►
From YouTube: Kubernetes SIG Windows 20200623
Description
Kubernetes SIG Windows 20200623
A
A
So
I
wanted
to
kind
of
follow
up
on
an
agenda
item
from
last
week,
which
was
what
do
we
do
about
some
of
the
changes
that
have
been
happening
that
require
signal
to
approve
them,
whether
they're
cherry
picks
that
have
to
go
to
previous
release
or
items
that
we
depend
on
right
now.
As
everyone
knows,
signal
has
been
incredibly.
A
The
immediate
relief
is
coming
from
a
couple
of
additional
folks
being
added
into
the
sicknote
community
that
will
kind
of
help
increase
their
throughput,
both
for
reviews
and
and
replies,
and
maybe
we're
not
going
to
see
a
lot
of
improvement
in
the
next
couple
of
weeks,
but
over
the
course
of
119.
I
think
that
we
should
be
able
to
see
a
noticeable
improvement.
A
What
I
want
to
do
is
maybe
not
today,
but
if
you
have
requests
that
you
won't
seek
not
to
approve
that
anything
that
hasn't
happened
already,
whether
cherry
picks
or
something
else.
It's
just
kind
of
stuck
just
send
me
a
private
message
on
slack:
I'm
gonna
collect
them
all,
and
I'm
gonna
create
a
single
email
that
I'm
gonna
send
to
the
new
folks
are
getting
added
into
the
sick,
node
approvers
list
and
see
if
they
can
take
care
of
it
for
us.
But
again,
I'm
not
gonna.
A
B
No
just
that
I've
been
also
feeling
pain
from
not
having
some
of
these
fixes
back
ported
and
I'm
glad
to
see
that
they're,
adding
or
adding
more
approver
bandwidth
here.
A
Yeah,
it's
again,
it's
immediate
relief
right
now
and
and
hopefully
it
will
translate
to
to
long-term
relief
as
well.
I
believe
someone
asked
me:
are
the
approvers
gonna
be
on
previous
branches
of
kubernetes
as
well
like,
for
example,
if
we
go
back
to
116
or
117,
are
the
approvers
gonna
be
applicable
there?
I
haven't
gotten
an
answer
on
that,
but
I
believe
they
will
be.
I
think.
A
So
one
of
the
things
that
we
wanted
to
to
discuss
today
were
very
very
quickly
was
you
know,
there's
a
lot
of
networking
changes
are
happening
upstream
and
david.
Thank
you
for
shepherding
a
lot
of
the
documentation
on
that
I
I've
seen
you
make
a
lot
of
changes.
I
approve
some
of
them
and
I
see
you
make.
I
did
even
more
things,
so
thank
you
for
taking
the
lead
for
both.
You
know,
driving
the
changes
as
well
as
adding
the
docs.
A
So
one
of
the
things
that
came
up
is
that
there's
a
lot
of
changes
are
coming
up
and
some
of
them
are
not
applicable.
Unless
you
have
windows
2004
so
can
we
accelerate
and
make
sure
that
for
version,
119
2004
is
supported.
A
A
Can
we
make
2004
to
be
one
of
the
two
sac
releases
we
support,
so
maybe
1909
and
2004..
Is
that
a
good
plan?
I
want
to
hear
from
mark
and
adeline
as
well.
B
Yeah
I
can
go.
I
think
that
that's
a
good
plan.
I
know
james
and
myself
been
working
on
making
sure
that
2004
support
is
stable
in
aks
engine,
which
is
what
we
use
to
deploy.
A
lot
of
the
test
passes
that
run
on
azure
for
sig
windows
and
that's
looking
pretty
stable.
We
wanted
to
make
that
was
good
before
we
start
to
light
up.
Some
test
grid
test
passes
on
that
and
I
so,
but
I
think
we're
getting
to
the
point
that
we
should
be
able
to
do
that.
B
So
yeah,
I
hope
that
in
the
next
week,
or
so,
we
should
be
able
to
have
test
grid
test
passes,
set
up
running
2004
for
the
same
segwindos
test
passes
to
get
a
more
comprehensive
test
signal,
and
if
that
looks
good,
we
can
just
have
a
conversation
about
which
versions
we
want
to
put
in
the
documentation
about
support
for
119.
A
Okay,
so
I'm
gonna
put
it.
It
is
a
plan
of
the
commitment,
so,
okay,
so
let's
get
an
update
next
week,
but
yeah.
If
we
can
support
204,
1909
and
1809.
I
think
that
gives
us
a
good
breadth
of
lts
and
sac
releases,
and
you
know
most
folks
are
probably
not
going
to
be
on
1903
anymore
right.
I
would
assume
that
most
folks,
if
they're
on
the
biannual
releases,
they
would
have
moved
out
of
1903.
