►
From YouTube: Kubernetes SIG Windows 20210302
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
march
2nd
2021
iteration
of
the
sig
windows
kubernetes
community
meeting.
As
always,
these
meetings
are
recorded
and
uploaded
to
youtube
so
make
sure
to
adhere
to
all
of
the
cncf
code
and
conduct
and
just
follow
that
anyway,
it's
good
advice,
all
right.
Let's
get
started
kind
of,
don't
have
a
super
long
agenda
today,
but
I
have
a
feeling.
Some
of
these
topics
are
going
to
fill
up
most
of
the
time.
A
First,
one
is,
as
we've
mentioned
previously,
we're
compiling
the
sig's
annual
report
for
2021.
jay
james,
and
I
met
a
little
bit
last
week
and
jay
graciously
kind
of
filed
an
initial
pr
for
that.
If
anybody
is
interested
in
looking
or
commenting,
please
just
comment
on
the
pr
not
going
to
spend
too
much
time.
A
Looking
at
that,
because
I
think
we're
going
to
talk
about
that
later
with
this
agenda
item
but
yep,
that's
open,
it's
gonna
get
published
and
we
want
every
feedback
from
everybody
about
how
we
can
continue
to
grow,
keep
an
open
community
and
accessible
community
and
and
all
that
good
stuff.
A
First
off
I
moved
this
agenda
item
up.
I
don't
know
if
jing
is
on
the
call.
It
looks
like
there's
an
issue
with
lsat
elstat.
Yes,.
B
It's
quite
strange:
suddenly
I
noticed
in
certain
hours
like
a
test
in
some
like
a
windows
version.
This
stat
will
not
work
for
socket
file
and
after
the
issue
I
added
it
will
show
like
when
you
use
os
stat
for
that
file
at
the
bottom.
There
is
an
arrow
showing
the
bottom
yeah
at
the
bottom.
B
B
C
We
we
had
to
do
a
similar
change
to
use
elstat
for
the.
C
For
the
work
on
the
device
manager-
and
so
I
can
link
to
this
pr
but
that
we
had
to
do
the
same
thing,
so
I
think
it
had
to
do
with
the
versions
of
os's
and
a
mixture
of
that
with
the
golang
versions.
B
B
So
this
is
only
for
checking
if
it's
a
directory
and
if
it's
not
that
there
actually
is
a
file,
it
will
go
ahead
to
check
the
if
it
is
a
domain
socket.
So
I
think
this
case
is
safe.
B
A
Or
related
under
the
hood
on
windows,
all
right,
if,
if
everybody
agrees,
is
there
anything
else
to
kind
of
discuss
here?
I
think
that,
let's
make
sure
that,
on
this
pr,
we
trigger
some
of
the
windows
test,
the
pr
tests-
I
can
do
that
after
the
meeting
and
just
to
get
some
extra
verification.
But
if
I
guess
that's
good,
I
think
we
have
reviews.
D
I
don't
think
I
have
any
expertise
in
the
internals
here.
I
just
tried
it
out
and
I
I'm
saying
the
same
thing
with
a
test
unix
socket,
so
I
think
it's
fine.
A
Okay,
yep
we'll
trigger
the
the
unit
tests
or
the
the
pr
tests,
and
then
I
guess
we
can
get
that
going.
A
All
right,
I
think
next
we
can
look
at
either
one
of
these
agenda
topics.
Unless
anybody
has
anything
else,
lori,
I'm
probably
going
to
have
you
kind
of
introduce
yourself
first,
if
that's
okay,
and
then
we
can
discuss
this
first
and
then
we
can
go
into
the
grow
in
the
community.
Would
you
like
to
introduce
yourself
laurie.
E
Sure
sure
so
my
name
is
lori
apple,
I'm
the
program
manager
for
sig
release.
I
have
been
that
since
last
summer,
I'm
also
a
staff
technical
program
manager
at
vmware,
and
I
work
with
jay
and
maybe
others
on
this
call,
and
I
have
felt
recently
that
the
work
in
sig
release
has
gone
well
enough,
that
I
can
step
back
and
apply
my
efforts
to
a
different
take
that
could
use
some
support
with
program
management
and
process
workflow
issues.
A
And
in
addition
to
that,
lori's
been
very
helpful
in
I
know
michael
michael
and
I
have
been
working
kind
of
behind
the
scenes
with
laurie
to
improve
some
of
the
process
with
sig
windows
and
a
lot
of
the
kind
of
intersec
interactions
with
sig
windows.
As
a
lot
of
folks
here
know
a
lot
of
the
sig
windows.
A
Pr's
and
changes
require
reviews
and
approvals
from
multiple,
usually
three
or
four
different
cigs
at
a
time,
and
that's
usually
has
been-
it's
been
a
little
bit
of
a
hurdle,
especially
for
for
new
contributors,
just
kind
of
having
to
deal
with
all
of
the
extra
process
with
that.
So.
E
Yeah,
I
feel
like
I
really
didn't
do
very
much
for
you
at
all.
I
just
had
a
chat
with
you
and
then
yeah
things
happened
over
the
winter.
E
Just
you
know
how
life
goes,
but
you
know
now
I
I've
what
I
was
actually
doing
was
spending
a
lot
of
time
in
the
enhancement
side
project
which
I'm
a
co-owner
just
to
try
to
get
that
to
a
place
where
we
could
start
making
some
of
the
process,
changes
that
are
needed
there,
and
so
that's
also
going
well
with
some
new
contributors
helping
out.
