►
From YouTube: Kubernetes SIG Windows 20211130
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
30th
2021
instance
of
the
kubernetes
sig
windows
community
meeting.
As
always,
these
meetings
are
recorded
and
uploaded
to
youtube
so
be
sure
to
here
to
the
cncf
code
of
conduct.
I
was
mentioning
before
we
started
the
recording.
We
have
a
very
light
agenda
today,
so
if
anybody
has
anything
to
talk
about
what
I'd
like
to
discuss,
just
feel
free
to
add
it
in
the
chat
or
edit
the
agenda,
if
not,
we
can
probably
get
some
time
back,
but
so
for
announcements.
A
The
only
announcements
we
really
have
today
are
that
today
is
when
the
enhancements
team
would
like
to
have
all
of
the
docs
vr's
for
enhancements,
ready
for
they'd
like
to
have
them
reviewed
and
lgtm
to
buy
the
subject
matter,
experts
so
that
the
docs
team
can
do
a
final
review,
for
you
know
anything
like
formatting
and
kind
of
structure,
so
I
think
we're.
A
I
see
that
we
still
have
some
of
the
prs
open
for
the
moving
host
process,
containers
to
beta
and
moving
the
or
and
adding
the
the
os
fields
to
the
pod
specs.
I
think
there's
a
couple
other
minor
updates
too,
that
we
can
track,
but
those
usually
aren't,
subject
to
the
same
enhanced
by
the
same
deadline
brendan.
Do
you
have
any
comments
on
the
the
post
process?
Containers
updates
one.
I
think
that
we're
kind
of
just
waiting
for
a
little
bit
of
feedback
from
tim
is
that
correct.
B
Yeah,
I
think
that's
about
it.
I
think
it's
a
good
idea
to
create
the
the
pr
into
the
122
branch
to
update
a
requirement
for
container
d
1.6,
so.
A
A
A
Pulling
up
the
notes
in
a
different
field,
doubling
here,
okay,
so
here's
this
one
so
yeah
as
we
were
just
mentioning
we're
host
process,
containers
and
123
do
require
container
d
1.6.
Now,
since
we've
removed
some
annotations.
A
We'll
just
kind
of
stay
on
top
of
that
ravi.
Are
you
on
the
call
today?
Do
you
have
anything
you'd
like
to
discuss
about
the
the
the
updates
for
the
pod
os
field.
D
Yeah,
I
think
they
have
addressed
most
of
the
comments,
but
I
think
james
had
one
more
comment
with
respect
to
effect,
so
I
think
we
just
need
to
clear
that
up
as
far
as
comments
from
tim
goes,
I
believe
I
have
addressed
most
of
them,
but
he
also
mentioned
about
the
cube
cuddle.
D
Like
say:
if
we
use
an
older
version
of
jupiter,
it
may
have
some,
it
may
feel
the
validation,
so
I'm
wondering
how
it
is
being
addressed
like
in
other
places
like
say
if
we
introduce
a
new
field
and
if
you
use
an
older
version
of
cute
cheap,
how
is
it
being
used
or
how
is
it
being
mentioned
in
the
document?
So
if
I
get
that
information
back,
I
think
I
can
include
it.
A
My
understanding
and
I
actually
ran
into
this
with
some
post
process
container
work
that
I
was
doing
is
that
there
really
isn't
anything
that
you
can
do
around
that,
and
that
is
the
reason
why
there's
the
three-staged
approach
for
adding
new
fields,
the
alpha,
the
beta
and
then
stable,
and
the
really
the
best
that
you
can
do
is
like
document
and
say
this
is
alpha
in
123.
A
You
need
to
use
a
123
version
of
cube,
ctl,
2
or,
and
also
the
client
apis,
if
you're
doing
it
programmatically
to
do
to
to
kind
of
interact
with
this.
A
I
think
yeah
claudia,
do
you
have
any
other
experience
with
this
or
like
because
I
know
you've
worked
on
some
other
features
too,
but
I'm
I'm
pretty
sure
that,
like
the
best
that
we
can
do
is
just
add
a
warning
and
probably
just
describe
some
of
the
symptoms.
People
may
see
if
they're,
using
mismatched,
if
they're,
using
like
clients,
libraries
or
tools
that
don't
understand
the
new
fields.
E
A
Oh
so
there's
some
comments
in
the
in
one
of
the
docs
for
the
new
os
fields.
I
think
this
one's
specifically
that
if
people
are
using
in
like
an
older
version
of
cube,
ctl
like
122,
that
doesn't
what
that's
using
api's
definitions
that
don't
know
about
the
new
fields,
it
won't
work.