Is
that,
like
kind
of
a
fair
assumption.
B
Probably
also
if
peter
or
you
are
on
the
call,
I
I
if
one
of
you
could
comment,
are
you
planning
on
adopting
or
have
already
adopted,
2004
supporting
gke.
C
I
I
can't
comment
on
gke
just
yet,
but
we
can
definitely
within
the
next
few
weeks
bring
up
a
test
job
on
on
gce
with
2004.
That
shouldn't
be
a
problem.
C
Awesome
is
there,
is
there
an
issue
open
to
track
any
of
this.
A
C
Okay,
I
can
create
an
issue
to
track
adding
the
gce
test
job
and
all
cc
the
right
people
on
it.
A
Okay,
thank
you
all
right.
Let's
move
on
to
the
next
one,
nilesh
are
you
on
today?
A
A
The
fact
that
it's
super
early
for
you
to
it's
like
3
30
in
the
morning
right.
A
All
right,
you
want
yeah,
you
have
the
floor.
D
I've
been
working
with
this
sig
for
the
windows,
implementation
with
a
big
financial
company
in
sydney
and
yeah
so
been
putting
some
of
the
pain
points,
we're
still
on
1903,
looking
to
move
to
1909
in
the
next
few
months,
but
yeah
so
at
the
moment,
they're
doing
an
on-prem
implementation
of
a
windows
host
hosting
all
the
containers
for
down
here
I
thought
you're
looking
to
get
involved
and
see
where
I
can
help
and
to
do
testing
or
to
even
help
fix
some
of
the
bugs.
So.
A
Cool,
thank
you
thank
you,
nilesh
and,
and
I
believe
craig
replied
also
to
your
email
to
the
mailing
list
about
you
know
connecting
together
for
the
sick
off
and
there's
a
policy
profiles
for
security
parts.
So
is
craig
on
the
call
today
I
don't
see
craig
today.
So
craig
is
didn't
join
us
today,
but
you
can
follow
up
on
the
email
and
you
can
connect
and
figure
out
next
steps
would
that
work
yep.
That
would
work
perfectly
awesome.
Thank
you
and
thank
you
for
joining
us
today.
A
I'm
glad
to
see
more
folks
from
from
down
under,
and
I
know
the
time
is
not
conducive
for
you
guys,
but
they
appreciate
the
effort.
B
B
B
Okay,
the
last
issue
that
I've
been
tracking
for
container
d
kind
of
feature
parody
with
docker
is
this.
I
link
the
pr
in
the
docs
here.
There
is
a
pr
open
to
add
named
pipe
mount
support
for
into
continuity
for
windows
and
that's
needed
for
primarily
the
csi
proxy
work.
That
pr
looks
like
it's.
B
Implementation
is
kind
of
approved,
but
they
want
to
see
more
test
coverage,
so
I'm
just
really
hoping
that
that
kind
of
lands
soon
so
that
we
can
validate
that
that
works
and
then
also
get
some
tests
for
csi
proxy
up
on
test
grid
with
container
d.
That's
kind
of
blocking
that
right
now.
A
Sounds
good,
I'm
assuming,
given
that
we're
like
quite
a
few
weeks
away
from
119
that
it
should
be
safe
for
this
to
make
it
in
right.
B
Hopefully
yeah
it
yeah,
it
would
have
been
tough
with
code
freeze
on
the
25th,
but
since
I
got
moved
out,
I
had
pretty
high
hopes.
A
Yeah
and
I'm
hoping
that
this
is
the
last
change
to
the
schedule,
because,
obviously,
like
all
the
other
changes
are
going
to
push
everything
towards
the
cubicle
in
boston.
If
that
ends
up
happening,
so
that's
going
to
make
things
a
ton
more
difficult
for
folks
that
want
to
prepare
for
confidence
plus
do
a
lot
of
work.
E
Hey
yeah
excuse
me,
things
are
going
pretty
well.
We
have
jing
and
kalia
helping
out
in
different
aspects
of
this
overall.
Looking
pretty
good
for
beta
in
119
things
we
are
probably
not
going
to
get
to
would
be
the
no
driver,
registrar
back
tests,
the
end-to-end
tests
for
windows.
Specifically,
so
we
were
thinking
we
can.
Maybe
you
know
leave
that
for
ga,
but
that's
just
like
an
external
sidecar
anyway.
E
So
it's
it's
it's
fine
to
maybe
not
deal
with
it,
but
we
have
like
all
the
e2e
pipelines
going
for
csi
proxy
now
integration
tests
unit
tests.