So
that
frees
up
more
time
to
be
here
and
actually
help
you
do
some
of
that
cross-sig
collaboration.
E
So,
as
I've
noted
on
the
agenda,
I've
actually
drafted
governance
docs
for
the
creation
of
an
official
program
manager
role,
I'm
not
sure,
that's
the
direction
steering
committee
wants
to
go
in.
We
were
discussing
that
last
night,
but
in
any
case
you
could
read
this
documentation
to
understand
what
that
role
would
do
and
some
of
the
aspects
of
program
management
and
so
to
get
a
sense
of
what
your
needs
might
be
and
how
I
could
help
you
might
see
something
there.
E
A
I
have
one
thing:
that's
kind
of
coming
to
mind.
I
think
irvin
was
on
the
call
earlier
it
also.
I
don't
know
if
he's
still
on
the
call
like
his
feedback
here.
G
A
I
think
one
of
the
things
that
we
I
feel
like
we
really
struggled
with
in
this
past,
or
this
current
release
in
121,
was
for
some
of
the
enhancements.
We
had
just
getting
some
of
reviewers
to
even
take
a
look
at
this
before
the
enhancement
freeze
deadline
was
there,
I
know
with
one
of
the
cups
I
put
forward.
A
One
of
the
major
reviews
didn't
happen
until
the
monday,
before
the
deadline
or
before
the
enhancement,
freeze
and
kind
of
it
was
very
tight,
and
then
we
had
to
involve
a
bunch.
A
number
of
other
people,
including
people
on
this
call,
and
it
just
kind
of
felt
a
little
bit
rushed,
and
I
think
I
don't
want
to
put
words
in
arvin's
mouth,
but
I
think
he
had
a
fairly
similar
experience
with
the
kept
that
he
was
trying
to
push
through.
So.
H
Yeah,
I
think
mark
said
pretty
much
what
I
would
have
said
too.
It
was
a
little
bit
of
a
struggle,
and
I
see
this
as
a
pattern
and
I'm
not
even
sure
how
we
could
really
resolve
it,
especially
when
it
comes
to
needing
some
review
regarding
apis.
Given
that
there
are,
I
mean,
if
you
look
kubernetes
wide,
it's
just
five
folks
who
can
actually
approve
a
new
api
change
or
edition.
H
However
minor
or
major,
it
is-
and
you
know
that's,
I
guess
just
general
roadblock
across
kubernetes,
so
I
I
don't
know
whether
it's
even
just
sick
windows
specific
it
could
be.
You
know
a
general
problem
that
is
there
today.
A
Okay,
so
I
guess
that's
kind
of
two
things
that
maybe
we
might
kind
of
propose
something.
E
Some
of
your
help
with
okay,
I
dropped
a
diagram,
I'm
working
on
to
actually
diagram
the
enhancements
process,
so
you
can
just
take
a
look
at
that
at
your
leisure
because
it
it
might
give
you
some
ideas
about
where
the
process
isn't
working
for
you
and
then
those
changes
that
you
might
suggest
or
pain
points
we
can.
I
can
take
those
into
the
enhancement,
sub
project
and
discuss
those
with
the
group,
because
this
is
what
we've
been
doing.
E
A
lot
of
in
recent
months
is
discussing
where
the
process
is
gaps,
where
it's
not
working,
where
it's
too
heavy,
where
it's
not.
You
know
not
complete
enough.
So
that's
one
aspect
for
the
api
reviews
yeah.
I
think
that
one
that
one
is
challenging
because
it
requires
people
to
have
the
ability
to
do
the
job
of
reviewing
and
approving-
and
I
imagine
that's
a
small
group
so
have
you
been
able
to
have
a
member
of
this
sig
become
part
of
that
group
or
at
least
have
a
liaison.
A
I
don't
think
we've
explored
that
yet
this
so
we're
kind
of
in
an
interesting
like
in
the
past
couple
of
releases,
the
sig
windows
team
was
really
finishing
up
a
lot
of
caps
that
had
been
introduced
in
like
116
and
117..
A
This
is,
I
feel,
first
kind
of
released
that
we're
introducing
new
caps
kind
of
from
the
beginning.
So
I
think
that
that's
a
little
bit
of
that
is
that
we
haven't
kind
of
built
those
contacts
or
with
people
recently
for
that,
but
we
or
and
haven't
really
run
into
this
issue
in
the
past.
So
I
can't
say
that
this
is
a
pattern,
yet.
A
One
thing
just
kind
of
off
the
top
of
my
head
without,
like
I
haven't,
looked
into
this
too
much,
but
I
I
almost
wonder
if
it
would
be
nice
to
have
a
separate
deadline
for
when
enhancements
kind
of
can
be
submitted
and
expect
to
actually
make
it
into
the
review
or
into
the
release
versus
when
the
all
the
reviews
needed
to
be
complete.
I
also
got
added
to
another
enhancement,
cap
or
cap
from
sig
node.
I
think
the
monday,
before
the
enhancement
freeze,
to
review
on
behalf
of
sig
windows
and
so
like.
A
I
would
also
that,
fortunately,
that
kept
was
in
previous
discussions.
