►
From YouTube: Kubernetes SIG Windows 20210323
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
All
right,
hello,
everybody
and
thank
you
for
joining
us
on
the
march
23rd
2021
instance
of
the
kubernetes
sig
windows
community
meeting.
As
always,
these
meetings
are
recorded
and
uploaded
to
youtube
so
be
sure
to
adhere
to
all
of
the
cncf
code
of
conduct
and
standards,
all
right,
let's
kind
of
jump
in,
let's
see.
First
of
all,
I
guess
we
can
check
and
see
if
there's
anybody
kind
of
new
on
the
call
who
wants
to
reach
out
and
say,
hi
and.
A
B
B
A
All
right
there's
a
couple
of
announcements
here:
the
first
ones
being
121
deadlines.
Tomorrow
is
the
test
freeze
for
the
121
release.
So
if
there's
any
prs
that
are
test
related
that
need
an
extra
set
of
eyes
or
are
related
to
windows
and
need
the
121
milestone,
please
reach
out
to
any
of
the
the
leads
on
slack
and
we
can
help
get
those
merged
in.
So
we
have
a
clean
test
signal
going
into
the
release.
A
The
next
deadline
is
march.
31St
is
the
doc
deadline
if
there's
any
doc
updates
that
are
specific
to
the
121
release,
there's
those
are
good
to
have
staged
as
soon
as
soon
as
possible,
because
the
doc
reviewers
like
to
review
those
too.
So
if,
if
anybody's
planning
on
making
any
kind
of
additional
doc
updates,
please
open
those
as
soon
as
you
can
just
to
make
sure
there's
adequate
time
to
get
those
merged.
A
I've
also
dropped
a
link
to
the
kind
of
early
schedule
for
any
other
deadlines,
including
when
all
the
different
rc
builds
are
going
to
drop.
If
there's
any
additional
testing,
that's
needed,
I
guess
one
other
announcement
is
that
we're
in
the
middle
of
planning
and
recording
the
kubecon
eu
talk.
A
A
All
right.
Next
thing
is
jay
and
james
and
myself
were
talking,
and
we
were
wanting
to
do
an
interest
check
to
see
if
there
would
be
any
interest
in
setting
up
a
a
community
meeting,
a
recurring
community
meeting,
that's
in
a
more
asia-friendly
time
zone.
I
think
we
both
know
some
community
members
who
aren't
able
to
attend
the
meeting
due
to
just
the
times
so
we'd
like
to
make
it
more
accessible
to
people.
A
I
think
we
will
probably
reach
out
on
slack
to
get
that
interest
check
going
and
see
when
there
is
a
when
that
when's
a
good
time
for
everybody
to
meet.
So,
if
you
know
anybody
who
kind
of
falls
into
that
bucket
ask
them
to
join
the
sick
windows.
Slack
channel
hopefully
later
this
afternoon
and
and
drop
a
comment
there.
Do
you
have
anything
you
wanted
to
add
to
that
jay.
C
I'd
also
be
interested
in
just
suggestions
about
topics
for
that
meeting
as
well.
I
have
you
know
some,
some
friends
in
asia
that
are
working
hard
on
the
ldap
gmsa
stuff,
and
so
maybe,
if
there
are
any
topics
that
some
folks
know,
some
people
in
the
asia
area
are
working
on
that
are
interesting.
I'd
be
interested
in
seeing
that,
maybe
that
we
could
have
a
little
bit
of
theme
for
it.
A
Yeah-
and
I
also
I've-
I've
noticed
just
over
kind
of
the
course
of
the
past.
Several
months
or
years,
there's
been
a
couple
of
folks
that
are
in
the
asian
time
zones
that
just
kind
of
submit
prs
or
issues
and
aren't
as
active
in
kind
of
slack
or
the
community
meetings
or
other
places.
So
if
anybody
has
any
suggestions
or
for
for
reaching
out
to
them
and
helping
them
get
more
involved
to
kind
of
in
the
community,
please
reach
out
or
and
we'd
appreciate
that.
