►
From YouTube: Kubernetes SIG Windows 20201215
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
another
sick
windows
meeting
it's
the
15th
of
december-
and
this
is
our
last
meeting
for
the
year.
So
a
quick
reminder
here
that
we
will
not
meet
for
the
next
two
tuesdays,
giving
everybody
some
time
to
decompress
after
1.20
shipped.
I
give
you
a
time
to
spend
time
with
families
and
and
have
a
great
holiday
season
and
then
we'll
start
again
at
the
beginning
of
next
year.
So
we
have
a
packed
agenda
today.
So
let's
get
started.
A
B
Yes,
so
there
was
a
kind
of
a
lot
of
activity
and
excitement
around
the
docker,
shim
deprecation
notice
and
that's
largely
been
focused
on
linux.
But
there's
been
a
couple
folks
that
have
been
wondering
what
this
means
for
windows,
because
docker
or
container
d
support
is
a
lot
newer
for
windows.
B
I've
met
with
a
couple
people
from
gke
before,
and
we
discussed
a
little
bit
at
the
sig
node
meeting
last
week,
and
I
think
everybody
is
kind
of
kind
of
on
the
same
page
here
where
most
of
the
people
involved
or
want
to
see
the
value
and
want
to
continue
moving
forward
with
container
d
support
for
windows
and
deprecating
docker.
B
B
We
have
a
lot
of
different
test
grid
passes
that
are
showing
that
container
d
with
windows
is
stable,
so
we
have
some
confidence
there,
and
so
we
were
kind
of
discussing
what
what
the
different
options
were
for
doing
that,
because
I
believe,
as
it's
scheduled,
the
docker
shim.
B
B
sig
node
was
very
kind
of
sympathetic
to
this,
and
they
understand
that
they
that
this
is
new
for
windows
and
they
don't
want
to
kind
of
put
any.
They
don't
want
to
add
any
more
friction
here.
So
one
of
the
things
that
dawn
chen,
one
of
the
signod
co-chairs
proposed,
was
that,
depending
on
the
timelines
of
when
different,
either
cloud
providers
or
large
deployments
can
validate
get
some
extra
validation
with
windows
and
container
d
deployments.
Signode
is
open
to
leaving
in
the
docker.
B
The
other
option
that
was
discussed
is
there
was
a
little
bit
of
a
kind
of
momentum
around
mirantis,
adopting
this
cri
docker
d
or
cri
docker
shim,
and
I
think
everybody
agreed
that
for
windows-
that's
probably
not
the
best
option,
because
it
would
just
be
there'd,
be
a
lot
of
effort
needed
to
validate
everything
with
windows,
and
we
agreed
that
that
effort
would
be
better
spent
in
just
focusing
on
getting
container
d
tested
and
validated
everywhere.
B
So
there's
no
kind
of
set
timeline
for
any
of
this.
It's
mostly
just
we'll,
wait
and
see
when
different,
when
people
start
adopting
continuity
for
windows
and
then
just
take
it
from
there.
But
I
think
that's.
A
pretty
good
solution
is
to
keep
the
docker
shim
in
the
code
in
the
cubelet
and
only
building
for
windows
if
needed.
A
Yeah,
absolutely
that
makes
sense
mark.
That's,
probably
the
the
best
path
forward
give
enough
time
to
continue
to
bake.
Now
that
is
ga
let
public
cloud
providers
and
others
to
start
putting
it
to
test
paddle
test
it
and
then,
as
as
when
we
feel
comfortable
and
everybody
feels
comfortable.
This
is
stable.
Then
we
can
start
the
deprecation,
so,
whether
that's
three
months
from
now
or
six
months
from
now,
that's
great
that
we
actually
have
the
option.
Now.
C
And
just
a
reminder,
I
think
I
think
we
shouldn't
take
that
as
a
signal
that
we
should.
You
know
just
to
continue
sitting
on
docker.
We
should
continue
moving
towards
container
d
and
trying
it
out
right.
Mark
like
this
is
by
no
means
like
an
encouragement
to
stay
on
docker.
This
is
just
like
you
have
a
little
bit
more
time,
so
don't
panic,
but
you
should
continue
thinking
about
it.
Correct.
A
So
so
so
miles,
one
thing
that
could
help
us
here
is
basically
since,
since
you're
the
primary
pm,
that's
driving
a
lot
of
the
continuity
work
is
maybe
come
up
with
a
plan
with
google
and
aws
and
obviously
azure
on
when
they're
gonna
roll
out
container
d
in
production
windows
for
the
windows
distributions,
and
we
can
actually
start
following
that
path.