It
was
agreed
that
that
wouldn't
be
into
or
in
121
it
would
be
reviewed
to
go
into
the
next
release,
but
I
also
could
have
been
on
the
like
the
the
cause
of
a
kept
delay,
because
of
that
too,
have.
E
A
E
So
the
tldr
on
that
is
that
it
might
solve
your
the
specific
issue,
but
you
have
to
opt
in
to
the
release
as
opposed
to
here's.
A
tracking
sheet
put
your
cups
in
there
and
then
a
release
team
will
come
and
ask
you
what
your
plan
is.
So
it's
making
a
deliberate
action
of
choosing
through
the
magic
of
yaml
to
go
and
say
this
cap
is
for
x
release.
You
know
this
release
so.
A
I
see,
I
think,
the
issue
that
we
ran
into
this
past
release
and
on
on
both
the
receiving
and
the
giving
end
of
this
or
I,
I
feel
that
kind
of
during
the
review
process.
Other
reviewers
can
kind
of
say
at
any
point
like.
I
think
this
other
sig
needs
to
also
be
marked
as
a
reviewer
or
an
approver
for
this
cap,
and
that
can
happen
pretty
late
in
the
review
process.
E
Yes,
maybe
we
should
actually
do
like
a
little
bit
of
a
audit
like
or
you
can
on
your
own
time,
or
we
can
do
it
together,
where
we
look
through
that
process
and
and
just
note
where,
where
you're
getting
held
up
and
that
could
at
least
make
some
momentum
happening
around
your
kep
process
and
then
I
think,
for
I
wanted
to
ask
about
your
general
workflow
for
prs
issues
moving
across
your
project
board.
Are
you
managing
that
fairly
well
or
do
you
need
some
support?
There.
A
I
have
my
ideas
on
that.
Does
anybody
else
want
to
comment
before
I
kind
of
monopolize
this
conversation?
If
not
I'm
happy
to
comment
on
that.
F
Keyboards
thing
that
we
need
to
grow
the
community
and
I
feel
like
that's
the
bottom
line
there.
We
need
people
interested
in
digging
in
that's
what
I
feel,
but
I
I'm
probably
outvoted
there,
because
there
are
probably
some
other
improvements
we
could
make,
but
I
think
that's
the
bottom
line
for
me.
F
Yeah,
like
that,
like
we
just
need
whether
it's
a
board
or
just,
I
feel
like
we're,
not
I'm
not
100
sure
where
yeah
like
yeah.
I
feel
like
looking
at
other
cigs
that
I'm
involved
with
like
sig
network,
I
feel
like
there.
We
need
maybe
a
little
bit
more
organization
because
there's
a
lot
of
people
and
they're
not
quite
sure
where
to
help
I
feel
like
here.
We
have
the
problem
of
everybody's
really
busy
and
we
just
need
people
who
we
want
to
get
involved.
H
I
mean
jay,
I
think
we
should
separate
this
out
into
into
sort
of
two
concerns
things
like
you
know:
backlog,
grooming,
backlog,
refinement
and
making
sure
our
project
is
in
good
order.
The
issues
are
being
tracked.
I
don't
think
we
need
a
lot
of
people
to
do
that.
Actually,
if
we
had
a
lot
of
people
doing
that,
it
would
be
a
little
hard
to
coordinate
a
small
subset
who's
willing
to
do
that
on
a
regular
basis.
H
B
A
I
I
my
kind
of
feeling
is
is
that
we
have
we
have
a
backlog
and
we
have
a
project
board
and
I
think
we
do
an
of
okay
to
good
job
on
keeping
track
of
all
of
the
work.
That's
actually
being
worked
on
on
that
board,
where
I
don't
think
we
do.
A
good
job
is
taking
and
triaging
a
lot
of
issues
that
come
in
that.
I
think
we
all
like
jay
mentioned
that
the
people
who
are
triaging
and
grooming
that
backlog
know
that
we
won't
have
time
to
work
on.
A
So
we
have
a
lot
of
things
on
the
backlog
that
I
think
kind
of
like
it
almost
feels
like
once
it
gets
put
in
the
backlog
it's
get
stuck
in
the
backlog,
but
for
work
that
we're
actively
working
on.
I
think
we
do
do
a
fairly
good
job
tracking
that,
on.
I
Feel-
or
this
is
just
my
son-
that
came
to
my
mind-
is
some
of
the
problem
that
we
have
in
sig
windows.
The
fact
that
nothing,
nothing
in
sig
windows
is
just
a
sig
windows
problem.
It's
like
nine
times
out
of
ten.
If
we
raise
an
issue
with
sick
windows,
it's
going
to
be
on
sig
node,
or
it's
going
to
be
on
sig
network.
I
It's
like
windows,
networking
or
it
could
be
testing,
which
is
obviously
you
know,
sig
testing
and
the
windows
kind
of
crosses,
so
many
different
things
that
that
causes
some
issues
as
well,
because
it's
never
just
a
sick
windows
problem,
whereas
with
sig
network,
it's
always
a
sig
network
problem.
If
it's
a
networking
issue,
does
that
make
any
sense.
A
A
E
That's
a
challenge
all
right.
Well,
I've
taken
notes
and
I'll
give
something
to
this,
but
we'll
also
catch
up
on
slack,
I'm
sure.
A
A
Yep,
all
right
jay
do
you
want
to
kind
of
take
over.
I
think
we've
kind
of
touched
on
a
lot
of
what
you
already
wanted
to
talk
about.
