►
From YouTube: Kubernetes SIG Windows 20190827
Description
Kubernetes SIG Windows 20190827
A
Hello,
everybody
and
welcome
to
another
sick
windows.
Meetup
I
think
you
all
for
attending
it's
the
27th
of
August
and,
as
always,
please
adhere
to
the
kubernetes
code
of
conduct.
We
have
a
few
things
on
our
agenda,
but
forget
started.
We
have
we're
presenting
at
the
community
meeting
next
week,
so
we're
just
chatting
a
second
ago
about
you
know
what
kind
of
things
we're
gonna
talk
about.
So
at
the
very
high
level
we'll
give
folks
an
update
on
GMS
a
went
to
beta.
A
They
run
as
user
name
work
and
the
cube
ATM
work
last
time
we
did
at
them
or
cube
ATM.
So
don't
think
money
to
demo
this
time,
but
we'll
do
a
demo
of
TMS
a,
and
this
cannot
drive
demo
then,
in
addition
to
that,
will
basically
provide
an
update
to
everybody
with
the
two
new
technical
introduced,
Windows
community.
Both
of
them
have
obviously
acted
on
the
role
for
quite
a
while
now,
so
it's
just
or
less
recognizing
them
officially
that
panem
deep.
A
So,
thank
you
guys
for
a
lot
of
the
work
that
you
have
been
doing
in
the
community
and
we
really
appreciate
it
and
wouldn't
be
here
without
you
guys
so
we'll
be
recognizing
you
as
well
in
the
bigger
community
as
well
and
Patrick,
Korbin
or
deep
anything
else
that
you
guys
might
want
to
highlight
for
that
community
update
I
think
we
only
have
about
seven
minutes
or
so
and
I'm,
assuming
that
this
should
take
about
half
of
that
for
a
quick
channel.
If
not
more.
A
A
A
A
B
A
A
A
A
C
B
B
B
E
B
F
They
basically
said
it
was
mostly
okay,
so
that
changed
a
lot
since
you
look
at
it,
but
I
still
need
them
to
you.
We
proved
it
so
I
think
that
mom
this
morning
in
slack
and
I
hope
it
gets
merged.
Today,
okay
and
there's
one
other
thing:
it's
gonna
be
an
ad
for
GM
SA
and
that's
the
end-to-end
tests,
the
PR,
for
that
should
be
up
to
date,
so
I'm
gonna
be
getting
windows.
So
it's
like
one.
That's
up!
You
know:
okay,.
B
B
G
B
B
B
B
B
B
C
Hey
yeah
I
can
go
ahead.
Let's
see,
there's
a
clock
tolerate
the
bell
going
really
high,
but
hope
you
guys
can
hear
me
still.
Basically
the
main
issue
that
I
ran
into
and
that
Peter
horny
AK
also
pointed
out
during
his
testing
with
the
plugin
watcher.
Well
that
for
some
reason
the
CSI
no
driver
registrar
is
creating
the
Windows
domain
socket
or
the
UNIX
domain
socket
in
Windows
in
a
way
that's
causing
that
FS
Cydia
that
we
depend
on
do
not
return
the
reparse
tag
that
we
are
expecting.
C
So
that's,
basically
what
I
am
looking
into.
It
seems
when
we
do.
It
were
usually
like
when
normal
program,
the
reparse
tag
is
correctly
associated,
which
is
what
the
P
R
was
based
on.
But
when
it's
being
passed
to
the
GRP,
C,
API,
City
and
registration,
it
seems
you
know,
something's
going
on
in
the
with
the
G
RPC
library,
it's
possible
socket
to
lose
its
rebar
stack.
So
that's
the
bit
I
am
still
investigating.
C
You
looked
at
1903
support
for
this,
where
there
appear
to
be
a
bug
which
seems
kind
of
similar
in
nature
and
other
rebar
stacks
are
not
getting
returned
by
the
FSC
GL
on
a
yes
and
if
he
just
searched
the
logic
there
to
just
try
to
do
a
socket
connection
on
the
file.