Closely.
Yeah.
C
Actually,
this
week
we
already
are
in
discussion
with
aws.
I
need
to
start.
I
can
probably
ask
peter
others
to
to
think
about.
You
know
I'll
start
that
you
know
so.
The
major
destroys
when
I
think
that's
a
good
question.
I'll
definitely
follow
up
on.
B
That
yeah
and
I
know
that
for
azure,
we
plan
on
doing
up
a
public
preview
with
container
d
for
windows,
node
pools
and
I
believe
gke
does
too.
I
don't
think
we
can
say
anything
about
timelines
for
that,
but
hopefully
it'll
be
kind
of
early
in
2021.
B
So
do
you
know
when
azure
will
have
the
public
preview,
I'm
not
entirely
sure,
okay
I'll
stay
soon,
yeah
I'd,
say
hopefully
early
2021?
It's
sounds
good.
A
Cool
all
right,
aravind
update
on
windows,
copy
discussion.
D
B
That
is
some
work
right.
There
shouldn't
be
many
changes
in
the
the
cni
plug-ins
I'll,
defer
I'll
comment
and
I'll
defer
to
david
shot
and
jocelyn
on
here.
The
one
difference
is,
though,
if
you
want
to
use
container
d,
the
cni
comps
they
normally
delegate
out
to
a
win
bridge
or
when
overlay
executables,
depending
on
what
networking
mode
you
have.
B
There
is
another
set
of
binaries
sdn
bridge
and
sdn
overlay
that
are
needed
for
container
d
support
and
there's
in
the
I
believe,
the
microsoft
windows
container,
networking
repositories,
that's
where
those
binaries
get
built
out
of-
and
there
are
example,
cni
config
files
using
those
for
a
couple
different
cni's,
including
flannel,
that
we
can
link
to
mark.
D
Mark
it'll
help
if
you
can
just
throw
a
link.
B
D
In
the
agenda-
and
I
can
take
it
from
there
awesome-
thank
you
so
going
back
to
the
update
on
the
capi
discussion,
so
we
had
a,
I
think,
a
fruitful
discussion
with
the
capi
folks.
There
seems
to
be
a
couple
of
approaches
to
for
enabling
windows
workers
on
on
clusters.
One
is,
of
course,
using
cube,
adm,
and
the
other
is
what
red
hat
is
doing
is
by
following
this
operator
model.
So
the
the
brief
takeaway
the
tldr
version
is
cappy,
wants
us
once
to
be
able
to
support
both
approaches.
D
So
what
I
have
taken
as
an
action
item,
is
to
for
red
hat
the
our
cappy
team
here
to
sort
of
rework
the
the
pr
they
have
opened
and,
if
needed,
even
open
up
a
proposal
just
describing
what
we're
trying
to
do
and
how
it
fits
in
such
that
you
know
both
the
approaches
will
fit
as
long
as
both
teams
are.
You
know
confusing
the
same
cluster
api
and
folks
don't
need
to
carry
downstream
patches
and
things
like
that.
D
A
Thank
you
any
any
any
comments
or
concerns,
or
any
questions
for
arvind.
E
So
I'll
just
throw
in
the
the
windows
support
via
the
cubeatm
component.
We've
got
the
test
passing
so,
if
anybody's
interested
in
seeing
how
that
works,
please
take
a
look
and
comment
on
the
pr.
It's
probably
be
reviewed
this
week
and
emerge
for
capsi
sometime
in
the
next
week
or
two.
So
take
a
look
if
you're
interested.
A
Very
cool,
thank
you
all
right.
The
next
topic
is,
you
know:
we've
had
a
few
folks
that
have
made
significant
contributions
over
time
on
windows,
and
we
have
a
few
folks
that
are
just
starting
up
and,
and
you
know,
we're
trying
to
put
them
full
time
on
windows
and
they're
starting
to
ramp
up
and
make
significant
contributions
as
well
and
mark-
and
I
were
talking
and
were
thinking
of
expanding
the
number
of
tech
leads
on
windows.
A
That's
heading
a
lot
of
our
storage
and
and
csi
efforts,
and
then
we
also
have
ben
moss,
who
hasn't
been
as
involved
as
of
late
and
will
kind
of
gauge
his
willingness
to
continue
to
put
in
more
time
on
on
windows
and
and
the
kubernetes
effort
here.
So
we
wanted
to
propose
two
additional
tech
leads
to
kind
of
help.
