►
From YouTube: Kubernetes SIG Windows 20210406
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
welcome
to
the
april
6
2021
iteration
of
the
kubernetes
sig
windows
community
meeting.
As
always,
these
meetings
are
recorded
and
uploaded
to
youtube,
so
please
be
sure
to
adhere
to
all
the
cncf
code
of
conduct
and
standards
all
right,
let's
kind
of
just
get
into
it.
I
do
see
a
couple
of
new
people
on
the
line.
Usually
we
just
kind
of
open
it
up.
A
If
anybody
wants
to
introduce
themselves,
there's
no
kind
of
obligation
to
do
so,
but
if
anybody
just
wants
to
say
hi
or
just
ask,
questions
or
kind
of
introduce
themselves
feel
free
to
either
raise
your
hand
or
just
go
ahead,
and
do
that
jamie.
I
see
you.
A
All
right,
I
guess,
is
there
anything
you
could
also
share.
Is
there
anything
specific
that
you
would
like
to
get
out
of
the
meeting
or.
B
Just
learning
about
the
current
state
of
windows,
support
for
kubernetes.
A
Okay,
cool:
we
have
a
pretty
light
agenda,
so
we
could
probably
go
into
that
some
more
if
there's
anybody
or
like
if
people
are
interested
anybody
else.
C
Hey
I'm
sladen,
even
I'm
new,
so
yeah
I
just
I
just
like
contributed
a
bit
to
the
release
team
for
kubernetes
release
engineering,
and
I
seen
that
I
can
maybe
hop
on
here
and
you
know,
help
you
folks
out
with
some
interesting
stuff,
so
yeah
looking
to
learn
more
about
sync
windows
and
you
know
chop
my
wood
in
the
community,
so
yeah.
A
All
right
great
yeah,
we
could,
I
guess
I
I
see
that
you
added
that
on
there
as
well.
We
could
probably
just
go
into
that
then
after
this,
I
don't
think,
there's
many
announcements.
I
think
that
the
kubecon
north
america
cfp,
the
call
for
proposals
got
pushed
out
a
little
bit.
A
So,
if
anybody's
interested
in
submitting
that
I
had
a
link
down
here
somewhere
yeah
check
that
out,
if
anybody's
interested
in
submitting
there's
some
guides
on
how
to
do
that
and
yeah,
it
looks
like
it
got
extended
to
the
end
of
may,
there's
some
examples
of
kind
of
what
would
make
a
good
proposal.
What
what
wouldn't
and
just
some
tips
here.
A
It's
a
good
way
to
just
kind
of
get
you
get
your
name
out
there
get
used
to
speaking,
and
especially
if
you're
interested
in
promoting
anything
windows
related
will
help
that
out
all
right.
Getting
to
this,
I
think
the
only
thing
we
had
was
slayton
here,
the
sig
release
liaison.
Is
there
anything
else
that
you
wanted
to
to
add
to
this
note
or
just
saying
hi,
and
then
can
maybe
talk
about
this
in
a
little
bit.
C
Yeah
yeah
yeah,
I
I
came
across
it
in
the
I
guess
it
was
laurie's
call
for
like
help
so
yeah.
That's.
I
just
wanted
to
add
that,
and
you
know
maybe
as
much
as
I
can
help,
although
I
would
need
a
lot
of
context
here,
but
yeah
yeah,
I'm
willing
to
learn
yeah,
that's
it.
A
Okay,
yeah,
I
was
going
to
introduce
that
I
think
laurie
said
that
she
could
join,
but
later
so,
for
anybody
who
is
kind
of
unfamiliar
with
that
laurie
has
been
working
on
a
proposal
for
a
new
role
for
the
kubernetes
community.
That's
an
enhancements
liaison
or
a
release
liaison,
and
this
is
a
new
I'll-
have
to
ask
laurie
for
the
link
to
that
or
find
it.
This
is
a
new
role
kind
of
that's
born
out
of
the
changes.
A
The
recent
changes
introduced
in
the
121
release,
where
sigs
now
need
to
kind
of
push
information
about
what
enhancements,
they're,
updating
and
pursuing
to
the
enhancements
team
lori
thought
that
for
a
lot
of
cigs,
it
makes
sense
to
kind
of
define
that
as
a
specific
role
that
could
rotate
release
after
release
just
to
kind
of
help
out.
Some
some
sigs
have
like
dozens
of
enhancements,
and
so
it's
hard
to
keep
track
of.