All
of
that
we
also
have
coverage
through
some
of
the
amazing
test,
infra
work
that
has
been
done
on
the
aks
side,
as
well
as
things
that
jing
has
been
doing
for
gce
pde
as
well,
so
lots
of
end-to-end
test
coverage,
which
is
one
of
the
main
things
we
wanted
to
focus
on
for
beta
yeah.
A
What
what
which
csi
providers
have
we
tested
that
are
gonna.
E
Work
so
right
now
it's
going
to
stay
azure,
disk
and
gcpd.
I
think
azure
file
has
also
been
tested,
but.
B
E
E
Okay,
perfect
yeah
and
there's
also
the
gcp,
the
google
cloud
smb.
You
know
the
the
equivalent
of
azure
file
in
google
cloud,
that's
also
being
tested
at
the
moment
so
to
smb
to
desk-based.
E
I
did,
I
mean
basically
the
point
they
said
they
do
not
have
cycles
like
at
least
this
quarter,
and
they
would.
E
Be
coming
back
to
see
like
pick
it
up
in
the
gfes.
A
Okay
same
thing:
for
this
year,
by
the
way,
I'm
having
a
hard
time
finding
cycles
but
trying
to
elevate
the
priority
there,
so
someone
can
do
the
work,
I'm
trying
to
see
if
ben
moss
can
come
back
and
and
do
it
for
us,
but
he's
been
pulled
into
other
things,
anything
else
on
this
dip.
No,
this
is
great
progress
by
the
way.
This
is
awesome
glad
to
see
this
going
to
bed.
So
this
is
a
huge
huge
for
us
all.
Right
windows,
containers,
github
repo,
oh
by
the
way,
deep
one.
A
Last
thing
I
don't
know
if
you
heard,
but
in
the
csi
world
they're
trying
to
create
enable
a
new,
a
new
way
to
basically
for
backup
to
happen
for
csi
providers
by
creating
a
snapshot
api
to
csi.
Do
you
know
if
that's
supported
by
any
of
these
providers
that
you
mentioned,
or
is
it
something
that's
going
to
come
later?.
E
So,
basically,
anything
that's
on
the
controller
side,
which
is,
like
you
know,
a
controller
driven
action.
We
are
not
focusing
on
these
all
focus
on
like
the
node
actions.
So,
like
you
know,
if
the
if
it's
the
controller
that
can
completely
take
the
backup
of
a
disk
by
invoking
the
appropriate
cloud
apis,
that's
kind
of
beyond
the
realm
of
csi
proxy.
E
If,
if
it
tries
to
do
like
you,
know
kind
of
like
a
flush
and
hold
or
request
semantic
on
the
node
itself,
that's
when
csi
proxy
comes
in
and
I
don't
believe
that
aspect
has
been
discussed
too
much.
Fortunately,
jing
who
is
helping
a
lot
from
google
is
also
quite
plugged
into
the
snapshot
api
side
of
things
so
yeah,
I'm
sure
you
should
bring
it
up.
A
Yeah
I
have
a
few.
I
have
a
couple
of
folks
that
are
also
plugged
into
that.
I've
asked
them
to
test
with
windows,
but
without
csi.
First
then
csi
next
and
I
want
to
see
if
they
can
do
that
sure.
But
okay,
thank
you
appreciate
it
all
right.
I
think
maz
you're
the
on
the
right,
the
last
one,
which
is
windows,
containers,
github,
repo
for
os
issues,.
F
Yeah
I'm
going
to
let
winisha
from
our
team
speak
to
that,
and
then
I
can
add
something
we
need
to
do.
One
announce
the
github
repo.
G
Hi
everybody,
my
name
is
vinicius.
I
work
the
same
team
that
was,
and
some
other
folks
here
in
the
group
work.
We
are
part
of
the
windows
container
platform
team.
Just
this
morning
we
released
the
new
github
repo
that
is
not
focused
on
having
any
code.
The
idea
is
that
we
are
creating
a
new
channel
for
having
customer
engagement.
Just
like
many
other
groups
at
microsoft
are
doing,
including
the
aks
team
on
which
we
are
going
to
allow
customers
to
come
in,
create
new
issues
that
could
be
bugs.
It
could
be
feature
requests.
G
G
So
we
hope
this
is
this
new
channel
will
help
customers
get
more
close
to
the
product
team
and
will
help
us
from
our
side
get
more
insights
on
what
our
customers
are
are
looking
for
and
what
we
should
be
prioritizing.
F
Yeah
and
just
to
add
on
so
this
is
based
on
the
on
the
feedback
we
have
gotten
from
community
and
customers
that
you
know
there
are
os
level
changes
for
windows,
containers
and
usually,
even
if
you
know,
michael
in
this
community,
if
you
have
an
issue,
you
normally
reach
out
to
mark
or
myself,
like
you
know,
we,
you
know,
we
triage
those
issues
and
you
don't
have
visibility.