Is
there
anything
that
we
can
do
or
should
we
just
document
that?
E
E
Yeah,
that
would
be
the
best
approach.
Actually,
it
would
be
a
good
idea
to
have
a
cup
city
that
matches
the
cube
api
server
version
as
well,
so
both
crypt,
ctl
and
qbp
server
will
have
to
recognize
that
field.
A
D
Yeah
the
way
I
used
to
handle
it
previously
is
like
we
will
have
usually
a
blog
post
associated
with
the
new
field
or
the
feature
that
we
are
introducing,
and
that's
where
we'll
mention
this
is
the
version
of
cubepetal
or
the
aps
that
you
need
to
be
on,
but
nowhere
I
have
seen
it
mentioned
in
the
original
dogs.
A
F
D
Anyways,
like
I'll
wait
for
tim's
response,
I'll,
also
ping
him
on
slack,
because
I
think
today's
the
the
code
freeze
right.
Sorry,
the
dogs
freeze,
yeah.
A
Yeah
yeah.
That
would
be
good,
I'll,
try
and
comment
on
here
too.
You
might
want
to
mention
too,
that
we're
planning
on
doing
a
blog
post,
where
we
can
describe
that
in
the
comments.
D
I
mean
I
have
not
planned
on
the
blog
post,
but
the
point
I
was
making
is:
oh
okay,
previous
scenarios,
I
used
to
write
a
blog
post
about
the
new
feature
or
the
field
that
I'm
introduced,
for
example,
for
the
min
ready
seconds
that
I've
been
working
on.
I
have
introduced
one
field
in
the
spec
and
another
field
in
the
status,
so
for
folks
to
know
how
to
use
these
fields.
I
have
written
a
blog
post
and
that's
where
I've
mentioned
cube
cutter
version.
This
is
the
minimum
version
that
you
need.
A
Okay,
yeah,
I
think
that
makes
sense
too.
Then
I
yeah,
I
don't
really
know
what
I
think
we
just
need
to
get
tim's
attention
and
ask
what
he
thinks
would
be
appropriate
here,
because
I
think
it's
kind
of
expected
with
alpha
features
that
there
is
going
to
be
some
version
compatibility
between
the
versions
that
don't
know
about
the
fields
and
the
versions
that
do.
D
Yeah,
it's
not
only
just
about
the
alpha,
but
any
new
field
that
we
are
introducing,
even
if
it's
beta
or
ga.
I
think
it's
going
to
have
because
we're
going
to
support
three
releases
and
say
pga
in
the
the
three
releases
and
if
someone
is
at
a
cubical
version,
which
is
n
minus
two
they're
still
going
to
have
the
same
problem.
A
Yep
and
usually
like,
I
think
that
they
it's
been
pretty
prevalent
that
or
heavily
documented,
that
your
tools
like
you
should
always
do
upgrades
from
like
from
one
version
to
the
next,
don't
skip
versions,
and
this
api
compatibility
is
like
plus
minus
one
version
yeah.
So
hopefully,
that's
sufficient.
G
D
So
I
think
the
tim
on
sick
dogs
and
then
see
if
I
can
get
a
response
from
him
I'll
also
james.
I
think
you
wanted
me
to
change
the
statement
a
bit
right,
so
is
that
this
one
yeah.
C
Yeah
I
just
maybe
it's
just
the
way,
I'm
reading
it,
but
it
rather
that
it
was
has
no
effect
because
it's
behind
a
feature
gate,
but
then
we
go
on
and
say
that
it
will
cause
failures
if
the
pod
goes
there,
and
so
I
just
I
just
just
like
the
wording
just
needs
to
be
a
little
clearer.
I
think,
but
if
I'm
the
only
one
with
with
that
particular
comment,
I
guess
it'll
be
okay.
D
C
Please
note
that
there
is
no
effect
well,
I
I
can't
make
it
off
the
top
of
my
head
here,
but
I
it
just
is
I
it
should
be
when
I
guess,
when
the
feature
gate
is
enabled
it
will,
it
will
cause
it
to
to
do
things,
but
when
it's
disabled-
and
you
set
this
field,
then
it
shouldn't
be,
then
there
won't
be
any
particular.
A
A
A
My
comments
were
addressed.
I
was
watching
the
ones
from
tim
to,
but
he's
probably
he
probably
has
the
best
comments
or
he
knows
what
he's
looking
for.
So
we'll
just
see
what
he
says:
okay,.