This
is
something
that
each
sig
is
kind
of
can
opt
into
doing
this.
A
I
think
we
definitely
think
it's
a
good
opportunity
to
help
people
who
are
interested
in
kind
of
pursuing
this.
If
laurie
does
join
later,
we'll
have
her
kind
of
introduce
it
or
a
little
bit
more.
If
not,
we
can
we'll
share
the
links
and
everything,
but
so
yeah.
Thank
you
for
your
interest
in
that.
A
Jay,
did
you
have
anything
you
wanted
to?
I
saw
in
the
sigma
windows
lead
slack.
You
wanted
to
talk
about
the
q
proxy
work.
Did
you
want
to
talk
about
that?
If
not,
we
can
kind
of.
D
Yeah
I
joined
a
little
late.
Sorry,
I
just
completely
lost
track
of
time
was
deep
in
something
else.
So
did
I
yeah
so
sladen
had
actually,
I
believe,
mentioned
something
about
the
coup
proxy
stuff.
Also
I
was
thinking
we
could
go
through,
maybe
that
an
image
builder
today
when
we
start
pairing
on
stuff-
I
I
didn't
put
this
in
the
agenda,
but
I
got
really
confused
this
weekend
when
I
was
playing
with
isos
and
image
builder,
and
then
I
realized
that
I've
always
been
building
images
using
a
special
iso.
D
That
perry
gave
me
and
I
never
really
tried
the
evaluation
versions,
and
then
I
got
really
lost,
and
then
I
and
I
realized
we
don't
have
anything
in
our
docs.
I
don't
think
about
the
difference
between
msdn
versus
evaluation
versions
of
windows
and
whether
there's
different
things.
We
need
to
do
to
support
them
and
we
don't
have
anything
about
that
in
image
builder,
and
so
I
wanted
to
ask
the
image
builder
gods
about
what
I
don't
even
know.
D
What
questions
I
should
be
asking
there,
but
does
anybody
have,
should
we
have
this
in
our
docs
or
should
we
have?
This
is
this,
like
a
series
of
issues
in
image
builder,
that
we
need
to
create,
or.
E
So
so
I
think
this
only
really
affects
people
who
are
running
on-prem
and
wouldn't
affect
any
cloud
provider.
I
guess
the
general
question
is:
when
will
when
we're
talking
upstream
in
image
builder?
Should
we
be
upstream?
E
Should
we
be
putting
it
so
configuring
it
for
eval
and
then
letting
people
do
other,
or
should
we
be
kind
of
so
what
the
way
I've
done
it
till
now
is
I've
kind
of
been
mandating
that
people
bring
their
own
licensed
media
to
be
able
to
use
image
builder,
which
is
what
jay's
talking
about
with
special
iso
just
means
I
actually
have
a
license
and
whether
we
should
be
upstream
in
github
whether
we
should
be
doing
that
as
an
eval
media
and
therefore,
and
then
just
changing
that
downstream.
A
Yeah,
as
perry
mentioned,
I
think
this
is
going
to
be
really
specific
to
the
platform
that
you're
building
the
images
for
like
in
azure.
We
don't
even
base
our
images
off
of
an
iso.
We
base
them
off
of
a
marketplace:
vhd,
that's
already
there
and
like
in
in
azure
that
that
you
can
bring
your
own
license,
but
if
you
don't
bring
your
own
license
the
vm
or
like
you're
just
by
running
the
windows,
server
license
you're
getting
kind
of
a
time
slice
a
time
shared
license.
D
E
It's
just
I've
just
got.
I've
just
got
an
msdn
subscription
that
I
just
downloaded
the
iso
from,
and
then
I
use
that
as
my
as
my
media,
because
it's
licensed
but
yeah.
My
general
question
is:
do
we
want
to
change
image
builder
so
that
people
who
are
building
upstream
are
using
an
eval
media
and
therefore
don't
need
to
provide
their
own
iso?
They
can
just
grab
the
it's
configured
for
eval,
or
should
we
be
presuming
that
they're
going
to
be
having
licensed
media
if
they're
going
to
be
using
image
builder
is.
D
A
E
E
It's
literally
as
simple
as
I've
had
to
write
an
unattended
or
used
an
attend,
dot,
js
or
xml,
and
I've
put
in
the
standard
product
sku
that
comes
out
of
that.
