backup
and
how
how
many
backups
to
keep
as
per
your
compliance
needs,
and
then
we
also
have
the
concept
of
hooks.
So
if
you
have
stateful
were
closed,
like
databases,
you
can
us
the
database
and
on
qss
by
using
the
pre
and
post
hooks.
Now
all
these
three
custom
resources
are
referenced
within
the
back-up
plan.
B
Back-Up
plan
is
the
overall
definition
of
what
you
are
backing
up,
where
you're
backing
up
to
how
often
you
are
backing
it
up
and
if
there
are
any
intricate
injections
that
are
needed
as
part
of
the
application
or
multiple
applications
that
will
be
backing
up
the
thing
to
know
to
be
heard
is
you
know
the
backup
plan
can
be
a
single
hell.
Map
can
be
a
single
operator,
app
can
be
a
single
label
based
app
or
can
be
multiple
held
multiple
label,
multiple
operators
or
can
be
a
combination
of
helm
label
or
operators.
B
So,
depending
upon
whatever
you
want
to
protect,
it
could
be
one
helm
application
talking
to
another
helm
application
which
is
talking
to
an
operator
based
application.
You
can
define
all
of
that
within
the
back-up
plan.
So
if
you
were
to
backup
your
namespaces
completely,
you
can
create
a
back-up
plan
which
defines
all
the
applications
within
it
and
it
would
be
able
to
backup
all
of
that
together.
B
For
you,
the
next
custom
resource
that
we
have
is
the
backup
custom
resource
in
which
you
provide
whether
your
backup
should
be
a
full
backup
or
an
incremental
backup,
and
obviously
it's
a
schedule
based
policy.
Then
it
would
be
using
that
schedule
to
take
that
back
up
periodically
as
well
and
once
you've
taken
your
backup.
The
next
operation
to
do
would
be
to
restore
now
your
restore
scan
happen
based
on
the
name
of
the
backup,
which
would
be
if
you,
within
the
same
cluster
or
same
namespace
or
you
can
restore
by
location.
B
Next
I'm
going
to
talk
about
the
protection
and
recovery,
so,
as
mentioned,
we
protect
the
metadata
of
the
application,
as
well
as
the
ecosystem
volumes,
and
we
do
that
for
home
operators
and
labels.
So
what
happens
is
now
you
can
back
up
your
application
from
one
name
space
into
another
name,
space
within
the
same
cluster
or
you
can
backup
your
application
from
one
cluster
to
another
cluster
and
a
completely
different
namespace.
B
B
Now,
let's
talk
about
what
does
truly
work
for
kubernetes
backup,
so,
as
mentioned,
we
backup
labeled.
Now,
if
your
applications
just
have
their
custom
applications
with
just
a
label
tag
on
them,
we
will
we
can
back
that
up.
If
your
applications
are
based
on
hell
week
and
back,
you
have
applications
up
or
if
they
are
operator
based
applications.
We
can
do
that
as
well,
when
we
do
the
label
based
backup.
B
What
we
do
is
we
look
at
the
spec
of
all
the
resources
and
we
back
up
the
spec
portion
of
it
and
all
the
resources
that
the
application
comprises
of
so
whether
it
is
pods
PVS
can
fake
map
secrets
we
capture
all
of
that,
every
PV
that
we
find
we
capture
the
data
out
of
the
PV
and
store
it
within
our
target
location
as
a
cube,
how
to
format
as
well
on
the
helm
side
of
it.
We
backup
all
revisions
of
your
application,
including
the
deployed
revision
of
the
release.
B
Then
we
parse
the
chart
we
identify,
which
are
the
persistent
volumes
and
we
back
up
those
process
in
volumes
as
well.
The
key
point
to
remember
here
is
that
when
we
back
up
and
restore
or
hand
based
application
using
Trello
board
for
kubernetes,
we
maintain
the
application
time.
So,
for
example,
a
helman
when
restored
is
still
a
helm
application.
You
can
still
use
your
helm
upgrades
your
hell
and
roll
back
commands
to
manage
that
application
after
it
has
been
restored
now
in
the
operator
world.
