►
From YouTube: Kubernetes SIG-Windows 20211102
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
Hello,
everybody
and
welcome
to
the
november
2nd
2021
instance
of
the
kubernetes
sync
windows
community
meeting.
As
always,
these
meetings
are
recorded
and
uploaded
to
youtube
so
be
sure
to
adhere
to
the
cncf
code
of
conduct
all
right
I'll
start
with
a
couple
of
announcements.
Code
freeze
is
coming
up
code
freeze
for
the
1.23.
A
My
release
is
two
weeks
from
today
so
november,
16th
yep,
please
work
to
get
all
of
your
pr's
out
early
to
give
reviewers
enough
time
to
take
a
look
and
for
you
to
be
able
to
respond
to
any
feedback.
That's
needed
too
kind
of
similar
to
that
docs.
There.
The.
A
For
any
enhancements,
the
docs
team
has
asked
that
there's
a
placeholder
pr
open
by
next
or
by
the
following
thursday
for
any
enhancements,
or
I
guess
other
docs
related
prs.
So,
if
you're
tracking,
an
enhancement
that
was
marked
as
required,
docs,
please
open
that
up.
That
is
mainly
so
that
the
enhancement
team
knows
who
to
kind
who's,
going
to
be
responsible
for
updating
the
or
for
making
those
docs
updates
and
just
for
tracking
purposes.
A
So
they
just
just
in
a
blank
draft.
Pr
is
usually
okay
at
that
point,
but
same
same
thing.
With
the
code
prs
give
the
docs
teams
as
much
time
as
you
can.
It
really
helps
provide
high
quality
docs
for
people
last
announcement,
at
least
for
me,
is
that
the
steering
committee
elections
are
currently
ongoing.
The
elections
close
on
november
4th.
A
So
if
you're
planning
on
voting,
which
you
should
vote
early,
there's
a
new
site,
it's
just
linked
here
that
will
have
you
sign
in
with
github,
and
if
and
it
will
let
you
know
if
you're
eligible
or
not
and
then
we'll
just
give
you
instructions
on
how
to
vote
and
show
profiles
and
bios
for
all
the
candidates.
So
it's
a
very
important
part
of
the
open
source
community.
Is
you
know,
voting
for
your
representation?
A
A
Okay,
so
the
first
agenda
item
was
going
to
be:
jason:
hall
was
going
to
do
a
demo
on
tecton
workloads
on
windows.
Unfortunately,
he
was
messaging
in
slack,
saying
that
there
is
a
he
has
a
conflict
so
and
it's
actually
a
weekly
conflict
at
this
time,
so
we
would
not
be
able
to
present
there's
a
thread
in
slack
that
I
will
link
to
right
now
where
he
offered
to
either
do
a
recording
of
the
video
or
could
work
at
another
or
could
possibly
come
at
another
time.
A
I
had
maybe
suggested
one
of
the
coming
15
minutes
before
one
of
the
meetings
so
and
we
could
replace
the
trash
session
that
we
do
with
with
that.
But
if
people
are
interested
in
seeing
this,
I
am
please
chime
in
on
the
slack
and
we
will
work
on
figuring
out
a
better
time
or
we
could
schedule
something
completely
like
separate
from
the
sig
windows
community
meetings,
if
needed.
A
So
yeah
yep
just
a
comment
in
there
if
you're
interested
and
we'll
try
and
get
that
scheduled
in
the
near
future.
Next
up
finishing
off
operational
readiness,
I
mean
is
this
you,
since.
B
Yeah
jay
will
be
late
today,
so
we
can
start
on
this
one.
Basically,
we
want
to
push
forward
and
merge
this
cap
to
start
work
and
maybe
discuss
here
what's
missing
for
for
this
move
forward.
B
C
Yeah,
so
I
I
read
through
it
and
it
seems
like
a
lot
of
feedback,
was
addressed.
There's
two
two
things
one.
I
really
like
how
it
was
kind
of
split
up
into
different
sections,
so
you
could
run
like
your
base
and
then
you
could
run
like
the
gmsa.