So
the
the
product,
the
the
generic
you
know,
product
key
that
you
can
use
and
it's
whether
I
should
be
doing
that
I
mean
I
would.
I
would
love
for
somebody
to
generate
something
that
stops
me
having
to
write
a
separate
unattend
file
for
every
version
that
I
put
into
image
builder.
But
I
don't
think
at
all
exists
of
that.
E
A
Other
problems,
but
so
like
the
the
this
is
more
just,
I
think
some
context
about
how
the
windows,
licensing
and
and
works
so
there's
so
there's
many
different
skus
of
windows,
server,
there's
like
server
data
center
server
standard
and
both
of
those
come
with
a
gui
and
a
non-gooey
flavor,
and
the
differences
between
those
are
really
just
like.
How
many
I
think
it's
how
many
cores
you
can
run
like
that
edition
of
windows
on,
but
then
for
each
edition
of
windows.
A
There's
also
what's
released
as
an
eval
version
of
that
sku,
and
that,
like
that,
is
just
with
an
evaluation
license
where
I
believe
you
have
30
days
in
order
to
activate
it
in
order
to
kind
of
comply
with
the
licensing
mandates.
The
normal
skus
are
they're
they're
armed,
assuming
that
you
have
a
license,
and
you
have
a
certain
amount
of
time
to
kind
of
register
that
license
with
whatever
your
your
licensing.
Server
is
sometimes
those
are
just
on-prem
licensing
servers,
sometimes
they're,
just
with
kind
of
the
windows
like
updating.
D
A
Pretty
much,
I
believe,
I
believe
so
now
that
there
may
be
some
features
that
are
like
some
backup
restore
features
that
are
different.
I
I'm
not
sure
what
the
latest,
but
in
like
windows,
server
2012
time
frame
there
used
to
be
more
differences.
I
think
they're
converging.
B
B
A
So
I
guess
the
question
like,
like
I
mentioned,
for
the
azure
image
builder
and
I'm
guessing
eks
to
there's
like
the
marketplace,
images
that
get
published
that
we're
using
to
start
off
with
an
ad
kind
of
the
kubernetes
special
magic
there's,
not
an
option
to
even
target
the
email
skus.
So
so
I'm
actually
don't
have
enough
information
to
comment
on
this.
Yet.
A
I
always
kind
of
viewed
this
as
like.
There's
reference
images
and
that's
what
we're
kind
of
checking
in
there's
steps
and
guides
on
how
to
build
tooling
on
how
to
build
reference
images
that
get
published
to
the
image
builder
repository.
But
then
the
recommendations
or
guidance
is
for
anybody
who
actually
wants
to
deploy
this
in
their
own
clusters
to
tweak
it
and
customize
it
and
build
their
own
images.
So
I
wonder
if
we
should
have
the
reference
images
be
based
off
of
that
eval
skus.
But
perry
are
the
images
or
the
questions
you're
asking.
A
E
I'm
just
wondering
how
to
handle
so
I
in
inside
of
image
builder.
I
just
need
to
know
whether
you,
realistically,
whether
we
want
to
specify
a
product
like
the
generic
product,
the
product
code
that
requires
activation
or
whether
I
should
just
remove
that
from
the
unattend
and
just
have
it
as
a
have
it
as
a
blank
product
key
which
then
would
allow
building
what.
D
Is
the
benefit
of
having
the
product
thing
in
there?
Actually,
a
big
shout
out
to
stewie,
he
was
the
one
stuart
preston
was
the
one
that
told
me
to
delete
the
product
queue
and
then
I
did
that
product
again,
I
did
it.
It
worked
and
I
was
so
happy
and
then
I
had
a
really
good
saturday.
After
that
I
was
like
okay.
This
is
it
I
feel,
like
I
accomplished
something.
E
D
E
You
to
download
an
an
iso
from
msdn
from
like
a
licensed
edition
and
then
run
it.
That's
the
only
that's
the
only
benefit
and
it
saves
you
haven't.
E
D
A
A
Yeah,
okay,
so
they
they've
changed
kind
of
the
model
of
what
happens
if
you're,
if
you're
not
doing
this-
and
there
are
like
if
you're
running,
if
you
are
running
with
the
license
version,
there
are
checks
and
windows
to
periodically.
You
know
contact
licensing,
servers
and
make
sure
you're
valid
and
they
will
start.
