►
From YouTube: kubeadm office hours 2020-01-15
A
B
B
A
A
C
Hope
there
will
be
not
not
shift
on
this
date.
Sorry
I.
If
there
will
be
a
shift
on
this
date
that
this
would
be
difficult
to
manage.
Are
you
thoughts
that
they're
going
to
shift
it?
It
happened
in
the
past
that
the
release
date
shifted
by
one
week
and
this
overlaps
with
kuvakin.
Basically,
oh
yeah.
A
So
it's
unlikely
that
they're
going
to
shift
it
before
in
time,
but
forwards
in
time.
It's
possible,
maybe
like
occasionally
they
shifted
a
couple
of
days.
But
yes,
I
saw
this
as
well.
If
it
overlaps
with
Q
Khan
is
going
to
be
very
hard
for
the
release
team
for
us
for
the
Canadian
people.
I,
don't
think
it's
going
to
affect
us
as
much.
C
A
All
right,
I
wanted
to
ask:
do
I
have
any
so
I
wanted
to
do
three
ash
of
the
enhancements
repository,
but
I
want
to
ask.
Does
anybody
have
any
other
keeps
aside
from
the
windows
kept
we're
going
to
talk
about
later
the
cube
a
they're
moving
out
kept
Fabrizio
rösti?
Do
you
have
any
caps
to
be
merged?
I.
C
A
C
C
A
A
A
A
A
B
For
in
large
part,
because
of
the
lack
of
like
host
network
and
privilege,
containers
on
Windows
so
doing
things
like
installing
cube
proxy
and
installing,
like
CNI,
but
also
still
working
around
some
things
on
Windows,
that
aren't
done,
like
still
aren't
done
by
cubanía,
and
so
the
approach
that
I've
been
exploring
and
now
like
whoa.
So
I've
talked
to
sick
windows
about
this
and
have
general
approval
there.
But
it's
using
a
like
what
I'm
calling
like
a
privileged
process
proxy.
B
But
so
what
this
kind
of
means
is
that
we
need
to
revise
the
cap
to
reflect
this.
Like
the
cap
was
written
in
such
a
way
where
it
seemed
like
there
was
sort
of
a
like
there's
a
bunch
of
stuff
about,
like
maybe
review
when
we
get
to
beta.
Whether
will
use
daemon
sets
it's
kind
of
unclear
like
but
I
think
now
we
want
to
switch
it
to
be
a
lot
more
explicit
that
we're
gonna
have
this
approach
of
using
this
host
service
that
will
enable
privileged
operations.
B
B
B
And
then
one
thing
that
I've
noticed
just
like
that
I
think
hasn't
been
resolved
is
the
cubelet
configuration.
This
was
like
called
out
in
the
cap
that
cube
a
teeny
cube.
Idiom
needs
to
support
like
a
separate
windows,
cubic
configuration
and
I.
Don't
know
this
wasn't
this
doesn't
as
far
as
I
know,
it
wasn't
ever
done
in
the
like
wrapper
script
that
was
created
right
now,
a
bunch
of
like
the
cubed
is
the
cubed
service.
That's
created
just
has
a
bunch
of
like
windows,
specific
command
line
flags
and
we
ignore
the
cubic
configuration
config
map.
A
Don't
think
that
we
should
integrate
all
of
that
into
comedian,
yet
so
there
still
has
to
be
a
wrapper
script.
Basically,
your
your
existing
work
here
in
this
repository
I
think
we
should
take
this
approach.
For
now
we
can
have
the
binary
the
proxy
binary
deployed
by
the
script,
and
also
the
script
can
manage
the
separate
commit
configuration
abstraction
as
well.
B
A
Yes,
so
this
yeah
this,
this
could
be
deployed
as
a
pre
step
in
terms
of
like
hey,
I
want
to
set
up
a
service
services,
and
then
the
users
in
the
docs
for
the
users
have
to
be
updated
to
instruct
them
to
simply
calculate
them,
and
it's
going
to
work.
I
think
the
cap
has
to
be
updated
to
modify
these
two
sections.
A
A
Honestly
I
have
an
argument
with
Michael,
who
is
the
chair
of
signatures,
that
we
should
not
push
this
feature
without
getting
some
of
these
details
clear,
because
we
have
to
secure
all
the
paths
if
they
are
not
secure.
We
cannot
graduate
this
feature
together,
but
I
think
for
for
a
question
like
in
terms
of
the
implementation
you
have
been
working
on.
