►
From YouTube: Kubernetes SIG-Windows 20230321
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
This
is
our
weekly
community
meeting.
It
is
a
cncf
project,
and
so
we
follow
the
cncf
guidelines
and
code
of
conduct.
So
if
not
familiar
with
those
that
just
means
to
be
nice
to
everyone,
and
if
you
have
any
concerns
or
questions
you
can
reach
out
to
one
of
the
leads
or
any
of
the
cncf
representatives.
A
So
otherwise
we
can
go
ahead
and
get
started
today
for
announcements
we
have
test.
Freeze
is
today,
so
if
anybody
has
anything
related
to
tests,
let
us
know
we
did
have
a
failing
test
from
GCE,
which
we
were
just
discussing
before
this
meeting
in
our
triage
meeting
and
I've
added
it
to
the
end
of
the
agenda
here.
So
we
can
discuss
it
here
in
a
few
minutes,
but
is
there
anybody
else
that
has
anything
related
accessories
that
we
should
know
about.
A
Cool
all
right,
then,
we'll
go
ahead
and
move
on
to
new
contributors.
Welcome
to
any
new
contributors
out
there
is
there
anybody
that
would
like
to
say
hello
and
let
us
know
what
you're
up
to.
A
If
not
we'll
just
move
on
here
great,
so
we'll
go
ahead
and
start
the
agenda.
B
The
kpng
direction
later
I.
C
D
B
A
B
I
will
I
just
wanted
to
quickly
mention
the
annual
reports,
so
the
annual
report
is
something
that
each
Sig
is
required
to
do
once
a
year
like,
and
it's
mainly
just
a
mantra
like
the
health
of
the
community.
The
annual
reports
draft
PR's
was
required
to
be
open
by
this
Friday
I've
got
a
draft
PR
open
for
the
Sig
Windows
annual
report
linked
here
everybody
or
anybody.
Please
take
a
look
at
that
and
leave
any
feedback.
You
have.
B
Yeah
I
think
now
we
could
probably
go
back
to
talking
about
this
issue
for
anybody
who's
new
to
the
or
just
during
the
meeting
there
we're
looking
at
a
there's,
a
slack
thread
and
an
issue
and
a
pull
request
to
hopefully
resolve
the
issue
around
a
change
in
behavior
for
the
DNS
resolution
for
Windows
that
we're
looking
at
it
looks
like
it
caused
some
issues
with
how
the
GCE
tests
are
configuring
things.
So.
A
Yeah,
so
I'll
drop
the
issue
into
the
chat
here.
So
this
is
the
issue
and
then
here's
PR
quadu.
Do
you
want
to
give
a
quick
overview
for
everybody.
D
Yeah,
basically,
there's
a
flag
for
cool
blood
called
which
typically,
it
is
supposed
to
be
a
file
which
is
how
it
works
on
Linux
as
well,
which
will
basically
contain
DNS
configuration
options
which
would
also
get
applied
to
pods.
For
example,
if
you
have
some
classes
specific
DNS,
config
options,
you
might
actually
want
to
have
that
option
set.
D
Typically,
let's
say
on
Linux.
You
would
get
the
slash,
UTC,
slash
result.com
file
and
apply
those
configuration
options
which
you
have
on
your
horses
hosts
as
well,
and
we
have
added
the
same
ability
to
get
the
host
DNS
options
from
Windows
as
well.
If
you
set
the
flag
to
dash
dash
resolve,
Dash
conf
equals
to
host
with
a
capital
H,
but
in
doing
so
we
introduced
a
small
regression
which
is
supposed
to
get
fixed
in
a
pull
request
which
is
sent
for
this.
D
Basically,
the
regression
is
that
in
doing
so,
we
removed
the
ability
to
have
the
same
users.com
like
files
available
on
Windows,
which
will
have
to
be
introduced
so
on
in
this
fully
pass,
which
is
shown
right
now,
but
yeah.
That's
pretty
much
the
the
gist
of
it
since
I
have
some
comments
to
address.