B
What
we
do
is
we
backup
the
resources
of
the
operator,
as
well
as
any
custom
resources
created
by
the
user
for
the
operator,
and
then
we
parse
are
any
application
resources
that
have
been
provided
or
that
have
been
created
and
we
backup
the
application
as
well.
This
is
the
application
that
is
managed
by
the
operator
and
again,
as
we
mentioned,
when
we
do
back
up
and
restore
the
operator,
it
is
still
an
operator
based
application.
B
So
you
do
not
lose
your
consistency
of
application,
tooling
that
you
had
originally
deployed
at
wire
next
I'm
going
to
talk
about
the
overall
backup
flow.
This
is
an
animated
slide,
so
it
should
help
everyone
to
understand
what
is
the
underlying
workflow
that
we
leverage
to
protect
a
particular
application?
B
So,
first,
what
we
do
is
a
metadata
backup,
so
in
order
to
do
the
metadata
backup,
the
first
thing
that
we
do
is
we
spin
up
a
meta
mover
pod.
What
this
meta
move
aboard.
Does
this
captures
all
the
metadata
information,
which
is
the
deployments
the
services,
the
config,
match
the
secrets
and
it
moves
it
into
the
target
location.
B
Then
we
look
at
the
data
backup
or
we
look
at
the
persistent
volumes
that
we
need
to
protect
and,
as
part
of
the
data
backup
we
spin
up
the
data
move
abroad.
We
spend
up
one
data
move
apart
for
each
persistent
volume
so
that
we
can
get
parallelism
in
terms
of
the
speeds
and
everything
that
is
being
moved
into
the
target
location.
B
We
take
a
snapshot
of
that
persistent
volume.
We
convert
that
snapshot
into
a
new
persistent
volume
and
then
we
mounted
or
attach
it
to
the
data
mover
pod,
so
that
the
data
mover
pod
can't
read
from
it
does
can
convert
to
a
cue
cow
to
format
and
stores
it
on
the
target
location.
We
detach
the
persistent
volume
from
the
data
mover
pod
and
then
we
delete
it.
We
keep
that
first
snapshot
that
was
originally
taken.
B
We
still
keep
that
around
for
incremental
backups
and
to
do
a
compare
of
the
next
snapshot
that
will
be
taken
for
the
incremental
backup
so
now
diving
into
the
workflow
for
incremental
backup.
What
we
do
is
again
first
part
that
is
fun
up
is
the
meta
mover
pod.
The
meta
mobile
pod
will
again
capture
all
the
metadata
information
and
put
it
into
the
target
location.
B
Then
we
get
into
the
data
backup
phase
and
as
part
of
the
data
backup.
As
mentioned
again,
we
will
be
spinning
up
one
pod
for
each
persistent
volume
we'll
be
taking
a
new
snapshot
leveraging
CSI,
we
will
be
converting
those
snapshots
to
persistent
volumes
again,
attaching
them
to
the
data
mobile
pod.
B
B
Restore
is
basically
an
inverse
of
how
we
do
the
backup
application,
so
we
take
the
incremental,
assuming
that
we
are
doing
a
restore
of
one
of
the
incremental
backups
that
we're
taking
real.
First
on
the
data
mobile
pod,
we
will
create
a
CSI
volume.
We
will
do
a
restore
of
the
pv
we'll
attach
that
pv
again
to
the
data
mover
pod,
and
then
we
do
a
chemo
image
convert
again
to
the
restored
PV
again,
the
TV's
are
detached
from
the
mobile
pod
to
complete
the
data
portion
of
the
operation.
B
Next
is
the
metadata
restore
where
we
spend
off
the
meta
mover
pod
for
the
meta
meta
data
restore
the
application.
Metadata
is
again
restored
back
into
your
namespace
or
particular
cluster,
and
then
all
the
application,
specs
and
TVs
pvcs
are
restored
to
the
restored
PB.
Obviously,
everything
is
done
by
the
CSI
interface.
B
Talk
about
the
use
cases
over
here
so
based
on
how
true
the
world
can
do
its
backup
and
restore.
There
are
a
plethora
of
use
cases
that
are
able
so
basic
backup
and
recovery.
You
can
schedule
on
the
mountaintops.