So
the
point
is:
if
there
is
an
os
level
changes
any
kind
of
networking
storage.
F
So
we
have
a
lot
of
features
in
kubernetes
community
already
I'm
trying
to
link
some
of
them
already
imported
some
of
them
there
so
we're
you
just
have
to
create
an
issue
and
link
it
to
your
to
the
one
in
the
kubernetes
and
then
the
windows
team
can
like
give
you
an
update
on
like
when
it
will
be
fixed.
That
venetia
was
saying
a
rough
timeline
and
which
version
would
be.
Would
it
be
targeted?
F
So
we're
super
excited
about
this
and
we're
hoping
folks
will
use
it
in
community
and
outside.
A
So
yeah
absolutely
would
they
be
able
to
like
we
have
a
getting
started
guide
for
windows.
Let
me
open
it
up
here
really
quickly,
so
over
here
under
sick
windows,
we
have
a
getting
started
guide
that
basically
kind
of
covers
a
lot
of
the
links
and
and
items
for
for
for
windows,
and
one
of
the
things
that
we
have
is
beyond
the
documentation.
We
have
a
contributor
guide
for
sick
windows.
A
A
So
what
do
you
guys
think
yeah.
G
F
And-
and
this
is
this
is
just
a
start
and
we're
hoping
that
you
know
the
community
is
going
to
get
engaged
more
and
you'll
have
more
visibility
on
windows
containers,
especially
the
changes
with
their
which
which
are
within
the
us.
So
if
there
is
any
feedback
on,
if
we
can
do
this
better,
if
if
there
is
a
way
of
doing
this
better
or
improving
we're
all
ears-
and
you
know
we'll
continue
to
improve
the
repo.
A
F
Well,
I
I
think,
just
a
quick
two
things
I
think
I
was
talking
to
mark
earlier
yesterday.
So
one
is
michael.
The
the
documentation
deadline
for
119
is
one
week
after
the
the
poetry
is
right.
Yeah.
A
Yeah,
I
think
that
you
know
big
items
for
120.
We're
gonna
have
to
I'm
trying
to
get
some
resources
for
cluster
api,
because
if
we
don't
do
that,
I
think
that's
gonna
hurt
us
a
lot
and
obviously
container.
They
need
to
continue
the
investments
because,
as
a
strategic
bet
for
us
to
enable
new
scenarios
like
hyper-v
isolation
and
and
you
know,
get
away
from
docker
demon,
so
I
think
that's
you
know.
A
It
goes
without
question
that
that's
a
huge
need
for
us,
but
if
we
don't
end
up
doing
the
cluster
api
and
cuba,
dm
and
and
moving
forward,
it's
gonna
hurt
us
yeah.
What
do
you
mean.
F
A
Cluster
api.
Well
so
so
we
had
a
few
folks
like
ben
and
gab
work
on
get
cube
adm
to
beta.
We
need
to
move
it
forward
right
and
then
we
need
to
make
sure
that
cluster
api
has
support
for
windows
notes,
because
I've
seen
a
lot
of
distributions
out
there,
commercial
distributions
that
are
trying
to
align
with
cluster
api
as
the
means
of
basically
integrating
with
kubernetes
for
lifecycle
management.
So
if
windows
is
not
in
that,
then
we're
gonna
be
left
behind
right.
B
Me
share
part
of
that
is
some
cloud
provider
implementations
for
cluster
api
and
I
think
we
were
targeting
a
couple
for
a
windows
day
view.
A
Yeah
I
mean
we
would
start
with
the
azure
and
and
gcp,
which
are
the
two
biggest
contributors
to
this
community
right
with
peter
and
and
all
the
folks
from
microsoft
and
then
from
there
on
like
aws.
If
they
come
to
the
table,
that
would
be
another
one
but
yeah
looking
at
this
and
this
year
as
well,
which
you
know
I
I
ultimately
have
to
own.
So
if
we
look
at
the
schedule
by
the
way,
since
you
asked
earlier
mass,
you
know,
dogs
are
july,
16th
right
now,.
F
A
Okay,
so
yeah,
so
if
we
get
close
to
code
free,
we
probably
need
to
switch
the
focus,
documentation.
Yeah-
and
you
know,
if
there's
like
you
know
it's,
it
doesn't
hurt
like.
Maybe
you
we
can
like
several
folks
can
follow
what
david
is
doing,
which
is
as
if
you
know,
you're.
Adding
changes
and
pr's
are
going
in,
just
maybe
add
a
placeholder
for
some
of
those
docs
right
like
something
simple
something
small
just
to
know
that
this
is
coming
and
target
does
work
in
progress.