►
From YouTube: Kubernetes SIG Windows 20191112
Description
Kubernetes SIG Windows 20191112
A
B
C
A
A
Ourselves,
so,
okay,
excellent,
so
if
we
actually
get
that
done
and
Patrick,
this
is
where
we
might
have
to
do
a
little
bit
of
catch-up
here
with
the
kubernetes
community.
Since
the
work
for
Cuba
diem
is
out
of
three,
the
only
thing
that
would
be
remaining
for
us
to
graduate
to
better
would
be
if
the
test
and
test
grade
are
passing
right,
which
is
what
Adelina
was
working
on.
B
A
So
Patrick,
who
can
deal
with
the
dogs
I,
think
the
dogs
change.
You
can
go
and
get
an
exception
for
I
can
I
can
talk
to
Janelle
I
think
her
name
is
Genevieve.
I
can
talk
to
Jenny
for
that
and
take
care
of
that
next
week
in
person.
My
biggest
worry
is
if
Ben
ngah
finished
the
work
or
Cuba
diem
upgrades
work.
Installation
works.
We
need
to
make
sure
the
upgrades
work
right
from
216
to
217
and
we
need
to
make
sure
that
the
new
work
is
is
working
well.
A
D
D
C
D
Yeah
that
call
out
was
just
so
that
we
were
aware-
and
we
had
already
settled,
that
with
the
cluster
lifecycle
people,
so
basically
they
just
agreed
that
we're
going
to
keep
it
as
Windows
services,
so
that
worked
as
far
as
I'm
concerned.
There's
no
work
to
be
scheduled
for
having
those
as
daemon
sets.
D
B
A
In
that
case,
then,
we
shouldn't
let
our
foot
off
the
pedal
then
continue
and
kind
of
finalize
this
and
go
down
the
path
gap
of
what
Kali
I
mentioned
around
the
upgrades
and
and
getting
that
working,
and
if
we
get
that
done
after
117
chips,
which
is
seems
reasonable,
then
you
can.
You
know
immediately
shift
our
focus
to
two
cluster
API
but
God.
D
D
You
pull
them
from
the
repo,
but,
like
you
have
to
get
the
new
binaries
and
then
you
have
to
replace
those
and
start
the
new
service
and
then
what
I
was
concerned
about
is
you're.
Gonna
have
to
get
a
new
set
of
scripts,
possibly
because
the
parameters
might
change
so
I
know
from
115
to
116
a
lot
of
things
broke.
If
you
try
to
use
the
same
deployment
scripts,
it
wouldn't
work
because
things
are
getting
deprecated
and
like.
A
D
Okay
and
then
also
we
I
guess
they
moved
the
main
scripts
out,
so
those
are
common
and
I
can
understand
why?
Because
there
are
like
there,
there's
a
lot
of
common
code.
So
if
you
find
a
bug,
you'd
have
to
fix
it
in
both
repos.
But
I
do
worry
that,
like
again,
because
of
how
arguments
change
and
everything
that
it
may
not
be
the
best
idea
to
do
that
so
right
now,
the
only
things
that
are
in
the
actual
version
folders
are
the
JSON.
Oh.
A
D
D
They
have
daemon
sets
for
a
lot
of
things,
so
the
thing
is
well.
Ok,
so
the
way
it
works
for
Linux
is
you
have
to
replace
cute,
let
you
sell
and
then
you
just
run
into
you
ATM
upgrade
and
then
it'll
update
the
config
map
that's
stored
within
the
cluster
and
I.
It
will
patch,
like
the
cute
proxy
daemon,
set
and
I'm,
not
sure
how
it'll
work
for
flannel,
but
I.
Imagine
well.
No
that
has
no
bearing
so
yeah
between
sheeplings,
you
proxy.
E
A
Yeah
I
kind
of
a
green
stuff
downloading
the
binaries,
who
can
actually
ask
the
user
to
go,
get
them
from
the
right
location
as
long
as
is
the
right
version,
and
then
you
come
validated
right
so
can
check
the
minor
version
like
if,
for
example,
if
you're
running
the
script
for
117
validate
that
117
is
in,
there
is
in
the
base,
is
the
binaries
are
being
used
and
and
fail?
Otherwise,
yes,
yeah.
B
B
A
C
B
So
the
up
one
of
the
other
things
I
was
graduating,
was
the
run
as
username
was
moving
to
beta.
There
is
no
designer
real
or
functional
change
there,
but
so
what
we
needed
to
do
is
basically
when
we
go
from
alpha
to
beta.
We
needed
to
turn
the
feature
on
by
default
in
the
feature
gate,
and
so
a
mark
made
a
PR
to
do
that
and
then
I've
got
a
corresponding
doc
PR
there
and
so
I
think
it's
just
okay.
It
looks
like
it
just
needs
an
LG
TM
for
me
called
you.
