►
From YouTube: Kubernetes SIG-Windows 20221129
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
All
right,
hello,
everybody
and
welcome
to
the
November
29th
2022
iteration
of
the
kubernetes
Sig
Windows
community
meeting.
As
always,
these
meetings
are
recorded
and
uploaded
to
YouTube
so
be
sure
to
adhere
to
the
cncf
code
of
conduct
for
anybody
new.
That
basically
just
means
be
nice
to
everybody
and
we'll
start
with
some
announcements.
First
announcement,
we
were
kind
of
just
talking
about
the
off
camera,
but
blog
posts
for
126
should
be
ready
for
a
review
today.
A
A
I
think
we're
just
tracking
the
one
for
host
process
containers
going
GA
we'll
talk
about
that
a
little
bit.
Does
anybody
have
else?
Have
any
announcements,
I
think?
Oh
one,
other
announcement
is
the.
A
The
maintainer
track
talks
for
kubecon
EU
submission
is
open.
I
think
that
they're
thought
it
was
January,
13th
or
January
17th
is
when
they'd
like
those
to
be
submitted.
So
we
can.
We
should
discuss
and
figure
out
what
we
wanted
to
include.
I,
know
I.
Think
Brendan
had
a
couple
ideas
of
things
to
include
that
were
more
technical,
like
do
another
Deep
dive
into
a
part
of
how
things
work
on
Windows
for
folks
I
think
folks
usually
respond
well
to
those
and
enjoy
listening
to
those.
B
We
are,
we
are
onboarding,
a
new
new
guy
prashad
to
help
with
the
operational
Readiness
stuff.
That's
one
thing
that
probably
is
worth
noticing
and
for
folks
that
don't
know
chinchi
who
was
who's
leading
that
effort.
She
just
left
VMware.
So
we
we're
going
to
try
to
I'm
going
to
try
to
help,
get
him
up
to
speed
and
he's
starting
out
by
figuring
out
how
to
set
up
Windows
clusters
and
so
on
and
so
forth.
So
I'm
going
to
encourage
him
to
say
hi
in
the
Sig
Windows
room.
I'll.
C
A
All
right
kind
of
speaking
of
that
we
can,
if
there's
any
new
contributors
on
the
call
who
want
to
introduce
themselves
or
say,
hi,
feel
free
to
do
so.
If
anybody
just
wants
to
say
hi
I
mean
feel
free
to
say,
like
what
you're
hoping
to
get
out
of
joining
the
meeting
or
what
you'd
like
to
work
on
or
what
you'd
like
to
see
be
worked
on
pause
for
a
minute.
D
Hey
it's
Andy
here,
I've
I
guess
been
lurking
a
little
bit
in
the
Sig
Windows,
but
I
guess
just
generally
just
interested
in
all
things.
One
day
is
and
windows
containers
I
have
no
idea
like.
What's
what
the
things
are?
People
were
working
on,
but
just
interested
to
learn
more.
A
Cool,
are
you
part
of
the
the
slack
Channel
too?
That's
probably
a
good
idea,
or
a
good
way
to
see.
What's
going
on.
D
No
I'm,
not
I've,
if
you
have
any
pointers
for
that
that'd
be
awesome.
A
Okay,
there
it's:
where
is
I'll,
does
somebody
want
to
find
a
link
to
the.
A
Yeah,
so
there's
a
slack
and
is
it
it's
probably
just
kubernetes
like
slack
that
kubernetes
dot
IO
you
can
put
in
your
email
address
and
you'll,
get
an
invite
to
join
the
slack
organization
and
then
there's
a
ton
of
different
channels.
Hashtag
Segway
doses,
a
good
one
to
start
with,
if
you're
interested
in
that.
A
Okay,
that
got
fixed
yeah,
so
normally
I
think
right
now
we're
kind
of
just
in
a
slow
spot
because
we're
after
code
and
test
freeze
for
the
current
release
and
but
the
branches
have
been
open
and
planning
hasn't,
really
started
for
the
127
release.
A
So
a
little
bit
just
kind
of
doing
odds
and
ends
and
cleanup
work
right
now.
I'll
probably
start
doing
more
planning
for
127
in
the
next
couple
of
weeks,
but
welcome.
E
A
Okay,
I
guess
we
could
start
with
the
agenda.
Now.
First
agenda
item
is
the
blog
posts
I
I
didn't
realize
it?
