►
From YouTube: Webinar: What’s New in Kubernetes 1.11
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 1.11 continues to advance maturity, scalability, and flexibility of Kubernetes, marking significant progress on features that the team has been hard at work on over the last year. This newest version graduates key features in networking, opens up two major features from SIG-API Machinery and SIG-Node for beta testing, and continues to enhance storage features that have been a focal point of the past two releases. The features in this release make it increasingly possible to plug any infrastructure, cloud or on-premise, into the Kubernetes system.
A
Welcome
everyone
welcome
to
the
CNC
webinar
series.
My
name
is
Roderick
ski
I
am
a
developer
advocate
at
cloud
native
computing
foundation,
feature
slated
a
cuban
artists,
wandered
11,
early
steam,
and
today
we'll
speak
about
cuban
I
just
wanted
to
live
and
latest
release
the
newest
one
from
the
cabinet's
release.
Jim.
It's
like
this.
A
A
B
A
B
Another
filter
is
a
framework
provided
a
pioneer
Ascanio
as
either
allows
customer
Catherine,
messaging
or
network
networking.
Riveter
appearances
such
as
packet
filter
and
of
tutoring
net
in
other
ISM,
is
another
worker.
Just
revision
for
the
traditional
Islam,
so
also
everything
things
ok
with
the
IPA
was
butter.
I
believe
doesn't
perform
very
well
in
larger
us.
You
not
just
air
factor.
C
B
B
There
are
some
optimizations
to
add
famously,
but
today
I
want
to
introduce
them.
I
will
I
will
introduce
a
IBPS
mode
that
you
deposit
before
going
through
the
people
walking
through
the
act.
It
was
murder.
Firstly,
let's
have
a
look
at
as
I
previously
was
abbreviated,
I
believe
the
same
as
I'm
famous
it's
peuta
owns
the
top
of
near
future
yeah.
D
B
B
Its
design
high
previous
offender
for
know
the
penalty
not
for
fair
war
and
the
IP
is
the
chorus
reload,
apparently
load,
balancing
murder
the
net
for
the
PR
motor
and
IP
Connery
model.
Only
net
murder
supposes
poor
translation,
so
in
karate
we
uses
net
mode
also,
not
very
fast,
but
now
the
modal
is
also
very
fast.
B
B
They
are
more
not
apparently
a
grinning
Olivia,
such
as
through
the
wrong
looping,
sauce,
harsh
Hoshi
destination
Hoshi
and
leave
it
all
those
for
the
icons
such
as
face
down
mr.
Lee,
Stallone,
okay,
mister
crashing
and
so
on.
We
can
also
attend
the
with
her
to
eat
your
servers,
so
so
could,
with
the
run
lubing
with
the
least
the
connection,
Salam
yeah
IBS,
also
supporter
server
highest
treatment.
That
nation
try
is
a
native
recorder,
big
station,
a
30
finalists.
As
we
stated
my
next.
B
B
You
we
needed
to
notice
actually
the
one
two
three
four
five
okay
fabric
in
the
mothers,
the
I
did
a
strainer
and
the
serrated
promo
things
as
I
read
is
model
is
not
as
still
another
for
the
model
it
winning
to
switch
the
from
forty
more
than
280
PS
and
the
making
sure
we
enabled
it
as
feature
anyway
before
modern
pants
before
Omaha
is
still
a
bus
station.
After
one
pollen
hand
it's
in
beta,
so
it's
enable
a
backbone
support.
Ip
has
promoted
to
true
sessions.
B
Service
installation,
firstly,
make
sure
to
meet
advice
to
me.
Device
interface.
Existing
know
that
the
the
for
the
name
is
Cooper.
I
did
yes,
I
did
yes,
you
know
so
time.
A
second
right
find
the
service
IP
address
to
the
to
me
device
for
carsabi
type
of
services.
The
server
IP
is
the
car
key
and
a
for
know.
The
father's
dimensional
it
does
address
the
kind
is
a
transaction
and
sorry
create
a
BBS.
The
virtual
servers
for
each
service,
IP
professor,
is
IP
and
the
crater
real
servers
for
each
endpoint.