In
this
conversation.
F
Organic
extension
of
what
we're
talking
about
so
lori
is
there,
like
you,
know
a
you
know:
here's
your
first
day
working
on
kubernetes
thing
that,
like
I
should
point
people
to
because
kind
of,
as
you
know
like
I,
you
know
we
talked
about
this
and
it's
like.
Well,
I
guess
one
of
my
goals
is
to
get
people
hands
on
keyboards,
doing
stuff
and,
like
I
think,
there's
going
to
be
a
few
people
who
are
it?
Well,
I'm
not
sure
how
this
is
going
to
go.
F
Maybe
there'll
be
zero
people,
but
maybe
there's
there'll,
be
nobody
or
there'll
be
some
people
that
haven't
done
anything
on
upstream
kube
that
try
to
get
involved
or
there'll
be
people
who
are
already
in
upstream
coop
and
they
just
need
to
understand
windows,
there's
kind
of
three
types
of
categories
that
I
could
see
needing
to
help
people
with,
and
I
guess
for
the
the
middle
category
right
the
people
who
have
never
really
worked
on
cube
before.
Is
there
a
here's
your
first
day
working
on
upstream
coop?
F
Here's
what
you
need
to
do
the
guide
that
I
can
point
people
to
so
that
I
don't
you
know,
have
to
spend
time
telling
people
github,
repos
and
kind
and
all
the
other
basics.
You
know
I
mean
like
the
10
things
that
everyone
has
to
do.
E
G
E
Working,
I
was
through
paris,
she
was
my
entry
point
into
the
project
and
giving
me
issues
to
work
on,
but
I
know
that
there
are
resources
and
contrabax.
E
I
know
that
the
last
release
team-
I
did
a
conversation
with
them
about
what
about
the
project
wasn't
clear
to
them
and
they
they
were
suggesting
that
the
documentation
about
how
the
ecosystem
is
laid
out
and
how
things
interact
with
each
other
and
how
the
the
project
works
at
a
systemic
level
that
that
information
is
either
they
didn't
discover
it.
So
they
didn't
know
about
it
or
maybe
it
doesn't
exist.
E
E
The
third
thing
is,
I
would
suggest
a
team
ways
of
working
agreement,
considering
that
that
enables
you
to
establish
a
set
of
protocols
and
how
how
you
like
to
work,
how
you
communicate,
how
you
prioritize
work
and
that
can
be
really
valuable
information
for
them
to
start
understanding
what's
going
on,
so
they
show
up
at
a
meeting.
Okay,
who
are
these
people?
Why
are
they
meeting
like
this?
All
of
that
is
described,
and
then
they
can
participate
more
effectively.
It
just
accelerates
their
involvement.
You
have
you
have
a
doc
for
that
somewhere
that
you
have.
E
J
A
I
was
all
right,
I
I
we
should
yeah,
we
could
take
that
and
update
our
contributing
guide.
With
all
of
this
information,
I
think.
E
E
But
we
did
the
ways
of
working
agreement
in
a
group
meeting
and,
and
it
brought
people
their
personalities
out
because
they
were
building
this
culture,
a
cultural
artifact,
and
so
they
they
sort
of
shifted
from
all
the
tech
stuff
to
like
okay,
who
am
I?
What
am
I
here
for
and
that
helped
them
to
connect
with?
I
think
the
meeting
and
the
group,
because
the
meeting
changed
a
lot
through
that
process
and
some
of
the
follow-up
processes
about
just
talking
about
what
people's
motivations
were
and
what
we're
interested
in.
E
E
E
F
F
A
I
need
to
drop
in
just
a
minute
too,
but
ireland
did
you
want
to
say
something
before.
H
Yeah
I
was
thinking
about
what
jay
is
trying
to
do
and
improve.
You
know
kind
of
build
the
whole
sig
windows
community
and
I
feel
like
the
main
missing
bit
for
us,
is
not
even
building
the
sig
windows
community.
It's
more
about
getting
sort
of
you
know
in
in
every
other
special
interest
group
out
there
signo,
you
know,
stick
cli.
It
started
to
get
a
windows
presence
in
there
because
the
main
issue
it's
like
when
I
was
at
cigot
right.
H
I
don't
know
much
of
what's
going
on
there,
but
if
I
feel
like,
if
there's
no
windows
presence
there,
they
would
make
a
decision
that
might
break
us
and
it's
not
they're
doing
it
on
purpose.
They
just
sometimes
don't
think
about
us.
So
that's
the
other
area,
where
I
think
we
should
put
some
thought
into
as
in
how
we
could
have
precise
presence,
a
windows
presence
in
these
different.
You
know
things
because
I
think.
F
A
I
always
just
need
more
people
right.
I
was
going
to
say
I
feel
like
we're
kind
of
rounding
that
corner
with
sig
node
a
lot
of
times,
folks,
derek
alana
will
message
or
like
include
james
and
myself
on
pr's
that
come
up.
I
think
it
took
a
lot
of
I've
heard
this
term.
A
lot
you
know
just
chat
would
carry
water
just
kind
of
doing
some
of
that
work
and
building
those
connections
and
sig
node.
A
G
F
Yeah
we
need
we,
we
we
yeah
so
network.
I
feel
like
we're
good,
because
we've
got
me:
we've
got
antonin.
We've
got
most
of
the
andrea
folks
that
are
deeply
invested
in
windows
there,
so
node
we've
got
mark.