C
You
could
run
the
networking
and-
and
that
was
really
nice.
I
enjoyed
that
and
I
think
it's
a
good
way
to
kind
of
break
break
up
all
the
different
things
that
windows
can
do.
The
second
part
was:
where
does
this
fit
into
conformance,
and
how
does
the
conformance
like
like?
Are
we
gonna
have
an
official
performance
reporting
somewhere
when
you
meet
one
of
these
sub
categories
that
you've
outlined
here
or
like?
Where
does
that
that
that's
what
I
seem
to
be
missing
from
the
cap
and
understanding
what
the
outcome
is.
B
I
think
it
makes
sense
to
have
a
conformer
specification
and
like
have
this
as
part
of
hybrid
multi,
node
spec,
so
I'm
not
sure
how
this
fits
and
how
you
can
push
forwarding
the
conformance
like
officially.
B
A
Yeah,
I
think
if
we
want
to
kind
of
interact
with
the
official
definition
of
conformance,
we
would
want
to
loop
in
sig
architecture.
Who
kind
of
is
the
overseers
of
conformance
for
for
this.
But
I
think
before
we
do
that
we
should
have
an
idea
of
what
we
would
like
and
hopefully
have
it
in
there.
C
Well,
so
there
is
no
windows
conformance
right
now.
You
can't
like
report
that
your
windows
node
is
conformant
and
that's
where
it
would
be
nice
to
be
able
to
come
in
and
say
you
know.
If
you
run
this
setup
set
of
tests,
it
would
report
as
conformant.
We
take
the
as
you
called
out
in
this
document.
Here
we've
got
like
the
linux
tests
that
are
marked,
as
conformance
that
aren't.
C
Linux
only
is
kind
of
what
we
run
as
our
suite
of
tests
right
now,
plus
a
set
of
windows
tests,
and
I
think
if
we
can
say
this
bucket
is
the
conformant
bucket
and
then
there's
a
couple
other
ones
that
you
would
maybe
mimic
to
be
conformant
in
the
active
directory
spec.
C
That
would
be
what
we
could
go
to
seek
architecture
with
and
and
and
say
there
as
long
as
we
can
agree
that
you
know
this
is
the
set
of
tests
as
as
a
sig,
and
then
we
can
go
to
them
and
say
this
is
what
we'd
like
to
have
conform
it
and
then
maybe
have
these
other
sub
components,
because
I
think
there's
there's
like
there's
another
concept
that
they've
been
trying
to
move
forward
with
conformances.
A
A
B
Yeah,
maybe
I
need
to
I
need
to
discuss
this
with
jay,
but
maybe
maybe
the
first
phase
would
be
have
this
as
a
redness
and
and
like.
B
A
And
it
might
be
easier
to
to
reason
about
this
if
there
is
like,
if
all
of
the
work
is
done,
and
it
is
for
operational
readiness
and
then
take
that
to
the
architecture
team.
B
E
A
I
I
was
just
wondering
I
think
I
wonder
if
this
is
just
a
markdown
formatting
issue.
E
Oh
and
secondly,
regarding
networking
does
that
include
both
network
overlay
and
network
sdn
bridge,
because
there
are
a
couple
of
limitations
when
it
comes
to
overlay
networks,
and
there
are
certain
scenarios
that
we
also
skip
as
tests
in
cis.
E
So
we
might
have
to
address
what
conformance
means
in
regards
to
networking
in
regards
to
those
issues.
If
you
remember
those
issues
mark.
A
E
A
E
Yeah
we
do
have
an
issue
for
that.
We
should
dig
it
up.
A
That
yeah,
but
that's
a
a
great
thing
to
bring
up,
is
yeah.
We
should
say
we
probably
should
differentiate
between
overlay
and
bridge
networking.
C
B
Claudia,
do
you
have
the
issue
number
later,
so
I
can
bring
us
a
comment
or.
A
Yeah,
so
I
guess
I
mean:
is
there
anything
else,
you're
looking
for
just
more
reviews
and
hope,
like
hopefully
some
approvals
or
I
mean
and
approvals
for
this
to
move
forward
or.
