►
From YouTube: Kubernetes SIG Windows 20191029
Description
Kubernetes SIG Windows 20191029
A
Hello,
everybody
and
welcome
to
another
sig
windows
meeting
it's
the
29th
of
October
today
and,
as
always,
is
a
recorded
meeting.
So
please
adhere
to
the
sin:
CF
code
of
ethics
and
guidelines
and
let's
get
started
so
in
our
agenda.
For
today
we
have
a
status
update
for
ran
time
class,
I
guess,
but
it
that's.
You.
B
B
Signo
to
do
a
review
of
this,
but
I'd
link
to
PR
here
for
everyone's
reference
in
we've
got
a
couple
questions
on
implementation,
detail
on
how
we
would
actually
make
the
CRI
update
and
we
were
gonna-
go
talk
to
Tim,
while
Claire,
hopefully
next
next
hour,
if
he's
available.
Otherwise
we
review
sometime
this
week
anyway,
so
it's
their
family
if
anyone's
got,
got
feedback
on
that.
But
that's
a
CRI
update
along
with
a
couple
corresponding
unit
tests
for
it
and
then
once
once
that
is,
you
know,
reviewed
and
merged.
B
A
B
There's
had
a
quick
note
on
that,
so
the
next
side
of
my
head
on
the
list
was
one
that
I
thought
that
there
as
well
around
the
cube,
ADM
stuff
and
so
right
now
I'm
working
on
updating
the
scripts
that
are
similar
to
the
ones
that
are
used
in
the
cube
ATM
repo
over
in
the
sig
windows
tool
tools,
and
so
I
should
be
able
to
have
a
PR
to
that
repo
sometime
next
week.
I've
got
this
stuff
working
today
between
between
a
Linux
VM
and
a
windows
VM,
but
I've
got
a
couple.
B
Networking
bugs
I
need
to
work
through
and
I.
Think
I
just
need
to
sync
and
build
the
latest
version
of
network
plugins,
because
I
deployed
an
old
build
on
there,
and
so
I
should
be
able
to
get
that
fixed,
and
so
when
there's
PRS
for
that
stuff,
I'll
drop
in
the
channels.
So
people
can
see
what
the
updated
set
up
scripts
are
going
to
work
like
so.
B
A
B
A
Essentially,
it's
getting
us
to
the
same
state
as
we
had
with
1.16,
but
for
continued
in
this
case.
Nothing,
that's
nothing
new
yeah!
So
we're
still
looking
for
someone
that
can
help
us
on
the
upgrade,
so
I
think
Luba
Mir
in
you
know,
and
I
need
to
see
if
you
can
actually
get
some
time
to
do
this.
What
do
you
mean
by
the
upgrade
so
so
Kalia
had
mentioned
that
no
she's
on
the
call
to
that.
One
of
the
things
that
was
missing
on
QB
DM
from
advancing
to
the
next
state
was
basically
handle
upgrades.
A
B
A
15
yeah
one
16
to
one
for
example,
and
then
the
second
thing
was
the
automation
and
Adelina
his
hand
is
basically
took
a
lot
of
the
automation
to
put
it
inside
our
our
testing
bench.
So
that
was
one
part.
The
upgrade
is
still
missing,
so
we
wanted
someone
that
it
made
may
very
well,
everything
would
just
work
is
what
Kalia
mention
is
just
we're,
not
sure
if
you
will
Kali
I
did
I
did
I
mention
all
this
correctly
or
is
there
anything
else?
You
can
add
nope.
D
Would
adhere
to
and
I
think
it's
in
the
notes
there
I'm
working
on
something
similar
to
what
budgeted
using
kind
and
of
yourself
to.
Basically
our
idea
is
to
create
to
have
a
method
chichi
deploy
these
clusters
faster
than
what
we
have
today
in
in
the
other
scenarios,
and
this
obviously
heavily
uses
cube
ADM,
so
I'm
still
gonna
go
ahead
and
fix
or
change
whatever
I
find
Inc
you
Bailey,
and
that
needs
to
needs
attention
and
I
would
include,
for
example,
setting
up
for
for
117
since
we're
testing.
D
We
should
be
testing
that
I
will
take
a
look
at
AB
grades
in
this
work.
Part
in
this
work.
No,
my
first
priority
is
obviously
to
set
to
say
this
stuff
up:
okay,
basically
I
guess
what
needs
to
be
done
here
is
just
to
run.
It
figure
out
if
it
works.
If
it
doesn't
white
I,
don't
see
that
it's
that
much
of
of
a
problem
but
and
I
said
my
prosperity
is
automation.