B
This
is
the
last
picture
of
this
year:
okay,
first
light
weight
great
service
in
him,
the
changing
services,
its
IP,
a
sec
here
is
the
boy
I'm
on
the
a-10
to
anger,
and
it
opens
a
portal
War
three
zero,
eight
zero
and
has
to-
and
upon
you
see.
Firstly,
we
can
see
here
is
the
Cooper
IPS
rain.
A
bit
is
0.
This
is
the
to
me
device
and
we
find
it
repented
as
the
Cossack
II
choose
to
meet
us,
hey
another
proposal
where
creature,
a
beautiful
service
and
service
service
and
anaconda.
D
A
A
E
D
E
A
CNC
F
incubation
project,
it's
used
as
a
plug-in
chain
architecture,
meaning
that
DNS
requests
come
in
and
it
passes
through
a
series
of
plugins
and
each
plug-in
performs
a
function
on
that
request.
It's
flexible.
You
can
combine
plugins
together
for
advanced
functionality
and
it's
extensible.
It
allows
you
to
provide
a
framework
for
writing.
E
So
that's
what
accordion
s
is
standalone,
but
how
does
it
fit
into
kubernetes?
There's
one
plug-in
called
kubernetes
that
enables
accordion
has
to
provide
DNS
based
service
discovery
and
a
kubernetes
cluster.
It
follows
the
DNS
based
service
discovery,
specification
that
is
defined.
It
can
replace
Kubb
DNS
because
it's
functionally
equivalent
and
core
DNS
is
an
approved
alternative
to
Cooper
coup
d,
NS
and
kubernetes
1.11.
E
Cordia
ness
fixes
a
few
issues
when
compared
to
coup
DNS
there's
some
open
issues
that
are
include
DNS
that
are
resolved.
Some
of
these
are
featured
requests.
Some
of
these
are
our
bugs.
I
won't
go
through
all
of
them,
but
like
allowing
custom,
DNS
entries
for
Koob
DNS
the
little
bug
like
missing
a
records
in
a
certain
case.
E
B
E
Three
containers
for
DNS
has
one
from
a
metrics
perspective.
They
both
report
metrics
to
Prometheus,
but
the
set
of
metrics
differ.
Part
of
the
reason
is
because
of
the
one
container
versus
three
from
a
configuration
perspective,
that's
the
biggest
change
you'll,
see,
there's,
there's
migration
tools
available
for
converting
the
Kubb
dns
config
accordion
s.
Accordion
s
is
fully
configurable
by
the
config
map.
Kudiye
dns
is
not
that's
another
difference
next
slide.
E
Here's
an
example
of
accordion
s
configuration
trying
to
keep
it
short
here,
so
I
won't
go
into
all
the
details,
but
the
bottom
section
there
it's
pretty
much
the
what
you'd
see
in
a
default
install
from
cou
bad
men.
The
top
section
is
an
example
of
what
a
stub
domain
looks
like
that's,
basically
saying
everything:
everything
that
comes
in
on
port
53
for
example.com
gets
proxied
out
to
one
two
three
four
and
we
cashed
it
for
300
seconds.
E
E
D
E
E
A
F
E
E
So
I
mentioned
that
it
were
built
on
plug-in
chain,
and
so
there
are
a
lot
of
built
in
plugins,
there's
34
when
I
counted
last.
A
few
interesting
ones
are
when
and
ones
that
are
used
a
lot
or
file.
It
allows
you
to
serve
a
zone
from
a
zone
file
rewrite
it
allows
you
to
rewrite
requests
as
they
come
in,
allows
you
to
do
kind
of
behind
the
scenes.
E
Aliasing,
that's
invisible
to
the
client
and
template
it's
great
for
debugging,
for
example,
it
allows
you
to
define
responses
based
on
a
regex
and
go
template
and
if
you've
got
other
needs
that
the
plugins,
the
built-in
plugins
don't
don't
solve,
you
can
build
your
own
plugins.
There's
some
cool
external
plugins
that
other
people
have
built.
Unbound
uses
the
unbound
library
to
do
recursive,
DNS,
PE
sequel
does
little
serve
records
from
a
back-end
database.