B
D
D
B
A
So
is
this:
if
you
set
up
setting
as
host,
then
it
would
look
it
up
in
the
registry
versus
just
looking
it
up
on
the
file
yep.
D
D
I
mean
I
in
the
original
discussion
on
the
pull
request.
We
basically
agreed
on
post
as
a
name
and
now,
apparently
that's
being
questioned
as
well.
A
Yeah
I
mean
I
I
guess
my
take
would
be
that
without
the
full
file
path
host
is
okay.
Yeah
typically
you'd
give
a
full
file
path
in
Windows.
B
E
B
He
asked
I,
don't
think
he
said
it
is
I
think
he
just
asked
us
to
confirm.
There's
a
relative
path
work
today.
Oh.
E
But
then,
if
you
just
say
it
equals
host,
do
you
just
assume
it's
SQL
and
slash
host.
A
D
We
could
check
if
there's
a
fire
path
for
that.
If
so,
we
can
just
shoot
it
as
a
regular
result.com
file.
D
E
B
D
And
in
regards
to
the
path
character
limitations,
while
using
abstract
paths,
I,
remember
that
Windows
you
used
to
have
or
still
has
a
limit
on
how
many
characters
you
have
as
a
path
unless.
B
B
D
D
B
E
Just
remind
me,
given
I'm,
still
coming
up
to
speed
on
all
the
suggest
stuff
and
all
the
there
was
a
request
to
send
out
a
video
right
for
that's
separate
to
this.
That's.
B
A
separate
request,
so
there's
also
I-
think
I'm-
should
have
mentioned
this
here
too,
so
for
kubecon.
B
So
I
guess
we
could
just
decide
right
now.
If
we
want
to
do
that
and
I.
C
Think
we
can
definitely
definitely
do
it,
I'll
do
it
unless
somebody
else
wants
to
do
it.
I
have
no
desire
to
do
it,
but
I
will
totally
do
it.
Unless
someone
else
wants
to
do
it.
Okay,.
C
Does
anybody
want
to
do
it
if
I
mean
because
I
I
could
totally
do
it
and
I
don't
mind
doing
it
I?
Just
if
somebody,
especially
somebody
new
to
the
community?
Here
we
got
a
few
people
hanging
out.
Somebody
wants
to
do
the
video
I'm
happy
to
help
you
create
the
content
and
we
can
get
arvind
and
everyone
else
and
and
Mark
and
James
and
Claudio
to
help
us
figure
out
what
we
want
to
say.
B
C
E
B
E
B
E
C
Anyways
yeah.
Definitely,
if
someone
wants
to
do
it
or
we
could
do
it
together,
I'm
happy
to
do
it
with
someone
else
too.
I
just
want
to
make
sure
that
we
don't
miss
the
deadline
to
do
it.
So
that's
why
I'm
saying
I'm
gonna:
do
it
yep.
B
C
Cool,
so
how
about
this
Mark?
You
can
start
the
thread
once
you
take
care
of
the
deadline
and
then
I
will
put
a
Google
doc
up
and
people
can
add
whatever
they
want
and
I'll
bug
folks
and
by
Thursday
whatever's
in
there
I'll
I'll
present.
Unless
somebody
says
they
want
to
present
it
and
I
think
it
would
be
great
to
have
a
new
face
in
the
sink
presenting
something
so
strongly
occur.
Anybody
on
the
call
who
wants
to
do
this,
okay,
cool.
A
Okay
and
we'll
start
a
thread
for
the
127
major
themes
too,.
C
Yeah
overall
I
think
what
what
we're
kind
of
looking
at
is.
We
don't.
We've
written
we've
written
a
cap
a
few
times
for
kpng
and
nobody's
really
had
time
to
finish
the
cap
and
like
really
have
all
the
debates
necessary
to
to
make
it
an
official
entry
replacement,
and
this
is
a
sick,
Network.
C
And
I
think
the
overall
direction
is
that
the
focus
on
Sig
Network
side
is
going
to
be
really
really
on
the
Linux
side
of
things.
I
mean
we
could
refactor
things
and
and
try
to
get
everything
in
tree,
but
I
don't
even
know
if
that's
the
right
direction,
because
I
see
so
many
people
making
out
of
tree
proxies
Andrea
is
one
of
them
psylliums
and
other
calicos
a
third.