A
E
C
A
That
explains
what
flags
are
part
of
the
corporate
configuration,
because
some
of
these
flags
unique
are
not
part
of
the
correct
configuration,
so
we
have
to
pass
this
command
line.
In
any
case,
this
is
the
setup
for
the
Windows
service,
and
currently
we
are
against
avoiding
so
Ben
I
think
we
are
writing.
The
couplet
configuration
on
disk
with
a
Canadian
join
incubate
am
is
using
this
same
configuration
so
I
don't
even
know
what
the
problem
there
is.
Why
should
we
manage
the
different
windows?
Oh.
A
C
I
proposed
okay
I
think
that
I'm
super
happy
that
I
see
is
moving
and
also
beginning
with
getting
rid
of
the
screen.
I'm
happy
so
I
think
that
we
should
iterate
on
on
the
cap
addressing
details
like
the
ex-wife
labels
or
considering
the
two
option,
which
is
provider
to
make,
who
a
mini
start
so
having
only
one
copy
of
the
config
map,
so
we're
in
the
cluster
and
distal
or
something
else
that
we
have
to
figure
out,
but
but
I'm
happy
that
there
is
a
way
forward.
For
this.
B
A
B
B
A
Wanted
to
mention
that
this
opens
the
very
easy
option
to
test
the
cube.
Adm
/
windows
feature
on
coaster.
Api
AWS
because
been
tested
is
already
with
the
other
solution.
We
entered
some
complications
that
you
have
to
execute
the
whole
previous
for
a
wrapper
script
using
a
part
of
coordinate
on
windows.
In
this
new
solution,
we
still
have
to
have
the
port
of
coordinate.
Maybe,
but
you
see,
the
script
is
much
simpler
and
basically
this
is
a
much
much
more
friendly
to
the
cost
API.
I'm.
A
F
F
A
I'm
not
sure
how
we
can
proceed
with
announcing
this.
There
are
multiple
channels
we
can
ask
it
preferences
the
Kubica,
the
key
beta
inspection,
though
we
can
also
ask
okay,
if
I'm
not
sure
what
the
consequences
of
this
action
is
going
to
be
like
who
is
one
to
short.
F
I
think
the
the
contributor
summit
is
a
good
place
to
talk
about
this
stuff,
as
well
as
the
ear
deep
dive
session
that
could
come
because
those
are
going
to
be
people
who
are
actively
trying
to
engage
in
the
community,
so
I
think
that's
probably
target
audience.
I
think
you
might
want
to
solicit
or
have
a
set
of
questions
for
folks
on
the
cuvette
DM
selection.
F
A
If
we
asked
people,
if
we
mentioned
this
to
people,
we
are
also
in
a
way
advertising.
It
and
I
wear
two
different
ritual
if
he
thinks
that
this,
in
its
current
state
is
something
that
we
should
even
mentioned
people
without
us,
getting
getting
it
ourselves
to
the
point
where
it's
you
know,
in
a
good
shape,.
C
So,
of
course,
I'm
pushing
around
on
the
patio
since
time
so
I'm
definitely
interesting
in
this.
Even
if
now
I
mean
I,
don't
have
much
time
to
work
on
it,
but
I
want
to
first
of
all
it's
something
that
we
already
talked
at
the
last
Kubik
on
and
we
get
and
we
got
interesting
feedback
from
the
people.
C
A
C
So
I
am
not
worried.
Is
this
part
of
the
alpha
beta
J
came,
or
at
least
I
consider
it.
So
what
I
think
is
that
it
is.
We
should
have
a
separate
meeting
and
try
to
and
pay
attention
on
on
how
we
express
our
question,
because
we
should
be
clear:
we
should
avoid
confusion
with
cluster
PII
with
the
multiple
approach.
So
it
is
not,
let
me
say
easy
to
fire
a
simple
question.
You
have
to
give
us
some
some
context
around
this.
A
A
C
A
A
E
C
A
To
promote
it,
yeah
I
just
just
had
a
question,
because
I
was
observing
some
of
the
discussions
around
customized
potential
because
to
make
this
feature
and
I
think
the
concern
there
was
that
the
cube
am
bootstrap
provider.
Hostel
are
now
passing
the
customize
flack
on
the
command
line
and
because
the
future
is
not
fully
graduated,
they
didn't
want
to
do
it,
and
also
one
of
the
problems
that
I
was
at.
You
have
to
write
the
files
to
disk.
A
C
I'm
not
sure
that
I
follow,
but
if
there
is
some
requirement
from
the
from
cross,
3
PI
IMAP
discussed
them
I
don't
see.
As
far
as
I
know
the
provider
I
don't
see
program
for
the
booster
provider
to
try
to
fight.