E
E
E
There's
some
links
here,
if
you
have
any
questions
or
issues
or
one
support,
we're
very
active
in
github
and
in
slack
for
any
security
related
issues
sent
to
security
at
core
DNS
at
I/o.
Each
plug-in
is
documented
and
you
can
find
that
documentation
on
the
core
D&S
website
and
also
thanks
to
John
Bell
America.
Pretty
much
took
99%
of
the
contents
of
this
presentation
from
his
blog,
which
I've
liked
here
next.
E
A
Progress
coordinates
made
during
the
last
like
a
spirit
of
time
from
there
from
the
easy
replacement
for
communities
built
in
DNS
or
works
the
default
one
days
days.
So
this
is
networking.
Another
major
error
in
kubernetes
is
not
functionality
and
I'd
like
to
might
help
him
from
signal
we'll
call
it
dynamic.
You
bleed
configuration
feature
hi.
C
So
why
why,
first
before
we
get
into
exactly
what
it
is,
why
do
we
want
to
do
something
called
dynamic,
qubit
configuration
so
just
for
some
history,
kubernetes
is
built
around
the
idea
of
offering
declarative
API
is
that
are
hosted
in
a
central
control
plane.
That's
your
API
server
in
kubernetes.
You
can
use
these
api's
to
configure
most
things
about
the
system
unless
the
thing
you're
trying
to
configure
isn't
already
represented
by
a
core
abstraction
of
that
system.
So,
like
most
things
run
in
pods,
but
most
cominius
deployments
don't
run
the
Kiewit
in
a
pod.
C
Some
do,
but
most
don't
so.
By
lifting
the
cubic
configuration
into
the
controller
plane,
we
make
it
more
visible
and,
in
many
cases,
more
convenient
to
manipulate,
but
there
are
certainly
ways
to
dynamically
configure
a
qubit
in
communities.
Today,
some
people
just
ssh
into
nodes
and
manually
mess
with
things.
Some
people
use
their
cardi
configuration
management
tools,
but
really
none
of
these
ways
are
built
with
career
IDs
in
mind
and
they
don't
have
a
first-class
user
experience
around
them.
I
can
have
the
next
slide.
Please.
C
So
what
is
dynamic?
Qubit
configuration
well
in
kubernetes
110,
we
launched
the
ability
to
configure
the
qubit
via
a
structured
version,
kubernetes
style,
config
file
API
as
opposed
to
using
command
line
flags,
and
so
to
use
this.
You
simply
pass
the
path
to
this
file
to
the
qubit
and
in
111
we
launched
this
dynamic
cubic
configuration
feature
to
beta,
which
is
basically
just
a
way
to
deliver
this
config
file
in
a
live
cluster
to
a
live
node
and
have
the
qubit
switched
to
using
the
new
configuration.
C
So
you
use
the
exact
same
structured
version,
curry
style,
config
file
format
as
the
file
you
would
normally
just
write
locally,
but
you
can
post
this
config
file
into
the
canaries,
control,
plane,
vi8
sigmap,
and
then
you
can
tell
nodes
to
refer
to
that
config
map
that
contains
your
config
file
and
once
you
tell
the
node
to
refer
to
it,
the
qubit
will
download
that
config
and
restart
to
use
the
new
configuration
can
have
the
next
slide.
So.
C
So
once
that
field
is
updated,
the
cubelet
associated
with
that
node
object
is
maintaining
a
watch
on
that
node,
and
so
it
will
see
the
update
and
it
will
then
go
find
that
reference
config
map
download
it
unpack
the
files
in
it
and
store
that
locally
as
what
we
call
a
config
checkpoint
internally.
Once
that
configures
check
pointed
EQ,
it
will
reach
start
to
try
the
new
config
checkpoint
and
then,
depending
on
what
happens,
the
Hewett
will
update
the
node
status
and
say
yes
I'm
using
this
new
config.
C
It's
great
everything
is
happy
or
the
new
config
was
didn't
pass.
My
validation,
for
example,
so
I'm
gonna
report
near
here
and
I'm
just
going
to
fall
back
to
the
checkpoint
I
had
previously
until
you
fix
whatever
the
problem
was
so
that
I
don't
disrupt
your
workloads,
so
this
workflow
extends
to
multi
node
rollouts.