Where
else.
If
there's
one
place,
arvind
that
you
felt
like
we
have
the
biggest
problem,
you
think
it's
off,
I
think
so,
okay
good,
so
that
could
be
a
place
where
we
could
try
to
maybe
get
somebody
to
start
being
involved
across
now.
Okay,
is
anybody
here
interested
in
doing
more.
F
More
more
stuff
getting
hands-on,
keyboards
contributing
getting
involved,
because
that
was
that
was
kind
of
what
we
were
going
to
do
after
this
meeting
was
over,
and
this
is
the
first
time
we've
kind
of
ever
announced
this
idea
of
spending
an
extra
hour
after
the
meeting
hacking
on
stuff.
F
F
Who
is
that
free
drink,
okay,
free
drink,
so
we've
got
okay
friedrich
and
looks
like
everyone
else.
James
jerry
army,
jocelyn
peter
everyone
else.
Here
is
a
sig
windows
person.
For
the
most
part,
it
looks
like
so
cool
friedrich,
so
we've
got
a
audience.
We've
got
one
potential
new
new
person
on
board
today.
So
where
are
you
at
in
your
in
your
what
so
so
so
I
guess
the
first
question
I
would
have
is
what's
your
time
allotment
and
what's
your
like
level
of
experience
with
kubernetes
at
this
point.
G
F
All
right
cool,
so
all
right,
so
if
anybody
else
doesn't
have
any
other
questions,
I'm
gonna
just
dive
in
and
start
onboarding
friedrich
for
today
and
anybody
else
who
wants
to
get
involved
and
like
ask
questions,
but
I
think
the
like
a
fish
and,
of
course,
hang
out
and
the
more
people
hang
out
the
better,
but
the
official
sig
window
stuff,
I
think,
is
pretty
much
pretty
much
over
at
this
point.
So,
okay,
cool
friedrich.
So
let's
see
how
we
can
get
you
started
so
I'll
share.
My
screen
here.
F
Okay,
so
you
see
my
window
right,
see
yes
see
you
can
you
can
you
can
talk
spanish
to
me
as
well?
If
you
want
but
then
what's
italian,
oh
yeah,
that
works
too,
so,
okay,
so
all
right
so
to
start
kind.
You
know
all
right.
So
to
start
like
the
number
one
thing
for
me
for
getting
started
is
you
know,
you've
got
to
have
fun
and
you've
got
to
run
somewhere
and
you've
got
to
try
to
install
a
cluster
with
it
and
try
to
delete
a
cluster
with
it.
F
C
E
I
F
D
F
F
G
Maybe
that
anyways-
maybe
you
can
give
me
your
material,
what
you
have
and
this
document
that
you're
showing
I
see
for
the
first
time,
but
sorry,
I'm
terribly
prepared,
I'm
basically
jumping
into
this
but
yeah.
I
can
take
all
the
material.
I
can
find
and
then
give
some
feedback
on
what
might
be
missing.
F
I
have
a
very
specific
thing
in
mind
here,
so
we
you
know
like
we,
we
basically
between
folks
like
james
and
perry
and
me
and
whatnot.
F
C
Just
down
way
at
the
bottom
is
says
so:
there's
the
link
to
the
kubernetes
contributor
and
then
there's
a
sig
windows
contributor
guide.
I
know
he's
contributing
there.
Oh.
F
Cool
cool,
so
we'll
go
here,
okay,
so
this
is
okay,
so
we've
got
block
slack
outstanding
pr,
okay,
building
kubernetes
for
windows
from
source;
oh
great
kubernetes
build
scripts.
Oh!
This
is
great.
This
is
like.
I
was
going
to
do
all
this
from
okay
yeah.
I
mean
oh
man.
Oh,
this
is
great.
Okay,
yeah.
C
F
F
J
F
F
F
I
just
you
know
I
can
make
a
pull
request
here
right
to
make
sure
we
link
specifically
to
that
so
where's,
our
I'm
gonna
go
back
to
the
onboarding.
K
I
I've
got
a
drop
guys,
but
I'll
catch
you
all
later.
Okay,
cool.
F
So
I'm
gonna,
I'm
gonna,
just
make
a
quick
pr
to
update
that,
because
I
think
that's
the
one
instruction
that's
kind
of
important
and
so
friedrich
just
fyi
here
this
these
when
you,
after
you
do
all
this
stuff
of
like
patching
and
and
all
that
stuff
that
they're
talking
about
in
here.
F
F
I'm
just
adding
that
link
just
because
I
think
it's
so
important.
So
after
you
read
all
this
stuff
in
here
friedrich,
it
talks
about
how
to
run
the
test
and
in
order
to
run
the
test,
which
is
oh,
where
is
it?
Where
is
it
here
in
order
to
run
these
tests
right?
You
go
into
this
sig
windows,
testing,
repo,
there's
a
bunch
of
instructions,
but
the
central
thing
there
is
building
the
n10
test.
Miner
got
it,
and
so
and
and
that's
in
there
that's
in
that
file.
F
But
what
you
do
is
you
build
that
end
to
end
test
binary,
and
then
you
point
it
at
your
cluster
and
then
it
runs
20
or
30
or
40
or
however
many
tests
you
tell
it
to
run
and
that
understanding
the
entrance
has
like
an
animal
in
and
of
itself
that
you
know
you'll
learn
about
eventually,
but
basically
kubernetes
comes
with
hundreds
of
tests
and
after
you
change
something
you
want
to
run
at
least
some
subset
of
those
tests.