A
All
right
next
perry
and
jay:
do
you
want
to
drive
this?
The
next
topic.
C
Yeah,
I
don't
know
if
perry
is
here,
but
I
can
drive
it.
We,
this
we've
gone
back
and
forth
on
this
15
character,
node
name
thing
in
the
upstream
community
and
in
the
upstream,
like
cappy
sort
of
side
of
the
community-
and
I
I
just
was
wondering
like
where
do
people
kind
of
stand
on
all
this,
because
it's
it's
going
to
affect
active
directory
and
it's
kind
of
probably
going
to
affect
cluster
api
provider
vsphere,
and
I
think
right
now
in
azure,
what
they
do
is
they
trim
the
characters
down.
C
They
trim,
they
trim
down
some
portion
of
the
character
and
then
add
the
auto-generated
stuff
at
the
end,
but
it
seems
like
we
just
haven't
like
collectively
come
up
with
the
actual
like.
What's
the
right
thing
to
do
here
solution,
so
I
think
james
might
probably
know.
C
Let's
see,
there's
another
link
here,
that
I
can
open
up
and
take
a
look
at
oh
yeah.
This
net
bios
on
host
names
is,
is
the
core
issue.
2217
yeah.
D
Yeah
and
we've
gone
back
and
forth
several
times
on
this,
and
I
think
I
think
the
latest
was
that
it's
needed
for
some
systems
to
tie
it
back
to
that
15
characters.
Is
there
any
chance
of
this
net
bios
restriction
going
away
yeah?
So
it's
all
in
this
issue.
As
far
as
to
where
that
goes,
I
don't
have
enough
experience
with
like
netbios
and
like
how
gmsa
works
with
that
to
be
able
to
make
a
conclusion.
There's
a.
I
think.
D
The
final
thing
at
the
bottom
was
that
if
you
have
systems
that
do
require
netbios,
then
you're
going
to
have
to
keep
it
to
15
characters,
but
there's
no
limitation
within
cluster
api,
as
far
as
like
the
names
being
different
in
different
sizes,
and
so
that
in
win
in
on
the
azure
side,
we're
not
even
able
to
create
vm
names
that
are
longer
than
15
15
characters,
and
that's
why
we
are
trimming
it
there
and
I
think
that's
due
to
azure
being
supportive
of
the
netbios
component.
E
D
Yeah,
I
think
at
the
bottom
of
the
issue.
There
jocelyn
did
a
pretty
good
job
of
summarizing
all
of
the
the
different
things
that
have
been
discussed
and
if
you
plan
on
supporting
netbios,
which
is
required
for
some
ad
some
older
ad
solutions,
then
you
are
going
to
have
to
limit
it
to
15
characters.
Now
the
question
comes,
I
think
it's
been
brought
up
once
or
twice
here
is:
do
we
do
something
inside
cluster
api
to
generate
those
names
so
that
they're,
unique
and
similar
across
all
of
the
different
providers.
A
A
So
I
yeah
we
can
maybe
get
joslyn
to
clarify
that
a
little
bit
too,
but
I
think
that
one
of
the
kind
of
use
cases
is
to
bring
some
infrastructure
into
a
cloud
native
ecosystem
like
kubernetes,
for
for
kind
of
large
windows
deployments
and
leave
some
pieces
of
infrastructure
that
aren't
able
to
kind
of
as
easily
migrate
on
prem
so
and
and
use
active
directory.
Your
group
manager
service
to
tie
those
together.
A
Theoretically,
this
says
that
the
net
bios
requirements
were
removed
in
2000,
but
I
think
that
that's
just
the
requirements,
I
think
that
there's
still
ways
to
configure
everything
to
use
netbios.
So,
if,
like
I
think,
a
lot
of
this
really
depends
on
what
exactly
each
individual
kind
of
cluster
workload
is
how
it's
configured
nate
here.