A
C
C
B
Yeah
this
is-
and
this
is
very
similar
to
a
setup
I'm
doing
for
dev
purposes
where
you
know
you'll
have
either
WS
l2
or
Linux
VM
running
a
control
plane
for
the
multiple
windows
machines
running.
So
that
way
you
know
running
cube
test
and
a
laptop
is
feasible
and,
of
course,
running
at
a
nested.
A
sure.
Vm
gets
us
closer
to
the
speed
we
need
for
PR
testing,
mm-hmm.
C
C
D
To
is
free,
and
we
also
can
pack
it
in
something
that
people
might
actually
use.
I
was
actually
planning
on
writing
some,
because
I
have
kind
plus
joining
manually
joining
a
Windows
node.
We
can
do
that
and
it
works
and
I
was
actually
planning
to
to
contribute
to
kind
documentation.
For
for
that
scenario,
I
still
need
to
make
it
work
with
you.
Baby
M
I
have
a
little
hiccup
there.
B
I'll
mention
it
if
it's.
If
it's
working
well
and
if
we
have
time
but
the
main
focus
there
is
on
being
able
to
run
like
the
kubernetes
conformance
and
in
unit
tests
around
it
before
submitting
a
PR
and
so
they're,
basically
trying
to
run
all
the
major
PR
tests
before
you
create
your
PR
with
the
argument
that
that's
going
to
give
you
a
faster
turnaround
on
bugs
and
waiting
on
proud
to
catch
up.
D
So
basically,
the
idea
there
is
that
we
run
the
test
inside
our
own
brow.
Everything
worked
fine,
the
test
press,
but
we
need
some
changes
in
the
deploy
script
for
webhook.
If
you
take
a
look
at
that,
PR
there's
nothing
much,
there's
some
assumptions
that
the
script
does
about
keep
CTL
and
where
it
is,
and
about
some
some
packages
that
need
to
be
on
the
note
I.
D
Well,
III,
don't
know
how
the
way
tip
CTL
that
variable
the
way
it's
inside
the
cube
desk
is
frankly,
very
confusing
and
for
reasons
that
I
don't
understand
it
probably
legacy
but
I'm
sure
some
test
somewhere
uses
it
and
that's
why
I
want
to
change
stuff
locally
I
mean
you
have
the
changes
contained
into
that
script
other
than
the
images.
So
we
don't
change
some
stuff
and
break
everything
for
reasons
that
we
don't
even
understand.
D
B
C
A
A
C
B
C
Hear
perspective
yeah
so
that
work
is
progressing
pretty
well,
it's
pretty
much
at
this
point
like
a
lot
of
what
was
done
around
the
code
generation
side
of
it,
where
we
have
the
support
for
supporting
different
versions
of
the
API,
and
all
of
that
is
like
in
a
fairly
stable
state
and
right
now
the
actual
host
side
api's
are
being
introduced
in
that
framework.
So
we
have
the
file
system
stuff
being
reviewed.
The
disk
and
the
volume
will
follow
pretty
much
it'd.
A
B
A
C
Sounds
good
someone
just
posted
about
Oh
Peter
did
about
CSI
windows
channel,
so
yeah
I
know
KK
and
Jing,
who
is
working
on
GCPD
and
obviously
Peter
are
all
there
and
we
have
so
yeah.
Whoever
is
interested
can
join
the
CSI
windows
channel
and
I
can
send
them
the
invite
for
the
meeting
and
they
can
get
pointers
to
where
to
pick
it
up.
A
A
It
looks
like
we'll
finish
ten
minutes
early
today,
I
kill
with
the
ten
minutes
back.
Well,
don't
forget
if
you
are
registered
for
the
contributor
summit,
I
believe
them
I
have
closed
the
registration.
So
if
you
still
want
to
go,
let
me
know
see
if
we
can
get.
You
in
I
know
that
Jeremy
just
got
approval
to
join
us
there.
So
we'll
see
him
in
person
at
the
the
conference
as
well.
A
We
don't
have
any
contributor
summit
specific
sessions
on
windows,
but
if
you,
if
there's
enough
of
us
they're
going,
maybe
you
can
ask
for
a
meeting
room
and
just
sit
down
somewhere.
Otherwise,
we'll
see
everybody
there's
a
meet-and-greet
or
me
and
dipper
going
and
then
there's
others
or
two
talks
on
intro
to
windows,
as
well
as
the
deep
dive
and
day
2
operations.
So,
hopefully
we'll
see
you
guys
there
cool
all
right
have
a
good
day.
Everybody
we'll
see
you
guys
next
week,
bye.