You
can
have
full
and
incremental
backups.
You
can
restore
into
a
new
cluster
or
into
a
new
name
space
into
an
existing
cluster,
and
you
will
have
full
or
selective
app
restore
as
well.
B
Now,
if
you
are
doing
disaster
recovery,
you
can
use
the
same
technology
to
take
your
data
and
recreate
it
at
another
cluster.
Again,
your
clusters
can
be
hosted
on
Prem.
It
can
be
in
the
cloud
as
long
as
it's
a
kubernetes
cluster.
You
will
be
able
to
be
a
disaster
recovery,
application
mobility.
You
know
from
a
testing
perspective
or
from
a
CI
CD
perspective.
B
And
now,
as
mentioned
today,
is
more
mostly
a
hybrid
world
where
people
have
on-prem
deployments
of
resources
or
kubernetes
clusters,
and
they
have
public
flower
deployments
of
the
kubernetes
cluster
and
various
distributions.
So,
in
order
to
maintain
your
costs
or
to
avoid
when
the
Lockean,
you
can
again
usually
evolve
for
kubernetes
to
take
that
data
from
one
environment
and
move
it
into
another
environment,
and
you
can
maintain
a
solid
TCO
and
get
more
for
your
buck.
B
Next
I'm
going
to
talk
about
monitoring
and
logging,
so
Julia
Walt
is
integrated
completely
into
Prometheus
and
prefer
know.
What
we
do
is
all
the
metrics
from
our
custom
resource
definitions,
whether
they
are
Target
backups,
backup
plans,
three
stores
we
send
all
those
metrics.
We
have
defined
a
few,
a
bunch
of
metrics
and
we
send
them
to
the
Prometheus
server.
We
do
the
visualization
through
prefer
nodes,
and
what
we've
done
is.
B
We
are
providing
almost
about
10
ports
as
part
of
the
product
which
will
be
useful
to
monitor
each
and
every
instance
of
an
application,
or
instance
of
a
backup
whether
in
the
Australia
want
environment.
So
we
will
provide
about
10
boots,
but
users
should
be
free
to
create
your
own
boots
as
part
of
the
Anglophone
instances.
Well,
well,
from
a
logging
perspective,
we
are
integrating
into
fluent
D.
So
all
your
logs
across
all
your
namespaces,
whenever
you
use
Chile
award
for
kubernetes,
will
be
available
within
fluently.
B
B
Now,
talking
a
little
bit
about
the
metrics
that
we
expose,
so
there
is
informational
objects
or
informational
items
that
we
provide
around
the
backup
plan
target
backups.
We
have
information
around
the
status
as
to
how
how
much
of
the
backup
has
completed
the
duration
of
the
backup
to
understand
what
the
transfer
speeds
were.
You
also
have
how
much
backup
storage
is
a
particular
backup
using
and
how
much
least
storage
is
available
on
a
particular
target.
B
Also
in
force
information
about
restore,
you
know
how
long
did
the
restore
take
if
the
restore
has
completed,
and
finally,
we
have
health
information
about
our
controllers
as
well.
So
not
only
can
you
manage
monitor
your
Trilly
of
Ault
for
kubernetes
application
itself,
but
you
can
manage
and
monitor
your
backups
of
your
other
applications
within
your
kubernetes
clusters.
B
Talk
about
the
security
and
permissions
so
from
a
security
perspective,
Trulia
world
does
not
require,
or
does
not
depend
on
any
admin
access
or
any
cluster
admin
roles
or
privileges
that
that
exists
generally
within
a
kubernetes
cluster.
We
leverage
the
existing
security
context,
constraints
that
are
provided
by
our
open
shift
and
from
a
service
account
perspective.
We
provide
an
additional
layer
of
security
by
creating
our
service
accounts
at
run
time
or
at
execution
time.
So,
as
a
result,
those
service
accounts
can
not
be
used
anywhere
else.
B
We
have
our
own
set
of
permissions
and
capabilities
that
we
define
as
part
of
that
as
power
of
the
product,
and
all
these
definitions
are
provided
in
our
publicly
hosted
gate
book
documentation,
which
is
a
docstore
chileo,
no
dialogue.