Did
you
have
a
comment
on
this
portion
of
it.
F
Yeah,
I
I
just
thought
late.
Jay
pointed
me
out.
We
were
chatting
about
this.
I
I
think
also
regardless
of
like
legacy
stuff.
The
sam
account
name
in
active
directory
is
restricted
to
20
characters.
That's
not
that
isn't
deprecated
anyway,
that
is
just
a
function
of
active
directory
and
it's
in
that
ldap
field.
So
we
do
need
that
cluster
api
itself
is
not
actually
responsible
for
any
of
the
host
names.
It
is
down
to
the
providers.
F
So
today
we
have
no
way
of
allowing
a
ec2
instance
with
a
node
name
that
isn't
instance
id
to
be
a
member
of
a
kubernetes
cluster.
There
is
a
there
is
going
to
be
a
flag
to
allow
that
it'd
be
like
a
opt-in
for
the
aws
cloud
provider,
with
a
big
warning
on
it
like
there
are
some
security
implications.
If
you
do
it,
but
we
need
to
actually
go
through
the
cloud
providers
to
check
those
bits.
F
The
bit
around
node
registration
and
then
it
will
be
down
to
the
individual
providers
to
figure
out
how
they
provide
names.
So
this
appears
in
respear,
because
we
we
say
the
id
identity
is
anything
so
we
we
basically
just
take
the
machine
name,
but
for
the
other
cloud
providers-
that's
not
the
case,
so
we
actually
need
to
investigate
both
the
cluster
api
providers
and
also
the
standard
kubernetes
cloud
provider.
Integration
make
sure
it
all
works.
A
G
G
Understood
it
basically
an
old
name,
so
the
object
itself
has
17
characters
that
are
unique.
I'm
sorry
that
randomized!
G
So
we
get
the
work
called
cluster
2
and
d,
something
something
we
have
a
couple
of
randomized
strings
after
that
separated
with
a
dash.
What
component
adds
the
17
characters?
Do
you
know.
D
G
D
Don't
yeah
and
in
cluster
api
the
machine
gets
auto
like
I
think
I
believe
the
characters
are
auto
generated
by
cluster
api
and
then
typically
like
in
a
linux
workload.
We
just
take
that
name
and
that
becomes
the
name
of
the
machine
that
the
infrastructure
machine
that's
created
in
windows
in
azure.
D
You
can
see
that's
linked
in
that
there
is
we
just
kind
of
chop
off
the
everything,
except
for
the
last
five
characters
and
then
just
append
that
so
that
we
get
some
randomness
to
it,
but
we're
just
pulling
that
that
random
bit
is
coming
from
cluster
api's
name
generation.
G
Has
to
be
some
sort
of
a
knob
in
cluster
api
to
allow
you
to
specify
the
length
of
this
these
extra
characters
at
the
end.
D
Yeah-
and
I
I
don't
know
that
those
are
that,
like
that's
exposed
currently,
but
that
could
be
part
of
like
if
we
wanted
to
make
this
generic
across
all
providers.
D
We
could
maybe
modify
that
in
cluster
api
and
then
all
the
providers
could
just
take
that
name
and
just
apply
it
to
to
the
machines.
And
that
way
we
would
have
you
know
from
a
windows
perspective.
Have
that
short
to
15..
I'm
not
sure.
If
that's
something
neater,
if
you,
you
could
say
like
that's
something
we
could
do.
I
think
it's
been
discussed
in
that
issue
that
we
linked
there
as
a
possibility.
F
F
F
Yeah,
so
it's
done
at
different
layers
depending
on
which
infrastructure
platform
you
want,
so
I'm
not
sure
how
generic
we
can
make
it
at
the
moment.
C
A
C
F
You
know
I
don't
feel
like
it's
a
huge
problem.
Is
it
I
don't
know
it
it's
it's
a
problem
because
there's
no
way
of
setting
a
host
name
for
aws,
like
from
the
cluster
api
level.