Typically
you'd
run
the
conformance
tests
that
you
know
like.
F
Yeah
so,
but
you
know,
there's
like
I
don't
know,
500
600
something
tests
or
whatever
in
kubernetes
itself,
so
the
conformance
is
a
subset
more.
F
Thousand
four
thousand
total
okay,
so
the
conformance
is
a
subset.
So
anytime,
frederick
did
you
want
to
see
if
something
is
a
kubernetes
cluster
or
not,
you
run
the
conformance
test.
That's
what
tells
you
it's
an
actual
kubernetes
cluster
and
that
it's
working
right,
so
that's
being
able
to
run
those
tests
is
central
to
any
change
that
you
ever
do
you
don't
necessarily
want
to
run
all
of
them.
F
Has
that
every
test
framework
as
well
the
ability
to
pick
a
specific
test?
So
that's
that
part
the
rest
of
this,
I
feel
like
you
could
roll
through
here
right.
I
feel
like
this
is
a
good
start
right,
we're
kind
of
moving
towards
only
testing
container
d
nowadays,
but
that
doesn't
mean
that
this
guide
is
obsolete
because
we
still
technically
like
support
docker.
F
If
you
wanted
to
run
container
d
v
and
james
have
a
separate
guide,
we
can
give
you
on
how
to
test
specifically
with
container
d
windows
and
like
what
that
involves,
and
I
think
that
might
be
better
for
you,
frederick,
since
you're
starting
from
scratch.
So
the
other
note
to
keep
here
by
the
way.
Are
you
taking
notes
here,
because
I'm
not
going
to
be
able
to
write
all
this
down.
G
I
count
on
I
try
to
start
writing
down
notes,
but
you're
too
fast.
For
me
I
will.
I
will
rely
on
rewatching
this
on
youtube.
F
G
F
F
That's
that
from
this
dock
is
that
when
you're,
when
we
get
started,
I
think
we
may
want
to
rather
than
running
darker
use
container
d
as
your
container
runtime,
because
that's
where
everything
in
kubernetes
is
moving
towards
right,
we're
not
really
running
docker
anymore.
It's
docker
shim
has
been
deprecated,
okay,
so
and
there's
plenty
of
resources
online
about
why
that
is
and
how
that
all
works.
But,
but
I
think
for
your
thing,
I
think,
for
a
quick,
tactical
win
on
your
end.
F
I
think
we
can
do
this
thing
of
like
getting
you
up
and
running,
getting
getting
up
and
running
building
a
recipe
for
how
to
build
a
container
d,
build
a
kind
cluster
and
then
bolt
a
a
windows.
Note
onto
that
cluster
I
feel
like.
Maybe
I
should
like
diagram
that,
like
does
that
make
sense
to
you
or
should
I
like
draw
it
out.
I
think
it
might
work.
G
Absolutely
I
will
be
super
happy
if
you
give
me
a
diagram
yeah.
F
F
F
I
have
a
github
repo
that
I
maintain
myself
and
we
can
just
use
that
for
this
to
start
and
we
can
literally
just.
F
Well,
actually,
this
is
actually
technically
a
sig
windows
issue,
isn't
it.
This
is
actually
a
sig
windows
issue
because
we're
going
to
be
contributing
this
to
sig
windows
tools
right.
So
maybe
this
should
be
a
github.com
kubernetes,
sigs,
sig
windows
tools,
and
then
we
should
make
an
issue
and
the
issue
should
be.
I
think
this
is
this-
is
how
we'll
do
it
right,
we'll
literally
say,
windows,
kind
instructions.
K
F
Okay,
cool
so
I'll
just
put
the
notes
in
here
friedrich
that
way
you
can
track
it
and
you
can
ask
questions
in
here:
issues:
145,
right,
perfect,
okay,
cool,
so
so
the
overall
architecture
right
you
have
xcd
and
you
can
stop
me
anytime.
You
want
okay,
so
I'm
just
going
to
sort
of
go
through
this
as
fast
as
I
can.
F
Okay,
it's
an
api
server,
so
this
makes
sense
so
far,
so
you
got
the
kx
api
server.
You've
got
the
scheduler
and
you've
got
the
kcm
right
and
that
all
talks
to
the
api
server
and
then,
after
that,
you've
got
you've,
got
a
kubelet
like
a
linux
one.
F
F
Normal
linux
clusters.
Look
like
that
right
so
and
you
might
have
multiple
of
these
right.
That's
kind
of
the
idea
right,
so
you
have
multiple
right
and
they
all
talk
to
the
api
server
right,
so
windows
cluster.
So
that's
really
easy
to
do
in
client,
right,
windows,
okay,
its
clusters.
Look
like
this
right.
F
They
have.
You
have
a
kubelet,
that's
like
a
linux
kubelet,
but
you
also
have
a
windows
module
right
well
and
actually,
technically
you
could
just
have
a
windows
kubler
and
not
have
a
linux
one,
but
in
kind
you
would
definitely
have
a
linux
one,
because
in
kind
you
run
the
you
run
the
api
server
and
xcd
in
the
container.
Okay,
that's
an
important
distinction.
By
the
way
I
don't
know
did
you
know
that
that
you
it's
normal
to
run
the
api
server
and
the
scheduler
and
the
kcm
inside
of
a
container
okay?