We
are
also
working
on
building
a
much
more
intensive
and
detailed
product
definition
page,
which
talks
about
all
the
security
aspects
of
the
product.
You
know
in
terms
of
how
the
application
is
built,
how
the
application
is
distributed.
So
all
all
intricate
stuff
and
useful
stuff,
our
customers
will
be
provided
within
our
user
documentation.
B
So
what
I'm
gonna
do
here
is
have
a
record
demo
of
how
we
do
a
backup
and
restore
or
film
applications
on
OpenShift,
and
what
I'm
going
to
be
doing
here
is
I'm,
going
to
be
showing
the
tli
or
the
non
UI
side
of
it
and
I'll
also
be
showing
the
OpenShift
UI
to
show
what
the
CLI
implications
were.
So
as
part
of
this
environment,
we
are
running
open,
chef
4.2,
we
are
using
heav'n
version
3,
we
are
using
the
open
shift
operator
for
a
storage.
B
We
are
just
using
a
host
pass,
CSI
driver
and
as
part
of
this
overall
them
or
what
we
are
going
to
be
doing
is
we
are
going
to
be
deploying
just
a
simple
demo
app.
We
are
going
to
be
backing
up
that
app
by
a
truly
award
for
kubernetes,
and
then
we
will
restore
it
and
make
sure
that
whatever
we
have
backed
up
is
still
available
and
running.
B
So,
as
you
see
first,
what
we're
going
to
be
doing
is
we
will
be
deploying
the
application,
which
is
a
hell
Mac,
and
once
we
have
deployed
the
application,
we
will
do
a
Helen
list
to
see
that
the
application
has
been
created
properly.
We
can
see
the
application
pods
that
have
been
deployed,
as
part
of
the
hand
come
on,
and
what
we'll
do
is
we'll
ensure
that
the
application
and
the
PVCs
are
all
up
as
part
of
the
deployment
and
then
finally,
we
will
ensure
that
the
app
is
running.
B
B
B
Next,
what
we'll
do
is
we
will
go
into
the
OpenShift
user
interface
and
just
confirm
that
we
have
our
backup
created
over
there
successfully
so
in
the
OpenShift
user
interface.
If
you
go
into
operator
hub
and
look
at
the
install
of
leaders
in
our
case
over
here
for
this
demo,
we
are
at
the
OpenShift
marketplace
and
you
can
see
it's
really
a
world
for
kubernetes
has
been
installed,
and
now
we
have
the
back-up
plan,
which
defines
the
Helen
application
created
moving
back
into
the
CLI
portion.
B
What
we
will
do
is
we
will
trigger
a
backup
of
that
backup
plan
or
the
protection
template
that
we
had
specified
and
next
we
will
do
a
get
backup
to
see
the
status
of
the
overall
backup
operation.
You
can
also
get
the
details
of
whatever
is
happening
as
part
of
the
backup
operation
to
confirm
this.
On
the
UI
side,
we
go
into
the
backup
tab
and
we
see
that
a
backup
has
been
created.
B
B
It
was
a
full
backup,
they'll
start
and
end
time
available
for
the
backup
as
well,
and
then
what
we
will
do
is
we
will
first
ensure
that
the
namespace
that
we
are
going
to
restore
this
back
up
into
does
not
have
any
work
load
running
already,
so
the
restored
namespace
is
empty,
as
you
see,
and
we
will
also
so
first
we
apply
three
store,
and
then
we
will
do
a
get
on
the
restore
to
see
that
okay,
the
restore,
has
started.
It
is
in
progress
and
the
objects
are
being
copied
over.
B
We
can
move
over
to
the
open
ship.
Ui
I
can
go
into
the
operator
operator
instance.
That
was
installed
an
open
shelf
marketplace,
and
we
can
look
at
the
these
stores
over
here
now
and
now
we
can
see
that
a
restore
has
started.
It
is
in
the
validation
step,
so
basically
we
first
validate
that
everything
should
be
able
to
restore,
and
then
we
start
the
restored.