It
has
to
go
into
the
bootstrap
user
data,
so
we
don't
have
a
mechanism
in
cluster
api
today
to
feed
the
host
name
in,
so
that
that's
a
change
to
how
we
generate
that
cloud.
F
A
Okay,
that
might
be
a
good
place
to
add,
or
it
might
be
good
to
add
those
here
and
say
that
this
would
kind
of
change
how
cluster
api
sets
the
name
names
or
that
would
be
required
for
some
of
the
cluster
api
providers.
C
A
Are
there
any
kind
of
next
steps
here
for
this?
It
seems
like
we're
kind
of
circling
around
some
possible
things.
We
can
do
to
make
it
better,
but
I'm
not
really
clear
what
the
next
steps
are.
A
A
lot
of
these,
like,
like
probably
a
couple
ones,
I
know
for
a
while
a
lot
of
s
protocols
relied
on
the
netbios
name.
So
if
there's
some
kind
of
smb
shares-
or
I
could
say,
storage
servers
set
up
on
with
resources
that
are
managed
outside
of
kubernetes
and
then
folks
are
using
like
active
directory
gmsa
to
access
those
resources
from
a
kubernetes
cluster
that
could
definitely
be
impacted
might
be
good
to
call
out
some
more
use
cases
in
here
too.
G
Yeah,
because
from
either
side,
if
not
none
of
the
services
in
a
question
depend
on
that
buyers,
they
will
just
get
a
warning
about
the
15
characters
limit
and
that's
all.
A
Yeah,
my
understanding
is:
there's
nothing
in
a
pure
kubernetes
cluster
with
windows
deployments.
That
would
be
restricted
from
this,
but
there
may
be
out
like
outside
resources
that
resources
outside
the
cluster
that
folks
are
trying
to
access.
G
D
A
A
A
All
right,
let's,
let's,
let's
come
back
to
this
in
one
of
the
upcoming
meetings-
I
think
that
those
next
steps
proposed
are
are
great.
A
We
can
try
and
find
an
owner
for
those
in
slack
or
just
by
commenting
here
and
on
the
issue
all
right
next
topic:
let's
have
a
couple
minutes
are
last
week
we
kind
of
walked
through
some
of
the
various
sub-projects
that
were
under
the
sig
windows
purview
so,
and
we
are
looking
for
help
and
contributions
for
all
these,
so
we've
gone
ahead
and
added
or
called
out
kind
of
some
of
the
areas
that
we
are
specifically
looking
for
help
with
here.
A
If
there's
anybody
who
is
interested
in
any
of
this
either
just
I
guess
a
lot
of
these
can
just
start
contributing
open
prs.
If
not,
if
you
always
feel
free
to
reach
out
to
us
kind
of
anywhere
jay,
do
you
have
anything
you
wanted
to
kind
of
add
to
the
center
you've
been
a
big
help
in
getting
this
effort
started.
C
Yeah,
I
mean,
I
guess
it's
it's
kind
of
a.
I
mean
it's
kind
of
an
interesting
thing
that
we're
trying
to
do
because,
at
the
same
time
that
we're
trying
to
grow
the
contributors
to
these
like
initiatives,
we're
also
open
to
growing
the
community
in
another
direction.
This
just
kind
of
felt
like
a
good
way
to
get
people
to
be
like
okay.
I
can
own
something
concrete,
because
there's
so
many
things
flying
around
in
sig
windows
and
and
in
kubernetes
in
general
that
for
somebody
coming
in
new,
it
might
be
a
little
overwhelming.
C
So
I
guess
just
the
two
points
to
make
here
is
experience
is
not
important
here
we
can
help
you
with
like
the
different
technical
details
of
things,
and
we
know
who
can
we
can
help?
We
can
get
help
for
anything.
C
So
it's
just
a
matter
of
if
you're
interested
in
something
and
if
you're
not
interested
in
any
of
these,
like
if
there's
another
area
that
we're
missing.