D
So
so,
the
other
reason
that
we
should
not
support
eval
is
that
when
you
run
the
evaluation
version
it
makes
you
pick
those
four
different
things
that
mark
mentioned:
data
center
standard
data
center
gui
data
center
whatever-
and
that's
not
something
that,
to
my
knowledge,
I
don't
know
if
that's
something
that
we
can
pick
from
people,
I
don't
know
so
maybe
we
should
just
sort
of
go
ahead.
Maybe
we
don't
support
eval
well,.
A
D
A
A
E
James
is
just
saying
it
should
be
configurable
and
documented,
which
is
true
so
yeah,
maybe
maybe
the
ultimate.
Maybe
the
ultimate
thing
we
need
to
achieve
is
how
to
configure
how
to
make
this
configurable
and
documented.
So
maybe
we'll
put
this
onto
that.
There
is
an
issue
in
image,
builder
repo,
so
try
and
work
out
how
we
can
configure
this
so
that
people
get
what
they
want
when
they
use
image
builder
and
document
what
would
be
deployed
by
default.
So
what
to
download
by
default.
D
Okay,
I
mean
I
agree
that
it
should
be
all
those
things.
My
big
question
here
is
priority.
How
important
is
this?
Is
this?
Am
I
the
only
person
that's
had
this
problem,
and
is
this
I'm
just
wondering
like
some
intuition
here
like?
Is
this
just
I'm
doing
something
stupid
and
it's
a
theoretical
thing,
we're
fixing
or.
A
I
think
this
depends
on
how
widely
these
imagers
are
being
used.
If
these
images
are
being
used
kind
of
in
any
sort
of
production
capacity.
I
would
say
this
is
fairly
high
priority,
but
if
it's
just
for
up
like
getting
tests
up
and
running
or
kind
of
development
scenarios,
then
it's
probably
not
too
high
of
priority.
A
But
I
would
make
sure
that
if
anybody
is
using
image
builder
to
build
images
to
run
cappy
clusters
in
any
sort
of
production
workload
that
they're,
just
following,
like
the
licensing
guidelines
and
everything
of.
D
G
Yeah,
so
to
be
clear,
we
sort
of
faced
this
issue
when
we
were
building
ci
images
for
openshift
and
we
did
not
pursue
the
route.
But
what
at
a
high
level,
I
think
jay.
We
sort
of
discussed
this
some
time
ago,
have
a
configurable
field
that,
and
that
is
what
james
is
suggesting
to
like
in
the
packer
script
or
in
the
subsequent
builder
that
you
are
using,
which
is
like
post
build.
G
You
can
say
that
this
is
the
environment
variable
which
contains
the
sku,
and
then
you
can
generate
the
auto
unit
and
dot
xml
based
on
this
queue,
and
you
can
use
that
for
the
new
machine
that
you
want
to
build.
A
Yeah
yeah,
it's
it's
all
related
tied
into
that
those
all
of
those
unattended
xmls
that
get
passed
to
set
up
to
when
you're
installing,
like
windows
from
an
iso
image.
Another.
G
Option
yeah.
The
other
thing
that
I
have
seen
people
doing
is
once
the
system
is
is
up
like
we
can
do.
We
can
unregister
and
then
register
again
with
the
new
skew
and
you
can
put
it
as
a
post
configuration
script
in
packer.
That's
something
that
I've
seen
people
doing
too.
A
Yeah,
I
was
going
to
suggest
some
something
similar
once
once
we
have
these
host
process
containers
or
the
privileged
containers.
It
might
even
make
sense
to
default
to
always
using
eval
skus,
and
then
somebody
could
schedule
just
a
demon
set
that
will
arm
or
up
upgrade
or
arm
any
licensing
mechanisms
there.
So
that
way
it
would
be
completely
configurable
and
end
users
wouldn't
have
to
muck
with
the
auto
and
attend
xml
files,
which
are
not
the
most
intuitive
and
definitely
not
easy
to
debug.
But.
H
I
mean
this
is
really
pointing
at
whether
we
should
be
asking
customers
to
use
images
built
using
image
builder
or
image
builder
images
in
production
right
and
the
answer
to
that
should
always
be
no.
It
doesn't
matter
whether
it's
configurable
or
not
would
be
my
take
because
you
don't
want
them
running
a
foul
of
any
sort
of
windows,
licensing
requirement
and
they
shouldn't
come
back
and
blame
us
said
you
told
us
to
do
this,
and
because
of
this
I
ran
a
fall
of
some
some
licensing
agreement.