B
So
once
the
restore
is
in
progress,
you
can
describe
the
restore
object
to
see
what
is
happening
as
part
of
the
overall
process
and
then
once
the
restore
has
completed,
you
can
do
again
a
helm
list
to
see
that
your
app
has
been
restored
successfully,
as
mentioned
earlier,
when
we
do
a
backup
of
any
application,
whether
it
is
a
Hell
map,
location
or
an
operator
based
application.
The
maintain
the
deployment
tooling
consistency
so
because
it
was
a
helm
application,
you
can
still
use
a
list
to
list
the
restored
app.
B
We
can
also
verify
that
all
the
pods
have
been
restored
correctly,
along
with
all
the
other
objects
like
deployments.
Replica
sets
persistent
volume,
names
TVs
themselves
and
then
finally,
final
test
would
be
to
expose
the
newly
restored
app
via
a
service
and
around,
and
then
we
can
ensure
that
the
app
has
successfully
been
restored
and
do
a
new
name
space.
B
B
So,
as
part
of
the
demo,
I
also
want
to
kind
of
move
you
to
a
quick
idea
of
how
we
also
can
be
manipulated
or
managed
to
via
the
UI.
So
this
is
a
photo
for
environment
that
I
have
in
front
of
me.
We
have
philia
work
for
kubernetes.
We
can
use
the
Yama's
within
OpenShift
to
manage
or
create
any
custom
resource
definitions,
and
we
also
conform
to
the
dynamic
form
startup
provided
by
our
open
ship.
B
B
Next
I'm
going
to
talk
about
compatibility
and
support,
so
truly
world,
as
we
mentioned,
has
been
the
ground
up,
leveraging
the
cube,
API
server
and
we
integrated
Atlee
into
you,
know
the
kubernetes
concepts
and
the
overall
architecture.
So
there
is
nothing
new
that
you
need
to
learn
or
understand
to
manage
or
operate
failure,
walls
or
kubernetes.
B
Now,
because
we
are
built
a
ground-up
addressing
or
aligned
with
the
kubernetes
constructs,
you
can
use
us
in
any
upstream
kubernetes
environment
that
supports
CSI,
so
CSI
was
supported
in
upstream,
at
least
from
one
node
1
2
onwards,
and
from
1.12
to
1.16.
The
CSI
feature
is
in
alpha
stage,
so
what
we
recommend
is
for
those
environments
to
use
trailer
for
kubernetes
for
test
f
purposes
and
then
once
when
CSI
did
move
into
beta
and
1.17,
you
can
use
that
for
production
as
well
from
an
open
shift
perspective.
B
Obviously
4.1
to
4.3
was
when
the
CSI
feature
was
introduced
on
board
one
to
four
or
three
started
using
coupon,
IDs,
1.1
or
and
higher
and
in
those
environments.
The
CSI
is
a
feature
for
snapshots
we're
still
in
alpha,
so
we
recommend
to
use
it
for
tests
of
environments
and
approach
from
4.4
onwards.
B
From
a
storage
perspective,
as
I
mentioned,
we
are
completely
agnostic
to
the
underlying
storage.
As
long
as
you
use
a
CSI
driver
to
manage
your
storage
environments,
TSI
is
the
de
facto
protocol
that
is
going
to
be
used
moving
forward
for
working
with
storage
by
kubernetes
interface.
So
we
are
aligning
ourselves
also
to
kind
of
keep
all
those
pieces
agnostic
and
ask
ourselves
to
be
able
to
move
and
maneuver
with
any
storage
path.
B
From
a
target
perspective
in
our
demo,
I
showed
you
that
we
had
an
Amazon
s3
target,
but
you
could
have
any
s3
compatible
storage
which
would
be
hosted
on
Prem
as
well,
and
also
we
support
file
based
storage.
That
is
why
our
NFS,
so
if
you
have
an
NFS
server
on
Prem
or
even
in
your
cloud
environments,
you
can
use
that
as
your
target
for
storing
ecosystem
tooling.
We
have
spoke
about
our
dashboards,
which
are
available
by
Accra
fauna.
B
The
first
thing
is
cloud
native
applications
are
completely
different
than
traditional
applications.