B
Exactly
I
would
change,
maybe
the
network
overlay
and
breed
and
do
the
bump
for
the
124,
and
maybe
we
could
be
ready
to
approvals
or
last
round
of
review.
I.
A
F
F
So
I
mean
that's
where
I
was
confused:
this
defined
the
verification
requirements
for
operational
readiness.
We
are
talking
from
an
infrastructure
like
somebody,
who's
running
a
kubernetes
cluster
right,
not
someone
who
is
about
to
deploy
who's
it
who's,
the
user
of
windows
and
kubernetes-
is
that
right.
F
Yeah,
I
definitely
see
the
use
of
it
being.
You
know
the
customer
being
like
the
end
user
being
more
confident,
because
sometimes
we
do
see
between
distros.
There
is
like
some
deficiency
one
place
or
the
other,
so
I
definitely
see
that
you
know
serving
the
purpose,
but
I
definitely
want
to
get
the
windows
conformance
test
kind
of
thing
going
as
well
for
the
larger
thing
so.
B
Yeah,
okay:
we
have
the
sub
tests
of
set
here.
If
you
go
to
the
spreadsheet,
we
start
to
map,
what's
missing
what
what
we
have
so
the
idea
is
to
have
at
least
this
amount
of
testing-
that's
guaranteed
that
you
have
the
spec
working
in
the
cluster.
A
E
I
do
have
an
answer
for
that
windows.
Support
was
first
introduced
in
pos
image
3.4.
E
So
that's
probably
the
reason
why
it
isn't
mirrored
on
the
microsoft
registry.
A
If
I
remember
correctly,
we've
the
the
pause
images
that
we
are
making
now
do
support
all
of
the
versions
of
windows
since
windows
server
2019,
so
it
should
be
safe
to
use
the
latest
pause
image
like
3.6
on
even
a
119
cluster
yep,
regardless
on
either
windows,
server,
2019
on
or
any
of
the
sac
releases.
Since
then,
I
think
there's
even.
A
G
A
Yeah
yeah
so
claudia
just
said
that
the
but
prior
to
3
4,
the
pause
images
were
always
built,
like
they
have
a
different
numbering
scheme
for
windows
and
are
published
differently,
but
for
for,
but
people
should
just
use
the
latest
pause
image.
Regardless
of
what
cluster
or
like
what
version
of
cluster
that
they're
on
so
use
the
3.6
pause
image
on
the
119
clusters.
It
should
work
and
it
should
have
performance
improvements.
D
Yeah,
so
I
have
broken
it
down
into
three
pr's.
The
first
one
is
related
to
introducing
the
api.
The
second
is
making
changes
on
the
cubelet
side
so
that
the
os
label
gets
reconciled.
The
third
one
is
related
again.
It's
on
the
cubelet
side,
it's
related
to
say:
if
the
os
field
does
not
match
the
node
label,
cublet
would
reject
the
parts
at
the
cubelet
admission
time,
not
at
the
scheduler
admission
time,
but
at
the
public
admission
time
so
the
first
we
are
related
to
api
introduction.
It
is
merged.
D
Thank
you
mark
and
obviously
jordan
helped
with
the
api
side
of
things
so
that
pr
is
merged.
So
we'll
now
have
a
os
field.
It
is,
it
is
behind
a
feature
gate.
D
The
changes
that
I
need
to
make
on
the
pod
security
plugin,
so
jordan
has
provided
an
interesting
point
there
where
he
says,
even
if
I
make
some
changes
to
cubelet
today,
since
the
cubelet
skew
support
policy
is
safe,
for
there
is
a
say
if
the
aps
server
version
is
n,
we
would
support
n
minus
two
versions
of
cubelet
so
because
of
that,
the
changes
that
I
have
to
make
may
have
to
be
back
ported
and
we
usually
do
not
backport
api
changes.
D
So
what
jordan
suggested
is
the
changes
that
I'm
making
to
part
security
plugin,
where
I'm
relaxing
certain
fields
based
on
the
os
field
that
may
have
to
wait
till
the
beta
is
hit
so
that
cubelet
supported
is
124
instead
of
123.