D
So
that's
all
I
have
as
far
as
the
comments
on
the
pl
moves.
A
Okay
yeah.
Thank
you.
We'll
keep
an
eye
on
that.
Okay,
I
think,
there's
a
couple
more
announcements
or
a
couple
more
agenda
items
here.
Does
anybody
have
any
other
questions
about
any
of
the
docs
updates
or
the
123
release.
A
If
not,
okay,
we
can
talk
about
the
docker
msft
provider
issue
who
who
added
this.
G
Hey
this
is
peter
here.
I
added
this
issue
and
just
wanted
to
have
a
quick
follow-up
on
this
brandon
already
commented
on
the
issue,
which
was
really
helpful,
so
thank
you
brandon,
for
that.
G
It
sounds
like
basically
docker
microsoft
provider,
which
is
kind
of
the
recommended.
The
long-standing
recommended
way
of
installing
docker
on
windows,
server
releases
brandon
clarified
that
that's
piping
at
this
time.
That's
now
piping
releases
from
mirantis.
So
I
the
thing
is
now
called
the
mirantis
container
runtime.
G
Does
anyone
have
any
other
any
other
options?
Besides,
you
know
container
d
and,
of
course,
we're
doing
both
is
anyone
already
building
mobi
for
windows
direct
from
upstream.
B
So
I'll
kind
of
jump
in
here
and
offer
some
additional
info
yeah.
So
basically
the
story
was
that
you
know
some
time
ago.
I'm
not
sure
if
you're
aware
of
the
history
but
mirantis
bought
pieces
of
docker
and
kind
of
acquired
them
and
which
included
the
docker
engine-
and
you
know
this
whole
time.
We've
we've,
the
docker
microsoft
provider
has
been
pulling.
B
The
you
know,
docker
or
mirantis
would
both
upload
their
their
versions
of
the
docker
runtime
binaries
to
our
kind
of
local
storage,
which
we
would
sign
and
then
provide
to
azure
customers.
So
that's
kind
of
the
history
of
the
the
docker
microsoft
provider,
api
and
that
basically,
is
just
pulling
you
know
whatever.
It
makes
something
a
docker
or
mirantis
make
available.
So
I
don't
know
if
it's
ever
been
in
sync
with
moby.
B
I
think
moby
is
you
know,
because
everything
is
kind
of
moby's
the
original
projects-
and
I
think
everything's
kind
of
based
on
that.
So
I
think
they're
always
going
to
be
a
little
bit
ahead,
but
my
recommended
recommendation
is
to
to
start
working
with
moby
and
try
to
start
building
those.
We
are
kind
of
aware
of
that
and
are
starting
to
work
on
improving
the
experience
for
making
binaries
available
for
moby,
so
that's
kind
of
an
in
progress
thing
on
our
end
so
yeah.
A
A
Do
you
have
any
comments
on
that?
Brandon.
B
H
H
B
H
H
Then
brandon
is
if
you're
going
to
build
the
binaries
right.
Microsoft
is
going
to
build
the
binaries,
would
it
still
be,
would
it
still
be
supplied
using
the
docker
msfd
provider
or
we
would
have
to
go
like
to
github
or
something
and
pull
the
binaries.
B
I
think
that's
still
an
open
question
the
dr
microsoft
provider,
like
so
as
of
right
now,
that's
being
deprecated,
because
those
are
providing
miranda's.
Binaries
and
mirantis
wants
to
have
people
install
the
the
mirantis
runtime
using
their
installation
process
and
their
installation
script.
B
So
when
our
kind
of
you
know,
contract
with
them
ends
at
the
end
of
next
year,
the
the
plan
is
to
deprecate
that
api,
because
it
technically
won't
be
a
docker,
a
binary
it'll,
you
know,
be
a
mobi
binary,
so
I'm
pretty
sure
at
least
that
that
api
will
be
deprecated,
we're
looking
into
putting
together
another
release,
method
or
possibly
another
api
that
people
can
use
to
install
the
mobi,
binary
or
container
d.
A
A
B
Yeah,
the
the
bottom
line
is
that
maranthus
is
kind
of
shifting
to
towards
running
their
own
thing.
So
if
you
want
to
use
the
traditional,
you
know
the
docker
engine
enterprise,
you
need
to
go
through
them,
because
they'll
have
a
whole
licensing
scheme
and
everything.
H
H
B
Yeah,
that's
kind
of
a
discussion.
That's
gone
going
right
now,
so
I'm
still
figuring
out.
What
is
what
it
is
we're
doing
with
the
with
container
builds
the
media
builds
so
that
that's
kind
of
an
improgress
thing.