I
was
a
bit
confused
because
it's
at
first
they
said
they
needed
a
draft
review
or
a
draft
up
by
today
by
November
29th,
but
it
looks
like
in
slack
they're
expecting
it
to
be
ready
for
review.
A
I
saw
Jay
did
a
quick
review
of
it,
but
Brendan
and
I
were
co-authoring.
It
I
think
Brenda
did
most
of
the
work
Brenda.
Did
you
have
a
chance
to
look
at
this
anymore.
F
Sorry,
no,
it
looks
like
it's.
My
changes
did
get
pushed,
but
there's
just
a
bunch
of
examples
in
those
sections
where
the
shoes
are
which
would
be
nice
to
have,
but
I
guess
they're
not
like
totally
essential
I
can
I
can,
you
know,
definitely
fill
in
a
bunch
of
those
but.
A
Could
probably
fill
in
some
of
these
I
didn't
check
it?
Did
we
link
to
the
the
YouTube
recording
for
host
process
containers
I,
think
that
would
be
a
good,
also
reference.
Oh.
F
A
A
Okay,
oops
I'll
link
to
another
PR
and
we
can
I
guess
we
can
coordinate
in
slack
Brandon
about
to
yeah
there's
in
the
docs.
Let.
C
C
A
A
A
But
we
probably
should
highlight
some
of
the
use
cases.
We
could
pull
those
from
the
YouTube
video
and
the
the
coupon
recording
and
things.
A
F
Yeah
I
mean
this
has
to
be
ready
by
today,
so
I'll,
oh
yeah,
I'll.
C
A
B
I
think
we
should
just
do
it.
I
I
mean
I
I
think
that
the
direction
is
that
we
just
would
do
it
because
I
don't
here's,
here's,
why
I
think
it's
low
risk
I
think
basically
nothing's
happening
in
the
cube
proxy.
At
this
point,
it's
a
it's
basically
a
dead
code
base
and
the
only
reason
that
it's
so
hard
to
change
is
because
of
all
the
Legacy
functionality,
especially
with
so
much
logic
then
points
the
old
endpoints
and
then
there's
so
much
stuff
with
contract
and
IP
tables
and
caching-
and
you
know
thousands
of
nodes.
B
A
B
So
there's
two
roads
we
could
go
down.
One
is
we
could
we
could?
We
could
basically
keep
the
grpc
functionality
of
kpng
and
say
we're
going
to
use
this
as
is,
and
then
we
literally
could
maintain
it
inside
of
kpng.
If
we
wanted
all
of
it.
The
other
thing
we
could
do
is
we
could
just
make
a
Sig
windows,
Coupe
proxy
repo
under
kubernetes
sigs,
and
we
could
build
a
proxy
that
did
exactly
what
we
wanted
to
like
and
like
whether
we,
whether
we
use
kpng
or
not,
is
really
an
implementation
detail.
B
We
could
decide
on
as
a
seg
right.
That
would
be
really
a
conversation
with
the
Microsoft
folks
I.
Think
of
like
do
we
see
the
grpc
strategy
as
strategic,
that
that
would
be
the
thing
and
then-
and
so
it's
like
we
could
rethink
this
whole
whole
thing
and
I.
B
Think
the
default
easiest
thing
to
do
would
be
to
just
sort
of
do
it
in
kpng,
because
we
don't
have
to
make
a
new
repo,
and
we
already
know
that
the
logic
there
is
pretty
pretty
solid,
because
it's
working
with
other
cool
proxies
and
at
least
there's
some
Community
there
that
we
can.
We
can
sort
of
rely
on,
but
there's
just
a
lot
of
ways.
B
We
could
think
about
this
and
I
think
it
really
depends
on
who's
going
to
work
on
it
and
who's
who's
going
to
be
able
to
have
the
time
and
it's
a
crazy
situation
right
now,
I've
seen
people
in
Upstream
getting
extremely
busy
with
with
so
many
things,
so.
A
C
B
He
has
that
document
and
I
forgotten
I
lost
that
document
that
he
had
sent
out.
He
went
through
some
of
the
build
details
right
in
there
yeah
so
I
guess
we
should
have
a
document,
then
that
actually
says
here's.
What
we're
going
to
do
right.
F
E
Yeah
I
don't
know
mine
talks
about
the
build
and
stuff
it's
more
about.
We
leaked
just
because
of
the
way
you
proxy
for
Windows
was
designed
and
and
grew
up.
We
leaked
a
little
bit
of
the
cni
implementations
into
qproxy,
and
so
mine
was
really
about
making
sure
we
remove
the
cni
aspect
of
that
yeah
and
so
I
think.