Instead
of
simply
updating
the
spec
for
one
node,
you
update
it
for
all
the
nodes.
You
want
to
refer
to
the
new
config
map
because
you
have
control
over
which
nodes
referred
to
that
you
can
stay
written.
You.
C
So
there's
a
lot
of
detail
on
this
slide.
This
is
an
example
of
the
actual
objects
in
the
control
plane
that
we're
dealing
with.
So
you
can
see
on
the
right.
We
have
the
config
map
that
contains
here
and
the
slide
and
truncated
cubic
configuration.
The
cubic
configuration
space
is
actually
quite
large,
there's
roughly
100
fields
and
on
the
left.
We
have
the
note
object,
that's
being
configured
in
this
example
and
in
the
spec.
C
For
that
note
object,
you
can
see
that
there's
the
config
source
config
map
field
and
in
that
field
we're
just
specifying
the
symbolic
information,
that's
necessary
to
refer
to
the
config
map
and
also
note
that
we're
saying
which
specific
key
the
cubic
config
file
of
the
config
map
is
in
and
if
you
follow
the
red
arrows,
those
match
up
with
the
config
map
there
and
then
in
the
status
of
the
note
object.
You'll
actually
see.
C
There's
three
sort
of
config
source
representations
there
and
those
are,
you
can
think
of
those
as
the
three
things
that
Kubik
cares
about.
So
the
first
one
is
simply
what
the
qubit
believes
it
has
been
assigned
to
use
and
that's
sort
of
like
an
AK
like
once.
The
qubit
has
downloaded
a
new
checkpoint
and
recorded
locally,
which
checkpoint
it's
supposed
to
use
and
come
back
up.
It
will
report
that,
so
you
can
see
that
it
at
least
acknowledged.
C
Oh
I
saw
you
reconfigured
me
the
next
one
is
called
active,
and
that
is
the
configuration
checkpoint
that
the
qubit
chose
to
use
at
runtime
for
the
current
runtime
duration
and
the
so
the
active
can
actually
be
equal
to
either
the
config
source
you
assigned,
or
possibly
the
last
known,
good
configuration
invalid,
for
example.
So
then,
the
final
piece
here
is
that
last
known,
good
configuration.
C
D
C
Good,
it's
it's
also
reported
in
the
status
and
then
finally,
you
can
see
the
config
map.
In
this
example,
slide
has
an
error.
Qapi
QGIS
is
set
to
a
string
that
is
not
valid,
and
so
here
that
error,
the
fact
that
there
is
a
validation
error
is
being
reported
in
the
status
as
well
for
pretty
much
all
the
validation
errors
and
other
kinds,
errors
that
could
be
reported
in
the
status.
C
In
addition
to
that,
the
fact
that
there
is
no
real
history-
that's
easy
to
query
for
like
Oh.
Give
me
this
resource
version
of
this
resource
means
it
would
be
very
hard
to
debug
what
the
actual
intent
was
in
a
live
environment.
If
you
do
require
versions
on
your
configuration,
we
recommend
that
you
use
you
know
inject
version
information
into
the
name
of
the
config
map
that
you
post
and
ideally
treat
that
config
map
as
immutable.
C
So
we
don't
live
in
a
perfect
world,
so
there
are
plenty
of
gotchas
with
all
kinds
of
things.
One
is
that
the
configuration
API
is
a
little
bit
low
level,
so
a
because
we're
using
the
same
file
format.
This
lets
you
set
the
whole
config
dynamically,
but
you
know
this
is
powerful
and
you
should
be
careful.
Some
things
are
very
safe
to
change
like
UPS
some
things.
I
shouldn't
touch
like
the
names
of
C
groups.
There
are.
C
There
is
inline
documentation
at
the
link
in
this
slide
for
per
field
advice
on
which
things
you
should
be
more
careful
with.
So
if
you
follow
that
link,
you
can
find
more
information
and
just
as
a
reminder,
the
intent
of
this
feature
is
the
target
system,
experts
and
service
providers
who
can
qualify
these
kinds
of
things
and
understand
the
system
very
well.