There
is
a
lot
of
independent
components
that
need
to
be
managed
and
protected,
along
with
the
metadata
and
the
data
and
traditional
approaches
that
we
have
all
been
used
to
before
a
cloud
native
world
to
not
satisfy
the
challenges
or
the
requirements
of
cloud
native
world
properly
we
are
till
award
for
kubernetes
is
purpose-built
as
an
operator
to
protect
loud
native
applications.
As
I'd
mentioned,
we
can
protect
your
helm
operator
or
label
based
applications.
B
We
are
kubernetes
native.
We
fit
into
the
ecosystem
by
a
Prometheus
of
an
affluent
deep.
We
provide
cutting
edge
features
in
terms
of
open
backup
formats,
not
having
the
need
to
do
or
buy
a
deduplication
appliance
for
your
target
storage,
along
with
a
lot
of
security,
fundamental
principles
that
have
been
applied
within
the
product.
B
As
a
result
of
these
features,
we
enable
a
plethora
of
use
cases
on
trailer
world
for
kubernetes
and
I'd,
mentioned
backup
and
restore
from
point
in
time,
recovery
perspective
or
to
protect
yourself
from
threats
like
ransomware
or
data
corruption
in
general.
You
can
use
it
for
disaster
recovery
for
moving
from
one
custard
to
one
other.
B
You
can
use
it
for
CI,
CD
pipelines
or
for
tests
of
application,
migrations
from
one
cloud
environment
to
another
flower
environment,
from
a
security
perspective
or
from
a
cost
perspective,
and
a
key
point
to
remember
here
is
that
trader
world
has
been
around
for
a
long
time.
We
are
the
leader
in
terms
of
data
protection
for
OpenStack
and
rad
virtualization
environments
already,
so
a
lot
of
the
copy
transfer
protocols,
the
underlying
way
of
how
you
take
data
from
one
environment,
move
it
into
another
environment,
has
then
matured
very
well
over
time.
B
And
with
that,
I
would
say
that
that
would
invite
all
of
you
to
try
out
to
evolve
for
kubernetes.
Today,
you
can
actually
log
into
our
website,
which
is
trailer
on
IO.
You
can
watch
a
demo
of
the
product.
You
can
run
a
test
drive
the
same
demo
that
I
wish
spring.
You
can
actually
run
it
in
your
flame
that
freetime
directly
from
the
trail
your
website,
without
having
any
infrastructure
created
for
yourself,
and
then
you
can
download
a
free
trial
or
a
basic
edition.
B
So
as
mentioned,
we
want
to
provide
developers
and
IT
administrators
an
equal
opportunity
to
use
the
product
and
their
own
scopes
and
in
their
own
landscapes.
So
we
have
a
basic
and
free
edition
or
the
free
edition
which
provides
unlimited
nodes
or
30
days
and
then
basic
Edition,
which
provides
10
nodes
for
an
unlimited
amount
of
time
and
then
finally,
we
have
the
enterprise
license
as
well,
which
provides
support
to
all
these
cutting-edge
features
that
trailer
one
for
kubernetes
provides.
B
A
Well,
Prashant,
oh
thank
you.
We
I
I
knew
that
I
had
seen
the
OpenStack
offering
that
you
had
and
I
had
not
seen.
That's
an
in-depth
thing
around
what
you
were
doing
with
kubernetes
and
OpenShift,
and
so
I'm
really
overwhelmed
with
like
how
awesome
it
is
to
be
and
I
feel
like
I'm
gushing
or
something,
but
it
I
kept
thinking
to
myself
man.
What
didn't
you
think
of,
but
you
answered
in
the
last
slide.
A
One
of
one
of
the
questions
that
came
in
from
the
YouTube
was
about
being
able
to
do
the
workshop
on
their
own,
and
so
the
you
know,
I
think
I
encourage
everybody
to
go
over
the
Trilly
Oh
site
and
take
a
take
a
look
at
their
online
lab
there
as
well.
There
is
one
couple
of
questions,
and
one
of
them
is:
can
backups
be
leveraged
with
third-party
tools.
For
example,
can
I
security
scan
backups,
okay,.
B
So
as
because
we
keep
our
backups
in
an
open
format,
which
is
a
coupon
to
format,
you
can
leverage
them
as
you
want
to
with
external
tools.