Let
us
know,
because
the
goal
is
really
to
create
some
new
owners
and
and
some
new
technical
expertise
in
the
group,
as
opposed
to
you,
know,
solving
one
of
these
specific
problems.
But
this
is
a
good
starting
place.
So.
H
Thank
you.
What
would
be
your
goal
for
the
next
three
months
on
this
like
if
you
could
paint
the
perfect
picture
where
you'd
like
to
be
with
the
different
subgroups
for
me?
Where
would
you
like
to
be.
C
C
C
A
Yeah,
I
think
more
contributions
from
more
unique
individuals
is
probably
the
most
generic
statement
I
can
make
on
on
that
we're
just
looking
for
people
who
are
interested
in
getting
kind
of
involved
in
getting
you
know
just
who
want
to
start
contributing,
who
are
maybe
a
little
bit
hesitant
to
to
do
so,
because
it
is
kind
of
a
just
the
nature
of
our
sig.
How
there's
a
lot
of
cross
second
involvement
and
that
can
be
off-putting,
so
we're
trying
to
carve
out
spaces
where
there
could
be
more
like
focused
contributions.
H
Right,
I
understand
that
so
I
think
that
with
when
you
have
an
actual
result,
because
I'm
hearing
a
lot
of
sigs
wanting
to
increase
their
con
contributors,
they
they
need
to
right.
So
if
you
were
able
to
say
like
in
three
months
time,
we
had
x
number
of
people
that
were
new
contribute,
something
or
10
new
contributions
from
different
people
that
hadn't
done
so
before
any
sort
of
reasonable
metric
that
you
could
shoot
toward.
That
might
also.
H
C
I
have
a
metric
that
I
would
love
to
see,
which
is
you
know:
10
commits
from
one
person
in
the
net
from
one
person
who's
never
committed
to
sig
windows
before
that
would
be.
I
would
throw
a
big
party
for
everybody
if
we
had
that
we'd
have
a
big
windows
party
and
you
and
you'd
be
invited
laurie,
and
so
would
everybody
else
on
this
call.
H
So
that's
a
good
goal,
any
others
that
we
could
think
of.
Like
maybe
x,
number
of
contributors.
A
C
Like
I
mean
that's
another
metric
that
has
meaning
is
it
the
exact
metric
that
is
meaningful
to
me?
Not
so
much
because
it's
like
it's,
you
know
and
updating
a
readme
file
or
updating.
Some
like
a
comment
is
like
usually
the
first
commit
you
make
to
a
project,
so
the
commits
don't
really
get
good
until
you
get
that
flywheel
thing
going
right,
but
I
mean
yeah.
If
we
got
three
new
contributors
in
the
next
three
months
I
mean
I
don't
think
anybody
would
be.
C
You
know
sad
about
that,
like
you
know,
it'd
be
weird
to
say,
like
we
want
five
new
contributors,
because
we
all
just
know
the
nature
of
this-
that
it's
very
hard
to
even
get
one
new,
really
active
contributor.
So,
like
you
know,
but
yeah,
two
or
three
I
feel
like
right.
If
we
could,
if
we
wanted
to
put
that
on
the
books
as
a
goal,
I
don't
think
that
would
be
bad
but
yeah.
I
agree.
A
I
think
we're
running
out
of
time,
we'll
follow
up
on
this
with
laurie
and
everybody
lumiere.
I
see
you
had
one
more
agenda
item.
G
I
saw
another
cubic
mpr
where
we
had
a
discussion
about
windows
control
plane
in
parallel
these
couple
of
pr's
also,
we
had
a
bit
of
a
discussion
at
vmware
or
whether
we
have
customers,
one
in
the
windows
control
plane.
Yesterday
I
spoke
with
cecile
and
craig
from
microsoft
on
private.
I
I
wanted
to
get
the
whole
to
understand
the
whole
picture
like
what
what
does
the
sick
think
about
eventually
supporting?
I
know
that
this
is
not
a
short-term
goal.