It
is
already
freighting
in
many
other
fights.
So
I
don't
see
proper
knee
in
having
it
pricing
the
the
page
as
well,
but
I'm
open
to
discussing
I
will
ping
Chuck
or
Vince
tool
to
see
if
they
had
clear
requirement
for
it.
A
H
F
F
I
do
know,
people
use
it
because
it's
super
useful
and
these
are
probably
in
a
very
limited
fashion-
they're,
not
using
like
all
the
features
and
capabilities
and
let's,
let's
hold
and
see
how
the
community
goes
from
customeyes,
because
all
we
really
care
about
is
the
ability
to
patch
or
manifests
that's
right
and
having
a
standard
way
that
we
use
this.
That
mechanism
inside
of
our
community
and
we
try
to
follow
Cuoco
does
lead.
But
if
they're
rethinking
this
and
if
they're
gonna
change
things,
then
that's
us
hold.
C
E
A
Yeah
and
I
some
point
I
promise
that
I
wanted
to
provide
a
completely
alternative
solution,
because
if
you
remember
the
discussion
around
fighting
our
own
sandbox,
we
had
when
this
feature
was
proposed,
but
I
never
got
the
time
to
propose
changes,
but
my
changes
required
modifying
the
eBay
DM
API
to
allow
this
customization
without
part.
These
customizations
without
patches
range.
A
A
A
D
D
The
pr
should
be
done
or
should
be
in
a
green
state
again
by
the
end
of
this
week,
and
then
I
will
revisit
some
of
the
questions
in
this
cap.
It's
notable
that
a
sizable
portion
of
the
feature
set
for
the
Animus
extensions
it
does
support
rom
manifests,
but
it
also
uses
a
heavy
adoption
of
customized.
So
that's
a
just
add
some
color
to
some
of
those
comments.
We
were
just
making.
A
A
I
A
A
E
A
A
A
A
Comment
so
that
I
can
try
address
in
the
comments
today
and
tomorrow,
but
there's
way
too
many
comments,
hopefully
I'm
going
to
give
an
update
by
the
end
of
the
week
on
this.
It's
not
specific.
This
is
this-
is
sequence
life
cycle,
but
I'm
going
to
argue
that
it's
epi
machine
areas
for
me.
What
do
you
think
I.
D
A
A
F
A
So,
let's
keep
this
false,
well
circle
top.
Is
this
ego
readiness
standards
process
for
proxy
help
I
believe
we
had
a
discussion
about
this
world
and
people?
The
conclusion
was
that
mr.
Wong
from
Google
said
that
we're
probably
is
probably
not
a
good
idea
to
make
this
change
because
of
the
restarts
it's
going
to
cause
so
I'm
also
going
to
skip
this
pier
at
this
point.
I
don't
really
want
to
proceed
with
it
anyway,
so
I'm
not
sure
what
the
auto
plans
for
it
both
move.
A
Booster
tokens,
you
do
sufficient
API,
so
I
try
to
explain
to
Dix
you
that
we
are
graduating,
boost
our
tokens
like
efficiently
to
GA
in
the
docks,
but
this
is
something
that
doesn't
really
like.
Most
of
this
PR
Jordan
did
not
approve
of
it
because
he
thought
that
I
mean
it
makes
sense
to
expose
these
facilities,
but,
like
the
maintenance
of
this
particular
code,
does
not
want
to
have
this
change.
So
it's
really
up
to
these
shoes
will
now
have
discussion.
He
has
not
responded
to
the
latest
comments.
A
A
A
This
is
a
PR
that,
ideally,
we
want
to
have
at
some
point
it's
from
a
BTech
who
has
been
contributing
to
the
working
group
component
standard,
the
TLDR
of
the
PRS,
that
it
takes
our
utility
to
split
yama
documents
into
jury
case,
and
it
puts
it
into
the
component
based
repository
I'm,
not
sure.
What's
the
state
of
this
I
see
how
she
came
up
roof,
sorry
it
it's
not
urgent.
I.
A
E
D
We
go
in
with
that
directly
programming
on
this
issue
and
he
put
together
google
documents
describing
the
approach
for
how
this
functions
in
multiple
areas
in
kubernetes,
including
cook,
cuddle,
and
we
found
other
issues
with
the
way
that,
like
llamo,
is
a
kind
of
guest
on
and
all
kinds
of
interesting
streams,
but
yeah
I
can
follow
up
with
him
if
that's
what's
desired.
Here,
he's
taken
on
some
other
work
already
in
the
sig
and
the
working
group
yeah.