C
Another
gotcha
is
that
we
are
moving
to
this
config
file,
API
from
command
line
flags,
and
so
there
are
still
a
lot
of
command
line,
flags
that
are
deprecated
but
not
yet
removed
and
can
still
specify
the
same
information
is
in
the
config
file
for
backwards
compatibility
reasons.
These
flags
take
precedence
if
they're
still
specified
on
your
command
line.
C
C
So
these
are
just
some
thoughts
on
future
work,
as
I
noted
in
the
previous
slide.
There
are
a
lot
of
low-level
knobs
in
configuration
today.
This
is
the
case
for
both
the
cubit
and
a
lot
of
other
components.
Most
users
actually
want
high-level
policy
when
users
tune
different
configuration
parameters,
they're
not
just
doing
it
like
for
the
heck
of
it.
They
actually
have
an
end
goal
they're
trying
to
optimize
for
something.
C
So
you
know
what
open
question
is:
can
we
have
or
offer
high
level
opinions,
or
you
know
named
strategies
that
satisfy
common
use
cases
and
to
new
cases?
You
know:
are
there
high-level
principles
that
work
for
everyone
that
we
can
just
bake
into
our
components
instead
of
exposing
knobs
to
tune
things,
the
higher
level
we
can
make
these
configuration
surface
areas,
the
friendlier
the
system
will
be
to
non-experts
and
also
experts?
C
So
another
item
is
that
like
dynamic,
qubit
configuration
is
a
manual
per
node
configuration
process,
so
there's
no
node
pool
type
orchestration
built
in,
but
the
greys
cluster
API
effort
is
focused
on
solving
this.
Among
many
other
issues-
and
you
can
follow
that
link
to
find
out
more
about
that-
and
of
course
we
are
continually
migrating
away
from
flags
in
today's
version
to
config
file
api's
and
then
just
one
more
slide
here.
C
Just
the
next
slide,
please
yep.
So
these
are
some
other
links
you
can
follow.
The
first
is
the
blog
post,
with
some
details
on
dynamic
Kubek
configuration.
The
second
two
are
the
official
cruise
documentation
for
using
the
feature
in
addition
to
the
file
based
config
API
feature.
This
third
link
is
a
dock
on
the
philosophy
behind
why
we
should
use
versioned
configuration
files
instead
of
command
line,
flag
api's.
C
So
if
you're
interested
in
some
of
the
rationale
behind
this
migration,
I'd
recommend
reading
that
and
then
the
final
dock
is
just
a
general
document
on
the
API
philosophy
behind
Peretti's,
given
that
you
know
we're
migrating
all
this
stuff
to
create
a
style
api's,
that's
one
of
the
primary
motivators.
Thank
you.
Everyone.
A
D
Hello,
everyone
payment
here,
resizing
volume.
This
is
something
we
we
introduced
in
1.8
and
we
are
introducing
as
beta
in
1.11
next
slide.
Please,
okay!
So
what
is
persistent
volume?
Processing
volumes
provide
the
storage
layer
in
communities
which
is
useful
for
reading
application
that
need
to
persist
their
data
and,
like
the
sample,
databases,
are
our
Prometheus
or
application
that
need
state.
D
There
are
three
building
blocks
of
purses
volumes
if
you
are
not
already
familiar
with
them
like
so
is
classes
they
are
like
they
are
like
temples
from
which
position
volumes
are
created.
Processing
volume
is
your
also
called
Peavey's
or
actual
cluster
scopes.
Volume
objects
which
tie
to
like
underlying
storage
and
then
there's
a
process
in
volume
claims
which
is
basically
a
user,
managed
object
which
user
can
use
to
create
persistent
volumes
again.
A
storage
class
in
position
volume
are
faster,
scoped
resources
are
possessed,
mode
claim,
is
name
space
resource.
D
That
is
something
that
claim
is
what
most
users
typically
work
with
next
slide,
please
so
what
problem
resizing
sols?
So
the
in
one
problem
is
that
once
we
provision
PV
and
PVCs,
they
remain
a
fixed
size
throughout
their
life
cycle
and
as
applications
that
use
them
need
more
so
as
it
becomes
essential
to
move
them
into
larger
previous
and
that
can
be
tricky.
You'll
have
to.