Scanning
them
connect
the
esterday
target,
and
you
can
do
all
those
intricate
post
Cup
operations
as
well.
So
because
we
keep
it
in
an
open
forum
and
you
are
able
to
do
all
of
that.
B
So
you
can
define
so
if
you
we
have
basic
scheduled
cron
jobs
that
you
can
define
as
per
the
back-up
plan,
the
cron
jobs
would
obviously
maintain
a
RP
o.
Not
an
RTO
RTO
would
be
more
based
on
your.
You
know
how
you,
in
an
organization,
what
is
your
SLA
in
terms
of
getting
the
application
backing
up,
but
from
an
RTO
perspective.
That
is
what
we
provide
through
policy
scheduling.
A
C
Prashant
Oh
Chris
short
here,
yes,
I
I,
want
to
like
thank
you
and
then
drive
home
the
importance
of
stayin,
kubernetes
native
right,
like
if
you
stay
in
that
Lane,
you
open
yourselves
up
to
all
the
availability
of
the
entire
ecosystem.
Can
you
talk
about
the
importance
of
staying
cloud
native
and
kubernetes
negative
a
little
bit
to
the
trillion
plan
form?
B
Because
everything,
the
concept
of
kubernetes
is
to
you
know,
make
sure
that
you
are
focused
on
deploying
applications
and
you
know
time
to
market
with
applications
for
customers.
That
is
what
kubernetes
is
kind
of.
You
know
that's
the
kind
of
the
idea
or
the
principle
behind
it,
so
there
are
a
lot
of
products
which
are
interfacing
with
the
central
API
and
are
leveraging
that
central
api,
which
is
the
cube
api
server
to
build
their
technology.
B
So
if,
as
a
product
you
aligned
with
that
central
api
that
and
our
community
kubernetes
native,
then
you
automatically
open
yourself
to
integrate
with
all
these
other
products
as
well.
If
you
are
not
integrated
or
if
you're
not
kubernetes
native,
then
what
will
happen
is
you
will
have
to
have
another
translation
layer
to
make
your
product
work
with
another
third-party
I
see
or
any
other
vendor
product?
B
So,
as
a
result,
in
terms
of
how
we
have
built
our
technology,
you
know
integrations
into
Prometheus
integration
into
flow,
and
these
become
actually
very
simple
and
tomorrow,
if
there
are
any
other
ecosystem
products
that
a
user
would
want
to
integrate
into,
it
would
be
very
straightforward
as
well.
If
they
are,
you
know,
just
leveraging.
Wyrley
api
is
and
managing
both
products
together.
C
B
And
that's
that's
one
of
the
reasons
right.
You
know
traditional
approaches
which
are
kind
of
vm
based
approaches,
state.
We
need
those
approaches
or
that
approach
needs
to
change
to
a
cloud
native
approach
so
that
you
can
have
a
complete.
You
know
just
kubernetes,
self-sufficient
environment
versus
having
good
on.
You
know,
kubernetes
on
one
side,
and
you
know
virtualization
on
the
other
side
and
stuff
with
that.
C
A
B
Thank
you
actually,
that's
a
great
question
for
me
to
eradicate
the
users
on
so
from
a
roadmap
perspective.
We
are
going
to
be
doing
that,
so
they
only
have
an
early
access
version
of
the
product
which
I
was
showing
you
then
my
demo,
which
is
available
then
operator
hub
and
for
upstream
environments.
Now,
by
the
end
of
this
month
or
early
July,
we
are
going
to
be
teeing
our
product
and
as
part
of
our
cadence,
our
target
is
to
push
out
code
every
month.
B
So
you
want
to
do
it
by
the
1st
of
every
month,
or
at
least
that's
our
target
or
pushing
of
new
features,
and
we
will
be
doing
that
cadence
from
July
on
words.
Now
from
a
future
perspective,
there
are
a
lot
of
cutting-edge
things
that
we
have
been
thinking
about.
In
terms
of
you
know,
providing
compression
encryption
and
one
of
the
most
important
things
that
I
would
like
to
talk
about
is
a
restore
plan
so
similar
to
a
back-up
plan.
You
know
how
you
specify
what
you
want
to
backup
as
part
of
your
application.