Maybe
the
the
one
thing
like
a
brief
document
that
defines
you
know
what
the
plan
is,
but
then
also
like.
How
are
we
going
to
handle
like
the
different
kubernetes
versions?
E
I
think
that's
something
that
would
be
one
of
my
concerns.
If
we
moved
out
of
trees
like
how
do
we
make
sure
that
you
know
we're
interacting
with
all
the
different
versions
that
are
supported
in
kubernetes.
A
Yeah,
usually
like
that's
just
taking
a
dependency
on
all
the
networking
apis
that
kubernetes
exposes,
so
assuming
that
everything
is
following
the
plus
or
minus
one
version.
I
think
it
shouldn't
be
too
hard
to
keep
like
I.
Think,
there's
just
going
to
be
like
a
rolling
kind
of
compatibility
like
compatibility
story
for
this
too,
like
the
current
releases
now
are
on
this
API
version.
So
they'll
work
with
clusters,
you
know
plus
or
minus
whatever,
one
of
whatever
that
API
version
is.
But
that
is
something
to
think
about.
B
Yeah
I
mean
well
the
other
thing
we
could
do
if
we
wanted.
If
we
felt
like
staying
in
tree,
it
was
valuable,
as
we
could
put.
We
could
just
go
out
of
tree
entry
right
and
we
could
just
have
our
own
directory
called
Windows
Coupe
proxy
under
the
sigwood
AutoZone
directory
and
maintain
all
our
code.
There
right
I
mean
there's
no
fundamental
advantage
to
us
being
under
PKG
proxy.
B
Is
there
I,
don't
think
there
is
like
if
we
just
have
our
own
implementation,
so
there's
like
so
many
it's
like
I
should
maybe
make
a
list
of
the
five
or
six
possibilities,
and
we
should
just
look
at
them
all,
because
I
I
suppose
there's
like
a
few
different
ways.
We
could
do
this,
you
know,
and
it
it
comes
down
to
how
much
we
want
to
do
it.
B
You
know,
because,
because
we're
not
going
to
get
it
for
free,
it
feels
like
I
mean
it
feels
like
the
KP
G
stuff.
We're
gonna
have
to
build
an
interface
before
between
the
grpc
stuff,
and
so
there's
just
going
to
be
so
much
time
for
that
to
happen.
A
C
E
A
We
expect
that
to
change,
though,
because
if
it's
the
grpc
interface
works
today,
right
so
like
we
could
just
is
it
is
it
going
to
is,
is
the
like?
Are
the?
Are
the
interfaces
going
to
change
a
lot,
or
is
it
just
going
to
be
replaced
with?
You
know
a
golang
interface,
which
is
what
I
kind
of
thought
was
happening
that
should
be
I'm,
not
I'm,
guessing
I'm
not
expecting
300,
but
should
be
pretty
much
a
drop
in
replacement.
A
B
Yeah
I
mean
the
the
replacement
is,
is
pretty
similar
to
what
you
originally
were
working
you
when
you
were
doing
your
stuff,
it's
pretty
similar
to
what
I
showed
you,
then
the
only
difference
is
that
you
know
there's
going
to
be
a
couple
of
extra
potential
methods
to
implement.
You
know
what
I
mean.
It's
it'll
be
something
like
that.
B
The
problem
is
that
everybody
who
was
working
on
it
got
super
busy
like
Andrew
Stokes
at
Red,
Hat
got
busy,
I
got
busy
like
and
then
but
I'll
be
back
in
the
US
next
week,
and
so
that
then
the
question
is
like
okay,
how
do
what's
the
most
efficient
thing
to
focus
on
so
it's
like
I
should
just
put
the
high
level
options
out
and
then
come
back
to
Sig
next
week
and
and
then
we
can
talk
about
it
because
there's
like
these
five
different
things
and
I
think
until
they're,
each
one
is
written
down
with
at
least
one
or
two
sentences.
B
B
Yeah
exactly
and
you
know,
because
I'm
I'm
also
okay,
with
the
idea
that
okay,
this
was
an
experiment
that
we
need
to
put
on
hold
and
just
get
back
to
doing
fundamental
Windows
stuff.
Again,
it's
just
there's
so
many
ways
to
think
about
this.
Yeah.
A
B
A
Like
the
code
was
mostly
lifted
out,
so
we
do
want
to
figure
out
how
to
remove
some
of
those
assumptions.