Typically,
you
have
to
like
manually,
create
PVCs
back
them
up
and
then
delete
order,
PVCs
and
then
restart
the
pod
that
we're
using
them.
D
D
So
what
does
resizing
person?
Volume
mean?
It
means
like
in
place,
expansion
of
volume
by
editing
the
process
and
volume
claim
object.
Any
user
can
go
ahead
and
edit
the
PVC
object
and
get
a
new,
updated
volume
that
has
new
size.
We
do
not
support
shrinking
or
reducing
size
of
precision
volumes
and
when,
when
we
say
resizing,
then
we
expand
both
the
remote
volume
object,
which
it
could
mean
like
EBS
volume,
or
it
could
mean
gcpd
or
safar
bitty
volume,
and
it
also
expands.
The
file
system
on
the
node
in
community
is
1.11.
D
D
So
how
do
we
expand
processing
volumes?
As
I
mentioned,
the
feature
was
introduced
as
B
as
alpha
in
1.8
and
1.11.
It's
going
beta,
but
question
admin
must
opt
in
for
for
this
feature
by
or
like
enable
expansion
of
PVCs
that
are
created
from
storage
class,
which
has
allow
volume
expansion
set
to
true.
Only
a
PVCs
created
from
the
storage
class
that
has
a
low
volume
expansion
is
allowed
to
be
expanded.
Other
PVCs
are
by
default,
not
allowed
to
be
expanded.
D
This
is
to
ensure
that
cluster
admin
has
hooks
enough
so
that
users
are
aware-
and
the
feature
doesn't
become
a
loyal
to
everyone.
Next
slide,
please.
So
how
do
we
expand
processing
volume?
It's
pretty
simple,
you
just
it
to
like
you
cut
off
edit
PUC
and
then
PVC
name
and
in
your
editor.
You
can
just
expect
resources,
requests
and
storage
like
you
can
see,
you
can
request
like
you.
Can
edit
the
size
make
it
from
by
default.
5
GB.
It
was
previously
and
I'm
requesting
10
GB,
so
it
will
expand
the
volume.
D
D
So
after
editing,
PVC
for
shared
file
systems
like
Lobsterfest
as
your
file,
the
if
you
have
a
part
that
is
using
the
PVC,
the
expander
storage
is
immediately
available
to
the
part.
There's
nothing
else
you
need
to
do.
The
part
should
have
the
new
capacity.
Everything
should
work
as
it
is
no
problem,
but
for
block
storage
volume
types
like
not
EBC,
sorry,
TBS
and
UCPD,
and
as
your
disk,
you
will
have
to
restart
the
part
so
that
the
volume
could
be
expanded
on
the
node
like.
D
D
Expanding
file
system
requires
restarting
the
pod
and
once
underlying
model
has
been
once
underlying
burden
has
been
resized,
so
the
steps
generally
be
do
is
like
edit,
the
PVC
to
decrease
more
space.
Once
underlying
volume
has
expanded
by
the
storage
provider,
then
PVC
will
have
this
condition
file
system
resize,
pending
condition
and,
as
you
can
see,
that
in
this
case,
has
this
condition
file
system
which
has
pending
and
then
wait
for
the
PVC
to
have
the
condition
and
least
at
the
part.
D
D
As
I
mentioned
earlier
in
coordinate
is
1
point
1
1
volume
resizing
gadgets
to
beta
as
a
result,
it's
enabled
by
default,
and
you
no
longer
have
to
enable
a
feature
gate,
but
when
it
is
1.11
also
introduces
a
new
feature
called
expand
in
use
persistent
volumes.
What
that
will
do
is
in
previous
slide.
You
could
see
that
we
have
to
restart
the
part,
delete
and
recreate
or
scale
up
and
scale
down
the
part
for
the
files
assembly
sizing
to
to
be
finished.
D
But
if
you
have
this
feature,
get
enable
which
is
alpha
feature,
then
you
don't
have
to
anything.
The
the
file
system
will
be
automatically
expanded
on
the
cubelet
and
and
the
new
size
will
be
able
to
the
part.
One
thing
to
keep
in
mind
is
that
automatic
file
system
expansion
only
supported
four
volumes
that
are
unused
by
a
running
part.