B
We
are
working
on
something
called
as
a
restore
plan
in
the
restore
plan.
You
can
basically
specify
how
you
want
your
application
to
be
laid
out
after
that
is
restored.
Are
there
certain
objects
that
you
want
to
manipulate
before
you
restore
it,
or
are
there
any
additional
workflows
that
you
want
to
inject
after
the
restore
has
been
done
so
that
actually
opens
up
additional
use
cases
and
additional
opportunities
for
customers
to
manage
their
data
and
manage
the
applications
in
a
much
more
succinct
fashion?
B
So
that's
one
of
the
cool
or
important
things
that
we
are
working
on
and
the
cooler
portion
of
it
at
least
who
I
feel
is
about
the
more
impactful
from
a
customer
perspective.
Also
is
we
will
have
a
separate
user
interface,
but
really
evolved
for
kubernetes
so
that
if
you
have
multiple
trailer
world
or
kubernetes
instances
and
running
across
namespaces
or
multiple
clusters
in
general,
you
will
have
a
single
source
of
truth
to
manage
all
those
tvk
instances,
so
that
will
give
the
ability,
for
you
know
doing
much
more.
B
A
B
Yes,
so
we
as
part
of
our
GA
when
we
do
GA,
we
are
going
to
be
flung
back
satisfied.
We
are
going
to
be
called
for
data
certified
and
we
as
a
stretch
goal.
We
are
also
targeting
MCM
certification
and
then
we
would
be
integrating
into
MCM
too
so
that
that
three
layers
or
the
three
steps
that
we
have
kind
of
figured
out
on
our
and
how
we
will
be
doing
the
integration
first
will
be
mostly.
You
know:
basic
UI
SSO
based
integration
followed
up
with
much
more
detailed
ground-level
integration
between
the
two
products.
I.
A
Think
one
of
the
things
that
Chris
and
I
were
chatting
separately
here
is
that
we'd
love
to
have
you
come
back
in
during
maybe
not
during
the
briefing
session,
but
actually
do
a
live
coding
have
couple
of
the
developer
and
evangelist
team
create
a
cluster
crack.
You
have
it
install
the
backup
and
restore
with
Trilly,
oh
and
then
walkthrough
restoring
lives.
So
if
you're
up
for
that
kind
of
a
challenge,
we'd
love
to
have
you
back
to
do
something
along
that
line?
Yeah
be
really
fun.
Yeah.
B
Definitely
what
we
could
probably
set
up
is,
you
know
we
can
do.
Label
based
and
base
operator
based
backups
throw
some
things
how
they
are
looking
in
from
its
disco
sauna,
along
with
logs
I,
think
that
should
provide
a
good
idea
of
how
the
flows
works,
and
you
know
basically
putting
that
over
to
the
gold
as
well
yeah.
A
Awesome
that
would
be
awesome,
because
this
is
like
I
kind
of
it's
like
magic.
It's
like
you've
thought
of
pretty
much.
Everything
and
I
can't
wait
to
see
it
in
action
live
as
well,
and
when
you
get
to
GA
we'll
have
you
guys
back
and
we'll
make
sure
that
this
is
something
that
we
all
can
take
advantage
of?
It's
wonderful
to
see
it
in
operator
hub
and
operator
hub
do
already
I
really
encourage
everybody
who's
listening
in
whether
it's
here
in
the
blue
jeans
or
YouTube,
or
twitch
or
Facebook,
to
take
a
look
at
this.
A
A
What
L
love
those
categories?
They've
done
lots
of
great
stuff
with
us
and
the
partners
and
other
comments
members,
but
it's
wonderful
to
have
you
guys
as
part
of
the
OpenShift
commons
and
in
operator
hub
and
looking
forward
to
the
GA
date,
we'll
all
celebrate,
but
in
the
interim
please
do
everybody
take
it.
Take
it
for
a
test,
drive
and
looking
forward
to
seeing
how
this
rolls
out
in
the
future.
So
thanks
again
for
taking
the
time
Prashant,
oh
and
for
the
other
folks
from
Trilly
Oh
who
are
online,
answering
questions
Justin
and
Carl.