C
We
would
not
be
the
first
people
to
be
doing
this
so
I
think
the
the
current
direction
the
way
I
see
it
is
just
to
keep
kind
of
doing
what
we're
doing
and
see
if
we
can
actually
get
it
to
work,
this
new
windows
out
of
tree
proxy
and
then,
if
so,
if
we're
successful
and
people
adopt
it
and
start
using
it,
then
I
think
we
can.
We
can
really
volunteer
to
like
hey.
This
seems
to
be
working.
C
Do
you
want
us
to
take
the
auditory
the
proxy
out
of
tree,
or
do
you
want
to
keep
it
around
forever
or
you
know?
I
mean
I
feel
like
that's
where
things
are
going
and
I
think
on
the
kpng
side.
I
think
kpng
is
a
convenient
tool
for
us
to
implement
this
with,
but
I
don't
think
it's
I,
don't
think
it's
essential,
I!
Think
long
term.
C
If
Sig
Network
comes
out
and
writes
some
generic
batteries
included
across
platform,
unopinionated
proxy
implementation,
that
does
what
kpng
does
I
think
I
mean
I,
don't
see
that
happening
anytime
soon
but
like
if
that
did
happen,
I
don't
see
any
reason
why
we
couldn't
just
use
that
for
for
our
windows
out
of
tree
proxy
right,
so
I
was
talking
to
Mark
and
we're
like
yeah.
C
Let's
do
let's
see
if
we
can
get
this
up
and
running
really
working
within
six
months
and
if
so,
we'll
just
we'll
just
keep
using
kpng
for
everything
and
it's
within
that
time.
Something
better
comes
out,
we'll
start
rebasing
onto
that,
but
I
think
the
hard
part
is
getting
it
all
working
getting
CI
around
it,
making
it
so
that
it
passes
the
node
Port
test,
which
it
fails
right
now,
so
so
we're
kind
of
leaning
towards
doing
the
hard
part
first,
as
opposed
to
doing
the
paperwork.
C
C
My
focus
is
mostly
on
seeing
if
we
can
get
developer
environments
up
so
that
anybody
can
contribute
to
it
easily
and
I
think
we
think
we're
getting
closer
with
the
qamu
ones
now,
but
I
know
Mike
you're
interested
also
so
what's
on
your
mind
or
anyone
else,.
F
F
Anything
you
know,
I
was
always
just
curious
about
Direction
I,
think
Dan
Winship
said
it
pretty
well
in
the
thread
that
you
started,
but
you
know
ideally,
you
know
the
cap
would
get
finished
in
some
shaper
capacity.
However,
just
getting
kept
through
is
you
know,
pulling
teeth
and
chopping
your
fingers
off,
so
it's
not
the
most
fun
yeah,
so
I
I
have
really
no
opinions,
I
mean
if
we
do.
If
the
cap
does
get
written,
then
in
a
way
that
Sig
Network
can
accept
that'll
be
great
because
then
it
can
be
presented.
C
Yeah
yeah,
I,
yeah
I,
definitely
actually
I
saw
that
threat
and
I
responded.
I
was
like
yeah
I,
definitely
think
it's
a
great
idea
to
write
a
cup.
If
someone
wants
to
go
off
and
do
it,
I
just
I
think
we're
four
cups.
Deep
at
this
point
and
I
don't
see,
I,
don't
see
a
fifth
one
being
any
more
successful
right
like
I,
don't
you
know
I
think
at
this
point
even
Sig
network
isn't
sure
that
they
want
to
build
a
generic
who
proxy
replacement
like
I.
C
You
know
I
think
it
benefits
people
who
are
using
non-linux
native
technology
more
than
it
benefits
people
using
Linux
native
stuff,
because
the
Linux
native
stuff
can
just
go
in
tree.
So
why
would
you
write
in
gen
generic
implementation
right?
So
I
was
just
talking
to
the
Andrea
folks
yesterday
and
like
they
have
the
same
problem
right
of
vendoring
and
all
that-
and
so
you
know.