A
Yeah
this
is
like.
I
believe
that
the
official
stance
is
anything
that's
in
the
image
builder
repository
are
only
for
reference
images
and
that
people
need
to
build
their
own
images,
but
maybe,
but
I
think
in
the
windows
case,
we
should
make
that
messaging,
stronger
and
kind
of
tie
that
to
licensing
things.
And
if
we're
doing
that,
then
I
think
what
I'm
hearing
is.
Maybe
we
should
like
always
use
eval
skus,
because
then
people
wouldn't
ever
run
amok
with
licensing
their
workloads
will
just
stop
working
after
30
days,
but
that's
probably
better
than.
A
B
I
A
Hey
laurie,
I
see
lori
joined
there
were
a
couple
people
interested
in
the
release
liaison
or
the
enhancements
liaison.
Did
you
want
to
talk
about
that?
Real,
quick.
A
Kind
of
introduced
it
briefly,
but
I
didn't
link
to
any
of
your
proposal
yet.
I
No
that's
great.
I
can
just
share
it
this
way.
So,
basically,
this
is
an
idea
that
spun
out
of
a
conversation
I
was
having
with
dims
many
of
you
knew
dims
he's
everywhere
in
kubernetes
fixing
the
things
and
on
the
architecture
sig
and
the
steering
committee
and
other
things
other
committees,
but
basically
over
time,
we've
noticed
that
responsiveness
to
the
release
team
can
be
tricky.
I
This
is
a
documented
issue.
There's
a
github
issue
from
117
about
it,
and
lots
of
improvements
have
been
made
since
then
to
improve
process,
but
we
still
would
like
more
collaboration
between
the
release
team
and
things
and
we'd
also
like
to
see
if
we
can
get
more
people
like
me,
involved
in
the
project
who
aren't
contributing
code
but
are
contributing
process
improvements,
keeping
things
running,
helping
cigs
break
down
topics
and
to
roadmaps
collecting
data
to
to
keep
things
just
moving
along.
I
So
this
would
free
up
the
chairs
and
tech
leads
from
doing
all
of
that
heavy
lifting
you
don't
really
need
to
have
detailed
coding
expertise
to
do
this.
You
basically
just
need
to
be
communicative
and
available
and
be
able
to
manage
your
time
so
that
you
can
respond
to
the
release
team
in
a
short
enough
period
where
you
help
them
keep
their
process
running
along
and
you
don't
leave
them
waiting.
I
So
if
you're
having
a
blog
post
on
an
enhancement
coordinating
with
them
to
make
sure
that
they
get
what
they
need-
and
you
know
just
basic
alignment-
collaboration
coordination,
communication-
and
we
would
time
box
this
role
for
one
release
cycle
so
that
you
know
when
it
will
end.
And
of
course
you
can
continue
doing
it.
If
windows,
if
you
like
doing
this
and
windows
needs
to
help,
but
you
know
just
to
release
pressure
on
folks.
I
I
E
A
Yeah,
I
was
just
going
to
add
that
any
of
the
leads
myself
james
or
jay
will
be
willing
to
help
out
as
well.
So,
if
anybody's
interested,
I
guess,
yeah,
let
us
know
we're
interested
in
helping
people
who
want
to
kind
of
take
on
this
role
for
our
sig.
D
A
So
for
this
release
we
didn't
have
many
big
enhancements.
There
were
for
sig
windows,
most
of
the
enhancements
that
are
that
were
kind
of
in
flight
had
wrapped
up
in
the
119
120
time
frame.
There
were
two
new
enhancements.
Neither
of
them
made
it
fully
into
this
release
as
an
alpha
release.
That
was
the
the
host
process
containers,
but
we
are
still
pursuing
alpha
in
122
and
then
the
log
viewer
work
that
our
event
was
writing,
and
hopefully
that
will
also
pursue
alpha
in
122.
I
What's
the
reason
for
the
slip
that
you
just
need
more
folks
working
on
the
cups
or
what
was
the
core
issue
there.
A
The
reason
for
the
host
process
slip
was
largely
just
reviewer
bandwidth,
both
the
kep
review
and
the
code
review
for
the
implementation
came
in
either
the
night
before
or
after
the
enhancement
or
code
freeze
deadline
respectively,
and
it
was
just.
There
was
a
lot
of
features
that
were
graduating
to
beta
or
stable,
that
all
of
the
reviewers
said
rightly
took
kind
of
precedent
over
that,
so
the
enhance
or
the
the
log
fewer
one.