Instead
of
looking
at
the
recourse
points,
then
that'll
work
in
1903
as
well
seamlessly,
but
I'm
still
kind
of
I
still
want
to
figure
out
like
why?
C
H
H
C
B
A
B
C
So,
just
for
tracking
purposes,
I
think
I
pointed
out
this
in
the
group
chat.
The
PR
from
Jeremy
I
know,
like
the
main
action
item,
as
Jeremy
just
pointed
out,
is
on
him
to
address
some
of
the
comments
from
Adelina,
but
we
need
bad
to
enable
the
e3
for
Windows
and
that's
in
the
Windows
testing
repo
yeah.
B
B
A
A
C
cluster
lifecycle
and
essentially
what
we're
gonna
do
is
now
that
we're
gonna
get
a
baby
em
out
there
for
the
first
release.
I've
already
met
with
them
a
long
time
ago
and
made
sure
that
the
data
model
that
they
had
for
cluster
API
would
be
amenable
to
include
Windows
sometime
down
the
line
and
it
wouldn't
be
completely
locked
down.
A
Just
for
Linux
I
said
the
OS
of
choice
so
as
cluster
API
is
basically
advancing
their
own
V
alpha
right
now,
they're
advancing
their
agenda,
and
now
that
you
have
cube
ADM
we're
gonna
work
with
them.
You
figure
out
how
to
provision
clusters
that
also
include
Windows.
So
it's
going
to
be
a
key
thing
that
we
can
start
driving
and
you
know
from
the
looks
of
art
say
it
looks
like
we're.
A
B
So
there's
there's
a
couple
people
that
are
already
starting
to
look
at
that,
so
one
of
them
is:
if
you
go
look
at
provider
imager
we've
got.
Let
me
just
jump
to
most
recent
messages.
I
know
that
Cecile
has
been
looking
at
this
from
an
azure
standpoint
as
well
as
Jack
Francis,
so
they're,
both
active
over
on
the
provider
Azure
and
then
I
think
either
Dinesh
or
Nick
was
going
to
start
attending
the
provider.
As
your
meetings
as
well
to
look
at
the
windows
side
of
cluster
API,
that's
great.
A
B
Okay,
all
right
sounds
good
yeah,
so
we
should
be
able
to
talk
about
that.
Hopefully,
in
a
little
more
detail
as
we
started
planning
for
4:17
and
then
I'm
hoping
to
have
an
an
updated
schedule
for
container
D
work,
then
I
know
that
they've
been
shifting
some
people
around
and
hiring
to
to
help
get
some
progress
made
on
the
container
D
side
you're
within
that
the
windows
engineering
team,
some
hoping
that
we'll
be
able
to
make
progress,
then
in
that
direction
as
well
I.
B
B
Yeah,
so
that
the
back
on
that
one
was,
I
know
generally
they
they.
The
guideline
is
not
to
add
new
features
in
back
pull
requests
and
since
I
guess
it
was
sig
AWS
at
the
time
hadn't
been
reviewing
and
Windows
features
that
it
was.
It
was
basically
for
version,
14
and
15
that
it
was
considered
a
new
feature,
and
so
that's
partially,
why
it
why
it
was
rejected.
B
A
D
Back
I
was
hoping
we
could
get
that
spear.
It
was
like
basically
the
same
yard
that
we
cab
has
open
for
vSphere.
I
was
actually
hoping
we
could
back
towards
those,
but
it
looks
like
precedent.
It's
now.
No,
it's
like
us,
you
know
straddling
the
line
of
bug
or
feature,
but
I
think
that's
what
that
whole
thread
was
about
well.
I
D
D
C
A
B
B
All
right,
well
I
gotta,
get
going
over
to
the
cig
note
meeting
was
there,
anybody
have
other
topics
or
a
rail
said.