A
Us
kind
of
increase
the
number
of
approvers
and
reviewers,
as
well
as
make
sure
that
folks
have
been
significantly
contributing
to
our
efforts,
are
recognized
and
and
have
a
technical
leadership
role
in
our
community
without
further
ado,
we'll
start
with
the
first
one,
and
everybody
here
has
an
opportunity
to
either
tell
us
we're
not
going
to
make
the
decision
right
now,
but
we
can
make
it
in
the
next
couple
of
days.
A
So
the
first
nomination
is
james.
So
james
has
been
shepherding
a
tremendous
amount
of
our
effort
around
sequindo's
tools.
James
has
been
instrumental
in
some
of
our
efforts
around
cluster
api
and
overall,
our
automation,
as
well,
as
you
know,
being
a
great
kubernetes
citizen
and
a
great
ambassador
for
for
for
our
work
here.
So
we
would
like
to
nominate
james
to
be
the
first
additional
tech
lead.
A
Do
you
have
any
use
the
slack
thumbs
up
or
thumbs
down?
If
you
like,
or
unmute
yourself
and
and
voice
your
your
approval
to
this
nomination
yeah,
I
agree.
A
All
right
so
we're
seeing
that
james
is
getting
some
thumbs
up
here.
So
so
that's
great
and
the
second
nomination
is
jay
javas
from
from
vmware,
so
jay
has
been
kind
of
shepherding
a
lot
of
the
folks
from
vmware
that
are
contributing
upstream
on
cluster
api
on
windows,
we're
looking
into
the
csi
support
for
for
vsphere
now
as
well.
As
you
know,
the
image
builder
work
and
kind
of
kind
of
wrangling
up
the
troops
that
are
kind
of
pushing
that
effort
forward.
A
So
can
we
see
the
nominations
here
for
for
jay
thumbs
up
or
otherwise.
B
A
Absolutely
so
we'll
leave
this
open
for
a
couple
of
days.
If
we
don't
hear
any
concerns
from
the
community,
then
we'll
start
moving
forward
and
making
it
official,
probably
with
the
beginning
of
next
year,
but
congratulations,
james
and
jay.
Maybe
the
next
person
needs
to
start
with
a
j
as
well,
so
I'm
just
kidding,
but
congratulations.
This
is
well
deserved.
Thank
you
for
your
for
your
contribution
so
far
and
we
look
forward
to
many
more
contributions
in
the
community.
F
Here
I
appreciate
that
I
just
wanted
to
make
sure
we
give
credit
where
it's
due
a
lot
of
the
stuff
or
almost
all
the
stuff
that
I've
done
has
been
piggybacked
on
top
of
perry's,
hard
work
as
well
so
perry
once
we
get
you
more
involved
in
upstream.
Maybe
you
can
join
us
in
this
in
this
circle
as
well.
E
A
Excellent,
all
right,
so
the
next
topic
is
device
plug-in
api,
the
gpu
support.
You
know
that
that
has
been
the
pr
that
has
been
kind
of
sitting
there
for
a
while,
like
if
you
remember
back
when
we
added
this
item
here
for
the
future,
once
one
21.3
is
open,
we
need
to
merge
these
prs,
the
948
and
the
285..
A
A
I
think
the
only
thing
that
they
might
need
is
to
be.
It
needs
a
remap
right.
The
rebase
is
there
anything
else
I
might
need.
As
far
as
I
know
and
mark
you
correct
me,
if
I'm
wrong
is
that
that
we're
able
to
validate
the
tests,
even
though
we
were
not
sure
if
the
test
covered
comprehensively
across
the
border,
everything
that
we
needed,
but
the
only
reason
why
we
emerged
back
then,
is
because
we
were
not
sure
of
the
risks.
But
now,
if
we
merge
it
early,
we'll
be
able
to
assess
that
risk
earlier.
B
Yeah
yeah
kind
of
near
the
end
of
the
120
development
we've
got
some
tests
that
are
running
on
gpu,
enabled
nodes
in
azure
and
we
were
able
to.
You
know
just
provide
a
lot
more
validation
and
I
think
we've
actually
identified
the
reasons
why
we
were
seeing
like
different
device
ids
and
everything
there
too.
So
I
think
it's
ready
to
merge
and
that's
probably
gonna,
be
one
that
we're
gonna
want
to
push
through
really
soon
early
early
in
121.
Once
everything
opens
up.
A
Yep
so
so
we'll
ask
them
to
rebase,
we'll
ask
them
to
to
make
sure
you
know,
since
the
tests
are
passing,
let's,
let's
after
the
rebus,
let's
make
sure
we
get
another
green
run
of
the
test
and
then
it
can
be
merged
all
right,
maz!
You
want
to
do
the
triage
now.