F
So
in
kind
not
only
so
you
do
that.
Not
only
do
you
do
that,
but
you
also
run
you.
You
have
a
docker
container
and
the
docker
container
is
essentially
like
a
vm
and
inside
that
docker
container,
you
run
multiple
docker
containers
right
and
those
docker
containers
are
implementing
kubernetes.
So
you
have
a
docker
container
inside
of
that
docker
container
you're
running
a
container
for
the
api
server
you're
running
a
container
for
the
scheduler
you're
running
a
container
for
the
kubernetes
control
plane,
whatever
right.
F
You
see
what
I
mean.
So
what
the
thing
that
the
thing
that
we're
doing
here
is
this
is
kind
of
your
project.
Is
here
right
instructions
for
this
right.
So
what
we
want
to
do
is
add
instructions
for
how
you
would
add
a
windows
kubernetes
right,
so
the
steps
are
going
to
be
right.
You
do
kind
create
cluster
over
overall
steps
kind
create
cluster
to
install
a
cni
provider.
F
Okay,
you
can
install
calico
or
andrea
as
a
cni
provider
on
that
cali
on
that
kind.
Cluster
I'll
show
you
how
to
do
that
in
a
second
and
then,
after
that,
you're
gonna
do
create
a
windows.
Vm
kubernetes
join
that
vm.
F
Or
install
the
kubelet
dot,
exe
and
other
windows
stuff
on
it,
who
radium
joined
that
windows
vm
to
the
kind
cluster
right?
So
you
could
basically
do
this
and
then
for
bonus.
You
could
like
wrap
this
all
in
a
you
know.
You
could
wrap
this
all
in
like
a
terraform
or
a
vagrant
recipe,
or
something
like
that
right.
Some
kind
of
infrastructure,
automation
recipe,
so
people
could
just
run
one
command
and
then
have
a
have
all
this
stuff
running
like
locally
if
they
wanted
to.
C
There
are
some
vagrant
scripts
that
patrick
was
working
on
where
he
left
about
a
year
ago,
they're
probably
fairly
reusable.
F
Okay,
cool!
What
that's
where,
where
do
those
live.
F
Yeah,
so
so
so
I
feel,
like
you
know,
once
you
get
this
working
like
manually,
you
know
frederick.
The
next
thing
you
could
do
is
this
link
to
james
is
going
to
give
us.
Maybe
there's
a
way
you
could
totally
automate
the
whole
thing.
Have
you
ever
used
any
of
those
like
vm
automation,
type
tools
before
no?
No,
okay?
Well,
they're
real
easy
to
use
they
just
you
download,
like
vagrant,
is
a
common
one.
F
F
You
know
what
I
mean,
but
I
mean
I
wouldn't
worry
about
the
bonus
steps,
because
if
you
just
get
this
five
things
working,
you
know
this
is
a
pretty
pretty
hefty
onboarding
project,
like
you'll,
probably
spend
several
weeks
just
getting
this
stuff
working
because
installing
a
cni
provider
in
a
vm,
you
know
that
works
with
container
d
and
all
that
stuff
is
non-trivial.
F
You
know
this
is
all
a
lot
of
this
stuff
is
sort
of
cutting
edge,
like
people
aren't
really
running
container
d
with
as
a
net
as
with
windows
very
commonly
kind
of
yet
you
know
what
I
mean:
you'll
you'll,
be
cutting
your
teeth
on
a
lot
of
interesting
issues
along
the
way
yeah.
So
that's,
why
kind
pretty
yeah?
F
That's
why
create
fine
cluster
is
going
to
be
fun
for
you,
because
that's
going
to
be
the
one
thing
that
actually
works
right
so
that
you're
going
to
be
able
to
create
a
kind
cluster
spin
up,
pods
do
stuff
with
it.
You
know,
and
and
that's
kind
of
the
starting
point,
that's
kind
of
the
anchor
for
all
this
stuff.
I
think.
C
I
see
somebody
else
just
joined
marcus.
Were
you
here
to
to
get
on
board
a
little
bit
too.
E
F
Okay,
so
I
think
yeah,
so
that's
a
starting
point,
so
this
is
like.
I
know
this
is
a
really
rough
starting
point
frederick,
but
I
think
I
think,
with
a
little
bit
of
googling
around
you'll,
be
able
to
get
some
of
this
stuff
sort
of
sort
of
hobble.
Your
way
towards
some
of
these
initial
steps,
because
once
you
read
the
docs
on
client
you'll
see
they
have
really
good
docs
and
they
show
you
how
to
spin
up
a
cluster
and
everything
else.
F
The
one
thing
that's
kind
of
since
I'm
telling
you
specifically
to
to
do
this
with
with
with
a
c.
So
normally
when
you
spin
up
a
kind
cluster,
it
doesn't
provide
you
like
a
real
cni
like
a
calico
or
andrea
cni.
Instead,
it
just
provides
you
like
a
cni
that
just
works
on
one
node,
so
I
have
a
script
that
you
can
use
here.
If
you
go
to
this,
if
you
have
like
a
mac
or
something
you
can
run
this
script,
I'll
put
the
link
in
here,
but
it's
to
achieve.
F
G
I
know
what
it
is,
but
I
don't
know
really
how
it
works.
It's
a
container
network
interface,
yeah
and.