C
B
My
concerns
about
this
are
still
kind
of
the
same
as
when
we
talked
about
moving
or
creating
the
out
of
tree
Cube
proxy
I
think
that
we
were
talking
a
little
bit
about
paperwork
versus
doing
work.
I,
think
we've
got
enough
of
the
paperwork
done
and
we
already
have
the
you
know.
Sub-Project
set
up
that
we
can
move
forward
with
the
attitude
proxy
without
the
Sig
Network
cap.
B
My
and
I
don't
think
anybody's
going
like
we
can
just
ask
for
forgiveness
if
we
need
to
I,
don't
think
or
I
still
I'm
kind
of
under
the
opinion
that
I
would
rather
not
have
to
have
Sig
Windows
maintain
all
of
the
generic
or
kpng
logic
around,
like
you
know,
respond
like
reading
the
API
server
responding
to
that
maintaining
the
grpc
the
grpc
kind
of
the
the
protobuf
contracts.
B
Just
because
it's
not
kind
of
our
expertise
but
I
am
I,
do
think
it
makes
sense
to
maintain
the
you
know,
the
wind
kernel
pracs
here
and
the
parts
that
actually
does
the
agents
programming
out
of
tree.
So
it
would
be
great
if
you
know
Sig
Network,
would
you
know
volunteer
or
like
support?
The
the
generic
parts
of
the
Q
proxy
and
the
API
server
watching
and
all
of
that,
so
that's
really
like
my
biggest
concern,
is
I.
Don't
know
like
what
know
what
to
do
with
that,
because
I
think
that's
just.
C
B
C
Yeah
I
mean
that's
for
sure.
Well
so
so
I
mean,
if
you
look
at
KPMG
I'd,
say:
there's
more
people
contributing
to
kpng
right
now
than
there
are
the
entry
Coupe
proxies.
So
if
you
go
look
at
open,
pull
requests,
you've
got
five
different
people
from
five
different
companies
right
so
yeah,
you
know
right.
C
Part
of
the
part
of
the
the
goal
with
kpng
is
make
it
easy
and
fun
for
people
to
work
on
the
cooproxy
and
I.
Think
that's
been
successful.
I
think
the
question
is:
how
do
we
Mentor
people
to
get
to
that
next
level?
Right?
Where
we
can?
We
can
get
more
and
more
people
other
than
Mikhail
stoichus
people
like
who
really
know
the
code
base
inside
out
and
so
I
think
there's
a
ton
of
community
work.
We
have
to
keep
doing,
but
it's
it's
definitely
far
from
you
know.
A
So
I
guess
it
seems
like
there's
always
going
to
be
a
place
for
some
out
of
tree
Q
proxy
implementations
like
right,
like
there's
just
like
the
injury
thing
is
just
not
gonna
meet
everybody's
needs,
but
it
makes
sense
to
have
an
entry
and
it
makes
sense
to
build
a
community
around.
You
know,
building
an
out
of
tree
one
for
those
folks,
like
you
mentioned
early
on
and
sick
Windows
would
just
be
part
of
that.
So
I
think
it
seems
like
there's.
A
B
A
B
C
Yeah
or
or
we
could
do,
the
other
thing
me
and
Mark
talked
about
is
if,
if,
if
people
really
want
us
to
stay
intrigued,
I
think
there's
I
think
it's
easy
to
make
an
argument
to
not
be
intrigued.
Given
the
direction
of
all
of
kubernetes,
you
know,
but
if
people
really
have
strong
opinions
that
kubernetes
needs
to
ship
with
an
entry,
Windows
Coupe
proxy
then
like,
like
you
know,
we
we
could
just
go
100
entry,
but
I
think
there's
a
there's,
a
middle
ground
where
we
could
say.
C
C
Stop
maintaining
it
yeah,
you
know:
if
people
don't
stop
maintaining
it,
it
doesn't
need
to
be
deprecated
that
you
know.