C
Yes
sounds
good,
so.
A
C
Just
quit
yeah
so
while
amber
and
I
and
mark
and
james,
I
think
a
lot
of
folks
were
going
to
this
networking
meeting
for
privileged
containers.
The
one
thing
I
really
liked
is
in
beer,
like
sig
networking,
bi-weekly
sync,
they
take
like
five
minutes
and
go
over
something.
They
call
issue
triage
where
they
look
at
the
they
filter
down.
C
Just
to
the
to
the
sig
networking
issues
we
could
do
for
windows
and
only
the
new
issues
and
they
don't
actually
go
and
try
to,
like
you
know,
get
to
the
root
cause.
They're
just
saying
hey.
Can
somebody
take
a
look
like
you
spend
five
minutes
for
the
you
know,
first-line
investigation
and
kind
of
you
know
either
assign
to
the
right
folks
or
kind
of
you
know,
specify
the
next
steps
so
just
for
that
they
assign
it
to
someone
to
take
the
first
step
kind
of
thing.
It's
super
nice.
C
It
doesn't
take
more
than
five
minutes,
but
usually
you
have
like
you
know
two
three
issues,
but
it
this
way
you
kind
of
you
stay
up
to
date
with
the
with
the
new
issues
that
are
popping
up
and,
and
that
way
the
whole.
You
know
the
group
kind
of
takes
a
look
at
it
as
well.
We
kind
of
seen
the
pattern
that's
emerging,
I
really
liked
it
because
I
mean
I've.
C
All
I've
seen
I
think,
we're
having
the
sick
windows
backlog
meeting,
which
is
this
thursday
by
the
way,
which
is
helpful,
but
I
I
think
the
we
often
don't
get
a
chance
to
you
know
go
over
all
the
issues,
so
so
they
stay
in
the
backlog
for
a
long
time.
So
I
just
wanted
to
like
float
that
idea.
If
people
are
interested,
I
can
take
that
responsibility
in
keeping
the
query
and
running
that
part
of
the
meeting.
Like
the
last
five
minutes
of
every
meeting.
C
We
just
look
at
the
newer
issues
and
somebody
volunteers,
and
you
know
I
say
like
oh,
this
is
deep.
Do
you
want
to
take
this
one?
It
doesn't
mean
you
have
to
resolve.
It
just
means,
like
you,
read
it
and
then
you
kind
of
think
of
what
whatever
is
the
next
step.
A
I'm
very
supportive
of
that
five
minutes
is
another
huge
investment.
If,
on
a
day
that
we're
super
busy
that
we
can't
get
to
that,
then
maybe
that's
the
day.
We
don't
have
it,
but
we
have
lots
of
days
that
are
meeting
center,
20
minutes
or
25
minutes.
So
putting
that
extra
investment
into
something
that's
gonna
move
the
ball
forward
is,
is
it's
gonna
be
advantageous?
E
Cool
and
also
just
I'd
like
to,
if
folks
want
to
come
on
those
thursday
meetings
too.
I
think
it's
been.
You
know
a
small
group
of
us
but
everybody's
welcome
to
to
join
us
there,
and
we
can
go
a
little
bit
more
in
depth
into
some
of
those
issues.
So
please
please
join
us
if
you
want
to
talk
about
them
in
more
depth,.
A
A
C
A
There's
a
reminder
on
slack,
but
let
me
let
me
go
back
to
sharing
my
screen
here
again,
so
we
have
so.
I
don't
know
how
many
of
you
all
follow
the
the
calendar,
but
there's
a
there's
a
calendar
with
all
the
kubernetes
communities
meetings
so
and
like
the
meeting
is
here.
So
if
you
click
on
that,
you
can
copy
to
your
own
calendar.
A
So
let
me
let
me
post
this
link
to
our
to
our
meeting
invite
and
that
you
can
do
the
same
thing
with
the
regular
seek
windows
meeting
as
well
by
the
way
yeah.
C
A
You
know
what
I
know
where
to
add
it
I'll
add
it
in
our
meeting
notes
all
right.
Let's
see,
matt
that's
image.
C
A
Cool
amber
floor
is
yours,.
G
Cool
I'm
hi
everyone.
We
were
just
chatting
about
like
the
actual
updates
for
privilege
containers
last
week,
but
today
I
kind
of
wanted
to
bring
a
little
bit
of
a
personal
update.
G
I
have
taken
a
role
outside
of
microsoft,
so
I'll
be
transitioning
off
of
working
on
privileged
containers
from
kind
of
the
you
know,
windows,
containers,
team
standpoint,
I've
had
a
great
time
working
on
this
project.