A
I
think
that
there
were
still
some
open
questions
or
concerns
from
some
folks
that
I
think,
are
still
trying
to
be
resolved,
mainly
how
to
handle
some
of
the
kind
of
linux
and
windows
differences.
I
H
H
I
have
also
gotten
distracted
and
not
refocused
on
it,
but
I
think
the
main
the
main
contention
was
trying
to
make
this
invisible
to
the
end
user.
This
new
feature
that
we're
adding
invisible
to
the
end
user,
whether
they
need
to
look
at
a
log
file
or
know
that
you
know
the
cubelet
is
running
as
a
service
and
that's
sort
of
like
a.
I
don't
know.
H
I
couldn't
think
of
a
good
solution
for
that,
and
I
think
we
like
mark-
and
I
discussed
with
with
moz
that
you
know
given
the
nature
of
the
feature
and
given
that
it's
a
cluster
admin
only
feature
that
sort
of
differentiation
was
okay
in
our
respect
and
we
were
hoping
to
get
some
more
feedback
with
with
folks
who
are
trying
to
use.
This
feature
add
that
to
the
cap,
and
then
I
was
going
to
respond
to
peter
that
hey.
You
know
this
is
an
ex.
This
is
a
advanced
feature.
H
A
Yeah,
I
think
that
there
I
had
tried
to
have
some
folks
comment
on
the
issue
and
I
don't
think
anybody
active
like
took
a
look
more
folks,
but
this
might
be
a
good
platform
for
that
to
see
if
anybody's
interested
with
this,
do
you
want
to
bring
this
up
again?
Should
we
talk
about
this
at
the
next
meeting
pretty
early
on
arvind?
I
think
we're
kind
of
running
out
of
time
here
yeah.
I.
A
Okay,
yeah.
I
think
that
the
the
biggest
kind
of
blocking
question,
though
to
summarize,
was
how
to
make
some
of
these
operations
more
transparent
to
end
users,
and
I
think
what
our
event
is
saying
is
that
we
don't
necessarily
want
to
do
that,
because
this
is
kind
of
advanced
functionality,
so
just
convincing
some
of
the
reviewers
that
are
providing
use
cases.
For
that,
I
think,
is
what
will
help
move
this
forward.
Is
that
correct.
I
We've
talked
before
about
some
of
your
old
backlog
items
like
the
ones
that
aren't
as
urgent
seem
not
to
get
touched,
and
so
how
could
we
integrate
some
of
those
into
your
122
activities
if
it
makes
sense,
given
your
other
priorities,
you
also
had
talked
about
some
of
your
contributor
goals,
like
how
many
commits
could
you
get
how
many
new
contributors
could
you
get?
So
if
you
saw
a
comprehensive
plan
for
122,
the
covering
all
of
your
goals
might
be
useful.
We
could
do
that.
D
Like
that-
and
I
feel
like
ravi's
I
mean
arvin's
thing
is
like
it
could
find
new
life
somehow
if
we
moved
it
into
like,
if
we
functioned
as
a
sig
more
like
a
startup,
we
could
make
a
plan
that
wasn't
going
to
get
interfered
so
much.
Why,
by
the
fact
that
we
can't
get
everything
in
linux
and
windows
to
be
the
same
right,
so
I
feel
like
if
we
combined
this,
I
feel
like
we
could.
D
A
So
this
one,
this
particular
cup,
is
probably
not
a
great
use
case
for
trying
to
you
know,
spin
up
a
separate
project
that
is
only
impacts
windows,
but
in
the
future
there
may
be
some
opportunities
for
that.
But,
yes,
that
sounds
great
lori.
Let's
maybe
devote
some
or
most
of
next,
the
next
community
meeting
to
kind
of
going
over
some
122
plans
and
and
priorities.
A
A
A
D
To
to
revisit
that,
though,
mark
the
whole
idea
of
the
whole
idea
of
being
able
to
easily
get
logs
from
windows
notes
is
still
a
generic
problem.
I
think
right,
it's
not,
I
feel
like
we
could
have
a
junior.
I
still,
I
feel
I
still
feel
like
we
could
think
outside
the
box
on
that
solution
or
something
and
build
off
that
cap,
even
though,
if
the
implementation
is
different,.