E
C
B
C
B
What
we're
kind
of
proposing
here
is
saying:
let's,
let's
give
the
outer
tree
Q
proxy,
based
on
kpng
six
months
of
to
to
stabilize
and
make
sure
it
can
kind
of
perform
on
par
with
the
entry
queue
proxy.
But
it's
easier
for,
like
Sig
windows
and
other
contributors
to
get
changes
in
and
if
we
can't
do
that,
then
we'll
do
something.
Then
we'll
kind
of
either
go
back
to
focusing
on
the
entry
queue
proxy
or
take
the
entry
code
and
just
make
that
the
out
of
tree
Q
proxy
code
and
maintain
that.
A
F
F
Aka,
like
you
know,
toss
the
ring
into
you
know
Mordor
kind
of
thing
and
get
it
done
so
then
we
can
address
some
of
Mark
Mark's
concerns,
so
we're
not
maintaining
everything
as
if
there's
interest
I
think
that's
all
we
need,
especially
you
know.
If
it's
it's
VMware
and
red
hat
and
I,
don't
know
who
else
you
know
three
four
and
five
was,
and
if
we
can,
you
know,
Wrangle
the
Andrea
folks,
celium
and
Calico
as
well.
To
kind
of
is
there
moving
away?
F
Oh,
they
have
moved
away,
so
it
does
sound
like
there
could
be
some.
You
know
tangible.
You
know
effort
here
and
there
is
interest
to
do
it.
I'm
just
kind
of
curious
if
that's
been
tried
or
if
it's
just
falling
apart.
B
F
C
F
F
However,
it
made
more
sense
to
kind
of
investigate
the
whys
that
was
happening
and
prevent
it
from
happening
or
allow
a
more
seamless
transition
for
that
to
happen,
because,
eventually
you
know,
if
I'm
not
I,
forget
I'm
sure
on
the
Linux
side
in
Windows
side
they
don't
use
KP
at
all
for
the
three
listed
here.
So
you
know
eventually,
I
would
have
just
assumed
they
would
all
everybody
would
leave
and
there
would
be
no
entry.
F
Unless
you
know
there
was
a
cap
to
kind
of
address
that
or
if
you
know
API
Gateway
was
the
You
Know,
the
Way
of
the
future
or
more
I
kind
of
never
got
an
answer
to
that
question.
So
that's
what
kind
of
was
I
gonna
ask
earlier
too.
C
I
mean
I
feel
like
there
needs
to
be
a
camp.
I
just
feel
like
the
people
who
have
to
write
the
cap
have
to
be
the
people
who
are
the
you
know
it's
like
Tim
or
Dan,
or
someone
needs
to
write
that
Gap
I.
Think
like
yeah,
you
can't
guess
what
what
the
right
answer
is
to
this
problem
right,
like
there's
a
thousand
possible
answers
to
it
right.
So
there's
it's
just
a
NP,
complete
problem
to
try
to
guess
what
people
will
agree
on
right.
C
That's
kind
of
what
kpng
gives
you,
though,
that's
right
like
you
get
that
for
free
with
KPMG
right
so
I'm
like
that's,
that's
why
I'm
like
we
just
go
with
what
works
and
then,
if
somebody
writes
a
better
interface
we
can
share.
We
can
re-implement
another
interface,
but
I
feel
like
our
problem
has
been
solved.
A
Time
here
and
I
don't
know
that
we're
going
to
solve
yeah
networking
thing
here,
but
I
think
we've
generally
agreed
to
move
forward
with
the
Q
proxy
for
Windows.
How
to
treat
for
now
with
the
timeline
and
then
we'll
check
in
in
six
months
and
see
where
we're
at
timer.
A
Awesome
so
yeah
we're
quite
a
bit
past
time,
so
I'm
gonna
wrap
up
the
meeting
here.
There
is
one
last
thing:
there's
a
PR
to
check
out
for
qmu,
go
ahead
and
click
on
that
link
and
give
a
review.
Otherwise,
we'll
see
you
all
next
week.
C
Okay,
I
can
hang
out
a
little
longer,
I,
see
Andy
and
other
folks
are
here
in
case.
Anybody
wants
to
I've
got
another
10
minutes,
foreign.