For
me.
I
don't
know
if
anyone
from
microsoft
here.
C
I
don't
think
there's
at
least
as
of
right
now,
there's
none
available
for
2022
yeah,
so
you
have
to
install
like
the
hyper-v
and
the
container
features
on
the
base
image.
That's
released,
at
least
in
azure,
for
for
2022,
okay,.
B
Yeah,
so
so
I
mean
we
are,
this
is
something
we
are
working
on
and
we
are
aware
of,
and
we
want
to
make
a
a
a
good
and
quality
experience
for
for
everyone
using
this,
so
yeah
I'll
I'll
definitely
have
more
info
in
the
next.
A
B
A
I
I
will.
I
can
also
say
that
the
the
process
that
makes
those
images
does
use
the
currently
used
that
docker
msft
provider,
so
those
images
would
have
the
latest
potential.
Like
the
latest
version
of
docker
ee
or
the
mark
marantis
container
runtime
that's
available
on
docker
msft
provider.
They
wouldn't
have
versions
that
have
not
been
released
to
docker
msft
provider.
A
D
A
docker
enterprise
edition
that
docker
engine
of
the
enterprises
have
moved
like
more
than
a
year
before
to
mirantis.
Actually.
A
D
My
current
understanding-
this
is
like
not
official
one
was
speaking
to
miranda's.
Employees
like
they
are
working
on
this
like
coming
up
with
a
docker
shim,
keeping
it
separate
and
bundling
it
so
that
they
can
continue
to
support
docker
engine
as
runtime
right.
They
are
working
on
it
I'll
check
with
them
about
a
release
date
or
something
when
they
are
trying
to
target
that.
B
They
are
working
on
making
it.
I
think
they
have
made
a
cry
compliant
at
this
point,
so
it'll
still
work
on
time
for
kubernetes.
D
Yeah,
so
I
think
the
idea
is
like
supporting
for
the
kubernetes
yeah,
I'm
not
talking
about
the
docker
enterprise
engine
separate.
This
is
like
still
continue
to
allow
a
docker
engine
as
a
runtime
in
kubernetes.
That's
the
world.
H
F
G
H
G
H
H
That's
why
there
is
that's
why
that
issue
itself
was
was
created.
So
it's
I
think
everybody
wants
to
move
off
of
docker.
I
think
eventually
and
move
to
continuity
as
far
as
windows
goes
but-
or
I
shouldn't
say
everybody,
but
there's
there's
momentum
behind
it
right,
but
it's
like
what
do
you
do
in
the
short
term
like
when
you
haven't
done
that
move.
A
Yeah,
it's,
I
think
the
big
difference
is
that,
like
docker,
ee
was
something
that
you
can
buy
paid
support
for,
and
that
is
very
valuable
to
a
lot
of
customers.
Whereas
container
d
is
a
community
project
and
like
what
are
they
said,
if
you're
paying
for
support,
you
should
be
able
to
expect
that
you
know
cve's
are
going
to
get
patched
in
a
timely
manner
and
the
bits
are
released.
A
A
Yeah
and
thanks
brendan
for
kind
of
providing
all
this
information
and.
B
A
Yeah,
okay,
the
operational
readiness
I'm
guessing
yeah.
This
is
jade
that
you
added
this
to.
F
Yeah,
I
think
we're
mostly
set
on
it.
I
just
wanted
to
see
if
there's
any
other
major
like
concerns,
I
went
through
the
issues
that
were
left
and
it
looked
like
they
had
all
been
resolved
like
they
had
been
addressed.
I
think
a
meme
addressed
all
of
them
last
week
or
the
week
before
and
whatever
ones
were
remaining.
I
think
I
addressed.
A
Yeah
I
saw
there
were
some
pings,
I
can
comment,
and
then
you
can
open
it
up
for
other
comments
too.
I
think
that
the
only
feedback
like
the
only
concerns
I
have
with
this
are
more
about
the
nature
of
the
cup
and
less
about
the
contents
of
what's
in
the
docs,
and
by
that
I
mean
like
I
was
thinking
about
this
recently
and
it's
like
how
do
we
fill
out
something
like
the
production
readiness
review
for
for
this
type
of
kept?
A
A
Around
so
that's
not
going
to
you
know,
stop
me
from
saying
you
know
this
looks
good,
let's
we're
just
at
least
in
you
know
the
provisional
state,
but
that
is
something
that
I
have
been
thinking
about
and
we
did
kind
of
have
a
similar
situation
with
you
know
say
with
saying
we're
supporting
container
d
for
windows
too,
where
it's
like
there's
not
really,
but
most
of
the
work
is
done
outside
of
the
main
kubernetes
kubernetes
or
you
know
six
repositories.