H
H
A
Yeah
until
we
have
the
privileged
containers,
I
think
anything
that
requires
services
to
be
installed
and
running
on
the
node
to
kind
of
light
up.
Functionality
is
kind
of
a
pretty
high
barrier
to
entry,
so.
D
A
But
yeah
saying
hey,
you
needed
to
do
all
this
stuff
on
your
nodes
before
you
wanted
to
get
the
logs
for
for
users
on
windows
where,
with
linux
you
can
just
do.
You
know,
cube
ctl
logs
the
node
it
it
highlights
kind
of
some
pretty
big
deficiencies.
It
would
highlight
some
pretty
big
deficiencies
in
the
windows
ecosystem.
H
Well,
mark
for
the
record:
they
don't
have
it
for
linux
either
for
the
cube
cdl
logs.
That's
what
this
feature
is
you.
C
D
A
Let's
continue
this
next
next
community
meeting,
I
think
already
over
time
all
right,
I'm
going
to
stop
the
recording
and
hand
it
over
to
jay
for
the
pairing
sessions
which
we
are
doing
each
week
after
this.
So
thanks
all
for
attending
and
hope
to
see
you
next
week,
bye.
D
Okay,
everybody.
So,
let's
see
I
I
was
thinking
I
didn't.
I
was
thinking
I
wasn't
sure
yet
about
whether
we
talked
about
two
things
this
week,
which
is
windows
proxy
out
of
tree
and
the
other
one
was
the
whole
deal
of
this
image.
Builder
auto
attend
stuff,
and
I
feel
like
since
perry
is
here,
but
I
don't
know
how
long
we're
going
to
have
him.
D
D
Does
anybody
want
to
talk
about
anything,
though,
before
we
start
jump
and
start
jumping
into
things?
I
see
friedrich
is
here.
Maybe
you
want
to
give
an
update
on
your
stuff.
F
F
I
don't
want
to
crash
this
thing
again
after
last
time.
Okay,
so
a
quick
update,
yeah
I
created
I'm
working.
F
Building
and
windows,
linux
virtual
machine
cluster
and
I
tried
to
optimize
it
as
as
far
as
possible,
and
I
had
a
lot
of
trouble
with
so
I'm
using
vagrant
in
a
vagrant
file
and
a
lot
of
trouble
with
pushing
around
windows
with
fragrant,
especially
if
you
have
to
do
reboots,
that's
no
fun.
So
what
I'm?
I
I
feel,
I'm
a
little
embarrassed
to
say
this.
So
the
the
point
is,
if
you
send
the
commands
directly
to
via
with
your
vid
rm
via
your
linux
console,
you
send
them
directly.
F
It
works
like
a
charm,
but
if
you
do
it,
if
you
write
it
down
into
a
vagrant
file,
you
get
errors.
So
I
thought
okay.
But
how
can
I
optimize
this
then?
And
then
I
again
have
a
feel
a
little
ashamed.
I
put
it
into
a
make
file
and
then
it
works
and
if
there's
still
some
issues,
because
you
have
to
escape
so
many
characters,
basically
from
the
mag
file
to
I
think,
vagrant
and
then
to
powershell
and
yeah,
you
have
to
escape
a
lot
of
characters,
dollar
signs
and
whatnot.
F
D
Yeah
so
awesome,
so
what
I
told
friedrich
was
it's
really
cool
that
you've
got
something
automated
that
you've
built
along
that
you
we're
gonna
finally
have
like
documented
and
stuff,
so
I
think
it's
really
cool,
even
without
cni,
to
get
this
thing
as
a
v0,
so
good,
good,
good
job,
frederick,
looking
forward
to
seeing
that
first
pr,
no
matter
how
ugly
it
is,
it's
going
to
be
awesome,
so
perry
I
feel
like.
D
Maybe
he
might
be
able
to
get
some
support
from
you
at
this
point
since
he's
pretty
deep
in
this
stuff,
so
friedrich
feel
free
to
reach
out
to
perry
he's
done.
This
is,
I
feel
like
the
point
where
you
can
start
getting
a
lot
of
help
from
him
and
and
maybe
some
other
people,
but
I
feel,
like
you
know,.
D
Yeah
so
cool,
so
friedrichs
project
updates,
so
okay,
so
the
big,
so
the
big
issue
was
is,
though,
for
you,
I
just
want
to
get
to
make
sure
we
get
this
down.
I
should
probably
be
recording
this
hold
on
one.
Second,
it's.