Of
course,
and
great
time
meeting
all
of
you
and
working
on
you
know
privilege
containers
has
been
something.
That's
that's
been
in
the
works
for
quite
a
while,
and
I
know
people
have
been
looking
for
it
for
quite
a
while
and
there's
so
much
excitement
around
it.
It
really
does
pain
me.
B
D
G
G
But
in
the
course
of
this
I
do
also
want
to
introduce
brandon,
who
is
another
pm
on
the
kind
of
windows
containers
group
who
will
be
transitioning
to
kind
of
take
this
over
and
and
take
over
kind
of
working
with
you
guys
and
making
sure
that
we
continue
to
have
support
from
our
team
for
this
particular
project
and
a
huge
presence
here.
G
Of
course,
you
know,
there's
muzz
and
mark,
and
so
many
other
people
who
have
been
in
james
and
and
all
these
folks,
who
have
been
also
incredibly
involved,
but
brandon
will
kind
of
be
stepping
into
things
where
I've
been.
You
know
situated
so
brandon.
If
you
want
to
give
yourself
a
quick
intro.
G
A
That's
okay,
amber.
I
would
like
to
say
you
know,
first
of
all
thank
you
for
all
your
contributions
and
kind
of
helping
our
community
and
and
having
champion
windows
and
especially
around
the
privilege,
containers
effort,
and
it
has
been
a
privilege
to
to
to
have
you
lead
that
effort
you're
going
to
get
that
a
lot.
I
think.
G
Yeah
yeah,
thank
you
guys,
and
it
has
also
been
a
privilege
to
be
involved
but
yeah.
You
know
it
is
an
open
source
community,
who
knows
maybe
I'll
all
be
in
and
around
at
some
point
someday
in
the
future.
If
anything,
just
to
see
all
of
your
guys's
amazing
work
but
yeah
great.
Thank
you.
Everyone,
hopefully
brandon
will
get
a
chance.
H
Oh
totally,
sorry
about
that,
my
computer
died
literally,
as
you
were
introducing
me,
so
I
switched
computers
but
from
what
I
missed.
You
know
thank
you
for
the
introduction
everyone
or
amber.
I'm
super
excited
to
get.
C
Cool
yeah
and
those
of
you
who
don't
know
brandon,
he's
he's
he's
very
technical
and
pretty
deep
into
windows,
kernel
and
windows
containers
so
good
to
have
him
on
board.
Of
course,
like
said
to
see,
amber,
go
but
excited
for
her
future.
One
quick
thing
amber:
do
you
want
to
just
talk
about
the
the
cap
state
for
privileged
containers?
I
know
it
needs
to
get
approved.
G
Yeah
yeah,
so
there
is
some
updates
that
we
might
want
to.
I
want
to
take
a
pass
through
it
to
make
sure
that
all
the
updates
that
we've
kind
of
been
throwing
around
with
networking
and
so
on
and
so
forth
and,
like
these
details,
are
kind
of
included
within
kind
of
my
time
remaining
here
and
also
it's
kind
of
an
intermediate
state
which,
like
I,
have
no
idea
what
occurred
in
my
github
account
once
I
exit.
G
G
I
guess
like
the
kept
ammo
file
or
one
of
those
email
files
that
I
think
was
introduced
somewhat
more
recently
included,
and
so
it's
kind
of
been
a
strange
in
between
state
and
that's
that's
what
I'm
really
looking
to
to
kind
of
update
yeah
in
this
time.
B
A
Excellent,
well,
that's
that's
it
for
today.
That's
it
for
the
year
folks
who
had
a
tremendous
year
in
windows.
You
know
we
took
a
lot
of
technology
to
ga
this
year,
container
d,
our
csi
effort,
you
know
a
lot
of
those
things-
would
have
been
possible
without
all
of
the
effort
and
work
that
every
one
of
you
is
contributing.
A
A
lot
of
you
are
contributing
on
your
company's
dime,
but
also
realize
that
the
lava
contributors
are
working
on
the
spare
time
nights
and
weekends
outside
of
their
regular
job.
So
we
appreciate
that.
Thank
you
for
for
your
effort.
Thank
you
for
your
contributions.
It's
been
a
great
year
and
we
look
forward
to
2021
and
hopefully,
windows
keeps
keeps
this
aggressive
momentum
that
it
has
had
in
the
kubernetes
community,
and
we
see
a
great
adoption
from
our
customers
as
well.
A
Thank
you,
happy
holidays
and
we'll
see
you
all
the
first
tuesday
of
the
new
year
yep
thanks
michael
bye,
yeah
everybody.