G
G
B
A
F
Hey
so
I
can
give
an
update
right
now,
if
you
want
so
basically
for
the
CSI,
it's
pretty
much
being
I
hear
and
GCPD
has
been
tested
with
a
prototype
version
of
it
of
the
of
the
CSI
proxy,
but
the
actual
CSI
proxy
changes
in
the
in
the
main
repo
that
it
will
come
out
of.
That's
still
in
progress
at
the
moment.
Did
you
say
I
sure
yeah,
the
azure
disk
is
being
looked
at
by
KK
from
Microsoft.
A
F
So
those
are
the
two
ones
that
left
us
out
with
the
proxy
just
showing
that
end-to-end
works
as
your
disk
and
GCPD
by
jings
ooh
from
guru,
but
the
actual
CSI
proxy
code
itself,
with
the
different
support
for
the
multiple
versions
and
stuff
like
that.
That
is
still
happening.
There
are
a
couple
of
main
PRS
that
needs
to
be
reviewed,
probably
over
the
next
two
weeks
that
should,
and
once
that
happens
in
118,
we
can
plan
for
a
beta
with
the
proxy,
not
waiting.
That's
working,
okay,.
B
Okay,
so
looking
at
the
container
D
stuff,
the
main
thing
we're
focusing
at
right
now
on
is
actually
move
in
the
test
passes
over
to
the
container
D
master
branch.
So
up
until
last
week
we
were
testing
on
on
Justin,
Terry's,
Fork
and,
and
he
and
Lantau
took
the
main
changes
that
were
needed
for
process
isolation,
support
in
a
container
d
and
got
them
merged
back
into
the
container
D
branch
for
41.4
and
so
a
whole
bunch
of
those
things.
B
But
you
know
the
kid.
The
key
thing
is
that
we
need
to
get
this
testing
over
on
the
master
branch,
because
that's
what
we
would,
what
we're
actually
going
to
be
using
on
on
test
grid
and
right
now,
almost
everything's
failing
because
of
that,
and
so
we're
focusing
on
getting
that
fixed
and
then
the
other
and
then
so.
B
But
yeah
we've
just
got
a
number
of
bugs
that
still
need
to
be
looked
at,
but
they're
kind
of
blocked
behind
that
other
one
we
just
found.
So
at
this
point
I
would
say
it's
still
sort
of
it's
not
a
sure
thing.
We've
got
a
lot
of
a
lot
of
bug,
fixes
that
we
need
to
get
done
in
the
next
few
days,
and
so
one
of
things
I'm
going
to
look
at
is
seeing.
B
Whether
or
not
it's
possible
to
carry
some
of
these
bug
fixes
past
this
Thursday
or
if
there
are
things
that
we
can
move
out
to
beta,
because
I
I
think
our
criteria
wasn't
100%
passing,
but
right
now
or
closer
to
100%,
failing
and
so
I
need
to
look
at
that
a
little
more
carefully.
You
know,
I
guess:
I'll
share
an
update
with
with
the
list
on
like
Wednesday
or
Thursday,
to
see
where
we're
at
before
code
for
use
to
see.
B
A
The
way
one
thing
to
note
some
of
these
bugs
there
were
I,
don't
know
if
they
were
all
in
the
list
that
you
just
went
through
there's
some
of
them
that
inherently
would
be
fixed
by
container
just
by
moving
into
container
D
right.
There
was
some
issues
around
you
know
the
networking
start
being
called
multiple
times.
A
There
was
one
about
slowness
that,
like
that,
basically,
the
performance
of
the
cubelet
was
very
slow
in
terms
of
getting
the
stats,
and
someone
mentioned
that,
just
by
moving
to
container,
they
will
be
fixed.
Are
you
gonna
resolve
those
ones
container
the
cats
in
the
state,
or
are
we
going
to
individually
go
back
at
them
later
on?
Oh.
B
Yeah,
so
some
of
these
are
in
here
the
stats
one
we
haven't
had
it
had
time
to
evaluate
it
yet,
but
otherwise
I
think
we
could
basically
resolve
those
and
say:
okay,
these
ones
are
going
to
be
fixed
with
container
D,
so
we'll
probably
want
to
actually
put
a
milestone
on
them
for
like
18
or
19,
for
either
beta
or
stable.
Once
those.
C
B
B
A
E
A
A
I
mean
for
the
ones
that
required
doggy
issues,
I'm
okay,
with
that
actually
they're
keeping
track
of
an
open
dog
issue
or
maybe
wait
keeping
the
bag
open.
But
for
some
of
them
don't
have
it
like,
for
example,
stats,
API,
slow
and
continually
fixes
it,
there's
nothing
user,
perceivable
or
user
impacting
or
the
other
one.
That
Dinesh
worked
on
like
almost
a
year
now,
where
the.