That
means
this
unit
is
itself
will
not
will
not
schedule
the
volume
to
a
node
so
that
it
can
be
reset.
D
It
can
be
resized,
but
no,
it
won't
do
that
you'll
have
to
for
the
automatic
online
expansion
to
work.
There
must
be
a
part
that
is
using
the
volume
and
next
slide
please.
So
we
are
also
working
on
enabling
volume,
expansion
for
CSI
there's
a
link
there,
and
then
we
have
1.12.
We
have
also
working
on
enabling
one
new
expansion
for
flex
and
we
are
trying
to
improve
the
general
stability
and
improving
the
file
system
expansion
support.
So
those
are
some
of
the
future
works.
We
are
planning
next
slide
and.
D
Have
like
we
have
some
like
the
persist,
volume,
documentation
and
in
communities
covers
volume,
expansion,
how
it
works
and
the
blog
post
and
there's
a
how
the
persistent
one
expansion
will
work
for
CSI.
There's
a
pull
request
that
you
can
have
a
look
at
it,
so
yeah
I
think
that's
all
from
me.
Thank
you.
A
C
I
know
the
answer,
so
I
can
answer
it.
So
typically,
the
qubit
in
most
communities
deployments
does
not
run
as
a
pod.
Q.
Proxy,
however,
does
I
can
also
answer
the
most
recent
question.
The
answer,
which
is
an
it
does
dynamic
cubic
config,
mean
that
we
run
qulet
in
a
container,
and
the
answer
is
no
dynamic.
Cubic
config
is
necessitated
by
the
fact
that
we
don't
run.
E
E
C
D
It
depends
like
when
we
expand
the
volume
and
like
what
user
has
requested.
Likewise,
a
user
is
requesting
10
GB
space
and
the
CEFR
bidi
was
already
after
in
GP
space.
Then
the
then
the
operation
is
basically
I,
know
of
and
not
know,
but
it
will
consider
the
expansion
is
successful
and
it
will
finish
the
entire
operation.
D
So,
but
if,
like,
let's
say
user
request
in
10
GB
and
the
underlying
safar
bidding
volume
has
8
GB,
but
you
have
used
only
4
GB
but
who
may
need
it
doesn't
know
that
there
is
still
for
a
week
freed
on
it.
So
it
will
still
try
to
expand
the
volume
to
10
GB,
that
user
has
requested
and
it
will
go
through
the
whole
resizing
the
volume
expansion
workflow
and
it
will
finish
the
resizing
process.
A
Ok,
the
next
question
is
marginal
question
and
about
the
performance
of
cabin
is
probably
someone
here
or
any
anyone
from
the
countries
and
just
may
have
an
answer:
did
any
performance
with
more
details
and
reports
available
for
production
use
urban
areas
like
what
is
the
maximum
number
of
physical
services
service
excuse
me
can
be
used
and
how
many
parts
it
supports?
How
many
VMs
can
be
used?
I.
A
F
So
we
so
we
don't
have
anyone
from
6
scalability
here
right
now,
but
there's
active
work
in
kind
of
expanding
that
and
building
tests
around
expanding
that
capacity.
There
is
a
link
that
I'll
pop
in
the
chat
around
building
large
clusters,
so
I'll
just
read
off
some
of
the
high-level
notes,
so
at
kubernetes,
1.11
kubernetes
supports
clusters
with
up
to
5,000
nodes,
more
specifically
with
support
configurations
that
all
the
following
criteria:
no
more
than
5,000
nodes,
no
more
than
150,000
total
pods,
no
more
than
300,000
total
containers
and
no
more
than
100
pods
per
node.
F
F
We
have
seen
we
have
seen
clusters
pushed
300,
500
pods
per
node
from
the
pods
per
core
perspective,
it's
more
arbitrary,
depending
on
the
applications
that
you're
trying
to
design
its
it's
more
of
a
let's
see
the
application
run
unbounded
and
then
decide
what
we
want
to
do
from
there.
That's
that'll
give
you
a
better
idea
of
what
your
benchmark
can
be,
but
we
do
I
can
say
from
the
scalability
perspective
that
we
do
run
tests
and
we
have
been
integrating
more
of
the
scalability
tests
within
the
test.