A
A
D
Yeah,
the
and
also
on
the
cubelet
side.
We
cannot
make
those
changes,
because
if
you
look
at
the
the
cubelet
change
that
I'm
making
on
the
os
spec
on
the
pod
spec
side,
I
have
to
look
at
the
os
field
and
that
will
not
be
available
in
122..
D
So
I
have
to
wait
till
it
graduates
to
beta
and
124
and
by
that
time
hopefully
we
will
have
cubelet
supporting
or
cubelet
understanding
the
osp,
because
124
will
be
the
only
version
that
is
supporting.
D
So
I
think
I'm
I'm
fine
with
that
particular
change.
But
at
this
point
of
time
I'm
waiting
on
the
note
team
to
review
the
changes.
I've
got
some
feedback
last
week.
I
have
made
this
changes
that
were
asked
of
me.
So
I'm
waiting
for
the
note
team
to
look
at
it.
A
Okay,
yeah-
and
I
do
know
that
the
node
team
was
quite
busy
last
week.
That's
when
they
set
their
their
reviewers
last
week
was
when
they
set
their
or
was
it
two
weeks
ago
they
set
their.
A
For
node
features
so
hopefully
that
there's
more
reviewer
bandwidth
but
I'll
start
also
trying
to
ping
the
reviewers
for
those.
D
Yeah,
so
that's
all
I
have
for
now
and
the
changes
that
erwin
mentioned
last
week.
I
think
he
can
give
it
a
shot
now
like
where
both
run
as
user,
and
I
forget
the
other
field,
so
if
both
of
them
are
set,
we
wanted
to
make
sure
that
they
get
rejected
right.
The
part
would
not
get
admitted
at
the
ap
server
admission
time
itself.
So
with
that
field
now
available,
I
think
we
can
give
it
a
shot.
D
A
Awesome
yeah
thanks
for
the
updates
yeah,
I
I
unfortunately
it
does
seem
like
we
may
need
to
wait
until
the
features
in
beta
to
have
all
like
a
majority
of
the
desired
behaviors,
but
either
way
I
mean
the
sooner
we
get
it
it
alpha
is
going
to
get
it
to
beta.
So
yeah.
D
I
think
the
next
few
days
I'll
try
to
push
it
to
beta,
we'll
see
how
it
goes
from
there
unless
I
receive
any
negative
feedback.
A
Okay,
yeah.
Thanks
for
the
updates,
I
can
provide
a
quick
update
with
the
host
process
containers
moving
to
beta
there's
a
couple.
I
think
most
of
the
functional
changes
have
been
merged.
There
wasn't
that
many
there's
one
outstanding
change
that
I
am
also
waiting
for
node
review
on
and
that's
to
add,
metrics
and
then
the
rest
are
test
changes
so
hopefully
we'll
be
promoting
that
to
beta
pretty
soon
for
the
123
release.
A
All
right
does
anybody
else,
have
anything
jay.
Are
you
going
planning
and
doing
that
yeah.
G
Yeah
definitely,
and
if
folks
I
just
wanted
to-
I
see
y'all
said
conformance
versus
readiness,
but
I
think
we
all
decided
we
just
we're
going
to
not
call
it
conformance
at
all
right.
So.
C
C
And
I
think
amin
kind
of
said
you
know,
let's
get
this
out
as
the
readiness
now
get
it
finalized
and
then
maybe
we
can
talk
about
moving
it
into
conformance
in
the
next
phase
of
this
whole
thing,
because
I
do
think
that
there's
some
level
of
importance
there
for
for
performance,
because
we
do
get
a
lot
of
questions
like
how
do
I
say
my
windows
nodes
are
conformant
and
what
you
know
there's
a
lot
of
confusion
there
and
so
it'd
be
nice
to
be
able
to
have
folks
feel
the
publish
and
say
hey
my
aks
cluster,
my
bm,
where
cluster
my
aws
cluster
is
conformant
for
windows
nodes
eventually,
but
yeah.