A
F
Yeah,
I
don't
have
a,
I
don't,
have
a
strong
opinion
there
I
mean,
if
there's
a
if
there's
a
more
appropriate
place
to
put
this,
I
I
kind
of
could
be.
You
know
that
that
doesn't
matter.
I
can't
I
just
don't
know
of
another
place
like
what
else
serves
as
an
actual
standard
for
defining
behavior
right.
It's
just
it's
tricky,
but
yeah.
F
H
A
A
That's
the
only
like
thing.
That's
not
quite
sitting
right
with
me
about
this.
Okay
cool
yeah.
I
don't
know
if
you
have
any
if
anybody
else
james
or
cardio,
do
you
have
any
comments
about
that
or
I
mean,
should
we
just
keep
going
ahead
and
even
if
like
and
just
maybe
push
this
problem
to
the
enhancements
team
and
say
or
or
like
the
community
and
say
we're
not
really
sure
where
this
fits
in.
F
E
E
F
I'm
100
sure
of
the
contents,
but
then
again
it's
I
I'm
selfish
here
so
yeah,
but
I
agree
it's
something
to
merge
and
iterate
on.
If
we
want
to
merge
it.
If
not,
then
I
think
it's
a
really
complicated
question
and
I
don't
think
we're
ever
going
to
get
an
answer
to
the
question
you
have
of
like.
What's
the
right
place
for
this
that'll
just
bounce
around
so
but
I
understand
yeah,
we
did
this
for
network
policies,
so
we
did
the
same
thing.
F
There
was
no
clear
thing
in
place,
so
we
did
what
we
did
there
was
we
did
it
as
a
cap.
We
rewrote
all
the
tests
that
took
about
a
year
after
redoing,
rewriting
all
the
tests
we
published
a
cncf
blog
post
about
all
of
the
providers
in
the
network
policy
that
support
network
policies
that
do
and
don't
support
the
tests
that
we
defined
as
the
standard
in
youtube
test.
So
we
kind
of
did
a
three
prong
approach
there,
which
is
that
we
wrote
the
cap
we
merged
the
cap.
F
We
implemented
all
the
codes
so
that
so
that
the
standard
we
put
in
place
was
transparent
across
providers.
Then
we
ran
them,
we
ran
all
those
tests
and
then,
after
running
all
those
tests,
we
filed
issues
with
the
different
providers
and
we
said:
look
you
either
fix
this
or
you
don't
and
we're
going
to
publish
a
post
about
it
and
we
with
tijera
and
the
entry
of
folks.
You
know
we
all
published
a
blog
post
about
it.
F
So
I
think
that's
the
trajectory
that
I
have
in
my
head
here
is
that,
like
we
create
the,
we
create
this
specification
and
then
we
run
it
on
eks.
We
run
it
on
azure.
We
run
it
on
tanzu.
We
run
it
on
openshift.
If
the
red
hat
folks
want
to
do
that,
and
then
we
say
like
like
this
is
what
these
providers
support
and
probably
azure
wins
right,
but,
like
hopefully
you
know
selfishly
at
vmware.
F
A
Yeah-
and
I
think
I'm
I'm
fine
with
that
approach,
I
think
we
merge
it
that
yeah,
the
the
hesitation
is
really
like
the
around
like
the
pr,
but
we
can
just
leave
all
of
that
blank
for
now
and
merge
it
and
then
add
that
later,
because
I
don't
think
that's
really
applicable
to
this,
because
that's
really,
how
do
customers
detect
that
the
features
on
how
do
customers
detect
there's
an
issue
with
the
feature,
but
you
can
just
omit
that.
F
Yeah
yeah,
it's
tricky
cool
and
if
we
have
to,
if
we
do
decide
to
merge
it
and
we
have
to
revert
it
later
or
remove
it
because
it
something
has
changed
about
caps
and
we're
not
allowed
to
put
things
that
are
standards
or
test
oriented
as
keps
tonight,
yeah
yeah.
I
think
we
can
do
that.
Does
anybody
else
have
any
opinions,
so
that's
me
claudio
mark.
Does
anybody
else
have
any
strong
feelings
about
this.
C
F
A
Okay,
I
think
we're
a
little
bit
over
time,
so
I'm
going
to
end
the
meeting
now.
Thank
you,
everybody
for
attending
and
contributing
and
hope
to
see
you
next
week.