Definitely.
A
Yeah
I've
thought
about
this
before
too.
I
can
share
my
comments
and
then
we
can
open
it
up.
I
think
that
this
is
like.
I
think,
that
we're
seeing
customers
and
end
users
who
would
like
to
see
this
there's
a
lot
of
people
who
are
kind
of
all
windows
based
and
are
finding
that
managing
the
linux
components
in
their
unmanaged
clusters.
A
Are
it's
a
bit
it's
just
a
different
skill
set
than
they
have
and
they
would
like
to
be
able
to
manage
fully
windows
cluster,
so
that
is
kind
of
on
like
on
the
radar,
but
I
think
making
that
happen
is
going
to
be
a
very
long
process,
because
even
if
we
get
all
of
the
the
standard
components
that
are
running
on
the
control
plane,
you
know
cube
api
server
at
cd
all
of
those
running
on
windows.
A
I
think
that
they're
still
going
to
be
kind
of
just
an
endless
number
of
other
components
and
add-ons
that
are
either
assuming
or
require
linux
functionality
that
people
may
want
to
use
like
even
helm,
I'm
not
sure
if
helm
run
can
run
on
on
windows
nodes.
So
it's
like
it's
something
that
I
think
would
be
nice
to
have,
I'm
not
sure,
I'm
even
ready
to
say
this
is
a.
You
know,
a
goal
that
we
want
to
even
set
and
kind
of
put
a
stake
in
the
ground
for
like
a
plan
for
this.
G
Yeah,
it
basically
will
trigger
the
whole
part
of
a
number
of
different
components
to
windows
as
well,
but
it
will
solve
the
use
case
for
users
that
don't
want
to
context
switch
between
the
operating
systems
in
a
isolated
windows
workload
cluster
that,
for
instance,
doesn't
care
about
prometheus
or
something
else.
G
A
I
Running
on
windows,
I'm
wondering
if
we
can
run
the
quantities,
control
salinas
containers
on
windows
using
hyperview
isolation.
I'm
wondering.
A
That's
something
else
that
I
was
wondering
I
one
of
the
things
that
I'm
I'm
wondering
about.
That
is
that
I
know
a
lot
of
the
like
some
at
least
some
of
the
components
I'm
guessing
the
core
dns
needs.
A
lot
of
like
does
a
lot
of
interaction
with
the
the
host
the
hosts
too,
and
I
wonder
if
linux
containers
on
windows
would
be
able
to
support
kind
of
any
of
the
the
privileged
or
kind
of
out
of
container
functionality
too.
But
that
is
a
definitely
an
interesting
direction
to
consider
and
maybe
pursue.
G
A
I
G
A
I
guess
if
there's,
if
there's
somebody
yourself
included
who's
interested
in,
you
know
trying
this
out
and
reporting
back.
That
would
be
a
great
way
to
kind
of
just
see
where
we
are
and
if,
if
we
are
like,
if
most
of
the
components
are
kind
of
just
running
on
windows,.
G
Yeah,
actually,
I
actually
wanted
to
try
this
last
year,
but
I
ran
out
of
time.
Maybe
at
some
point
I
can
try
it
with
a
cube
ideal
setup
to
see
if
what
changes
are
required.
Yeah.
I
can
give
this
a
shot
eventually.
A
J
Sorry
mark
I
was
late
to
the
meeting.
I
could
next
time
around
talk
about
the
short
names
and
the
15
characters
in
long.
I
was
part
of
the
discussion,
but
I
was
late
to
this
meeting
today,
so
I
I
didn't
have
the
opportunity
to
jump
on
it.
I
can
talk
about
it
a
little
bit
next
week.
A
Okay
and
nadir
joined,
and
he
had
some
comments
about
getting
the
support
and
cluster
api
and
he
was
going
to
update
this
issue
2217
with
some
more
details
that
that
he
had
that
were
kind
of
relevant
here.
So
I
think
you're
already
on
the
issue,
so
you
should.