This
is
this
is
solid
work
and
it's
code.
A
D
There
there's
some
unknowns
as
to
whether
this
is
the
one
shared
abstraction,
but
this
abstraction
is
existing
in
the
code
and
it's
mostly
a
refactor
that
cleans
things
up
but
yeah.
There
Alex
did
additional
work
to
understand
what
the
different
abstractions
are,
and
that
is
documented
and
I
can
ask
him
to
follow
up
on
that
before
we
decide
whether
to
merge
this
or
close
it.
Okay,.
D
A
A
C
A
And
my
approach
was
this:
IR
ed
get
mode
name
from
hostname,
which
is
a
rapper
that
is
now
used
in
a
couple
of
locations,
but
it's
not
using
your
code
for
the
couplet
perfect,
so
in
qadian
we
still
have
this
mood
these
multiple
ways,
so
I
was
kind
of
okay
with
to
proceed
with
this.
This
way
for
this
particular
use
case
and
maybe
refactor
your
way
in
the
future,
but
that
was
more,
like
my
only
concern
about
this
PR
and
if
I
can
change
the
error
message
that
Rossi
requested.
A
So
with
this
is
an
interesting
something
interesting,
that's
going
to
happen,
the
cube
idea
when
we
move
it
to
a
new
repository
somebody
modifies
the
coolest
configuration
because
people
add
new
fields
to
the
current
configuration
all
the
time
without
changing
the
version
of
the
API.
This
means
that
maybe
some
of
our
unit
tests
in
the
cube
Emily
are
going
to
fail
if
such
changes
are
made
without
us
seeing
them,
we.
F
Really
need
to
like
take
the
component
configuration
or
talk
with
component
standard
people's
that
particularly
tough
up
my
top
and
about
making
the
API
be
a
first-class
venerable
thing
like
the
contract
for
componentry
should
be
an
external
thing
in
reality
that
just
gets
used.
So
if
it's
in
staging
it
should
be
posted
out
all
the
other.
F
A
Yeah,
a
sort
of
problem
with
API
machinery
as
well
is
that
they
are
completely
fine
with
adding
fields,
for
instance,
to
something
that
is
alpha.
They
add
new
fields
to
the
structure,
but
they
do
not
bomb
the
version.
I
think
they
do
not
agree
for
beta,
but
the
couplet
configuration
has
has
a
history
of
adding
fields
without
wanting
the
versions
which
is
very
popular.
That's
super.
F
A
This
is
some
of
these
changes.
We
have
to
evaluate
basically
when
we
move
out
cube
Adium
but
I'm,
starting
to
think
that
when
we
move
out,
we
are
going
to
start
seeing
some
of
these
problems.
It's
completely
expected
to
see
some
of
them
and
we
have
to
think
of
solutions
predicting
the
whole
future
is
not
going
to
be
for
us.
That's
a
point
that
I
tried
to
make
in
the
clip
that
we
just
have
to
do
it
at
some
point.
A
Okay,
so
this
is
moving
packages,
it's
my
Clayton
I.
This
is
a
lot
of
us.
I
mean
it's
starting
to
be
a
game,
but
import
was
reverse
import.
This
is
I.
Don't
have
context
from
this,
but-
and
my
assumption
is
this
touching:
it's
touching
the
cube,
a
DM
import
misfile.
Maybe
yes,
and
it's
not
being
approved,
so
we
can
skip
it,
implement
structured
output.
A
C
A
It
is
really
the
implementation
needs
more
all
right.
So,
let's
move
to
the
final
one.
We
have
two
minutes
whitelisting
of
static
pusimos,
and
so
this
is
something
that
I
wanted
to
do
for
1:18,
but
the
original
work
in
the
Cape
was
so
cold,
mostly
because
the
last
cycle
was
a
short
one
and
the
sick-out
folks
didn't
have
the
time
to
address
most
of
the
comments
on
a
certain
pending
PR.
In
any
case,
we
are
not
going
to
make
this
change,
maybe
in
118
I'm,
not
sure
but
I'm
just
having
the
piano
halt.
A
It's
following
the
logic
of
the
cap,
so
I
think
we
are
out
of
time.
Let
me
mark
this.
Let
me
copy
this
I'm
going
to
mark
where
we
are
so
that
we
can
continue
next
time.
I
just
want
to
finish
the
three
hours
of
the
PRS
and
do
some
theorizing
of
the
cube
ADM
issues
as
well,
because
we
are
very
behind
in
this
aspect
all
right.
So
thanks
everybody
see
you
next
week
see.