F
Yeah,
if
you
create
a
pod,
the
cni's
job
is
to
plug
an
ip
address
into
the
pod.
That's
it
right:
okay,
plug
in
an
ethernet
cable
into
the
container
right.
Okay,
that's
what
it
is,
and
it's
like
so
there's
kind
of
three
real
popular
ones:
calico
interesting.
The
clouds
like
microsoft,
like
azure,
has
its
own
cni
for
cloud.
F
You
know
container,
like
you
know,
google
has
their
own.
Aws
has
everybody's
got
their
own
cni.
You
know
what
I
mean
and
they
have
different
fancy
features,
but
at
the
end
of
the
day,
the
core
thing
is.
It
gives
a
network
to
a
container.
So
the
container
can
talk
to
other
containers
in
the
cluster
and
talk
to
the
outside
world.
Okay,
so
so
this
script,
if
you
run
it,
you
can
just
run
one
of
these
commands.
You
know
what
I
mean
and
it
will.
F
But
that
said,
you
should
try
to
understand
what
this
script
is
doing
a
little
bit
if
you
use
it,
because
what
it's
doing
is
it's
creating
a
yaml
configuration
for
kind
and
that
yaml
configuration
is
going
to
result
in
three
docker
containers:
okay,
one:
that's
simulating
a
control,
a
kubernetes
control,
plane
and
two
that
are
simulating
kubelets
right,
just
like
raw
kubelets
right,
and
so
we
create
these
like
configurations
and
then
all
we
do
is
you
do
kind,
create
cluster,
and
then
you
give
it
that
yaml
file.
F
You
see
that
okay,
so
we
just
give
it
that
yaml
file,
and
then
it
uses
this
topology
to
create
this
cluster
okay
and
then
afterwards
it
does
a
couple
of
little
hacks
that
allow
docker
and
docker
networking
to
work
by.
You
know
disabling
xdp
and
stuff
like
that,
but
you
can
kind
of
ignore
those.
The
thing
that
actually
matters
is
that
it
it.
This
is
the
thing
that
matters
it's
called
calico
now.
F
The
thing
to
realize
here
is
that
when
you
install
calico
you're,
actually
installing
containers
on
your
on
your
cluster,
whose
job
is
to
give
ip
addresses
to
other
containers,
so
that
introduces
you
to
this
concept
of
a
host
network.
Okay,
some
containers
can
run
as
most
network
pods.
Other
containers
can
run
as
and
I'll
just
let
me
share
it.
I'll
show
this
like
sort
of
concretely
like.
E
F
K
F
F
Okay,
so
if
I
go
coupe
ctl
edit
pod-
and
I
it's
just
a
way
of
looking
at
the
contents
of
a
pod
right
and
I
say-
dash
name
space,
coupe
system,
it's
going
to
pop
up
this
coupe
controller
pod
right
and
if
I
go
to
image
you'll,
see
it's
running
the
you'll
see
that
it's
it's
running,
this
calico
cube
controller
image.
F
F
F
Host,
no,
would
you
see
this
host
network?
True
free,
yes,
see
that.
So
what
these
these
node
containers
right,
they
come
up
and
they
are
you're
not
going
to
understand
all
this,
because
I
know
you're
new
to
this.
But
this
whole
concept
right
is
yeah.
It's
like
it's.
These
containers
come
up
on
the
host
network,
they're,
not
using
a
cni.
F
The
reason
they're
doing.
That
is
because
there
is
no
cni
at
the
point
that
these
containers
are
installed
right.
So
watch
this
caught
kind,
get
clusters
kind,
delete,
cluster
dash,
dash
name
equals
calico.
So
I'm
going
to
tear
this
cluster
apart
right
and
I'm
going
to
rebuild
it.
Okay,
cdks
prototypes,
okay,
see
kind
ls,
okay
and.
K
F
F
F
So
this
is
a
watch.
This
is
just
running
in
a
watch
here.
So
now
it's
starting
it's
it's.
It
hasn't
written
a
cube.
F
F
So
the
reason
these
are
init
is
that
what's
happening
is
these?
Are
these
aren't
up?
Yet
these
have
some
initial
containers
that
run
before
the
actual
calico
node
containers
run
and
then
eventually,
these
calico
node
containers
will
start
once
that
happens,
all
three
of
our
nodes
will
have
a
network
got
it
and
as
soon
as
that
happens,
then
you'll
see
these
other
containers
come
online,
core
dns
and
coop
controllers,
because
those
actually
need
a
network
to
start,
you
see
what
I
mean
so
like
all
these
containers.
F
D
F
It's
a
very
critical
thing:
okay,
yeah
these
are
host
network.
These
are
host
network,
all
the
ones
that
came
up
after
calico
node.
Those
are
all
regular
ones
right.
So
now
I
can
do
coop
ctl
get
pods
dash
oh
wide,
and
I
can
actually
look
at
what
this
actually
means,
because
you
see
this
right.
The
reason
this
networking
stuff
is
important
is
because
that's
like
always
the
hardest
part
of
any
windows
cluster
is
getting
the
networking
right,
because
it's
totally
different
than
the
way
it
is
in
linux.
F
You
know
so
at
least
understanding
it
on
linux
is
a
starting
point.
You
know
so
so
you
see
these
different
ip
address
ranges
how
some
of
these
are
192's
and
the
others
are
172's,
so
the
172's
you
may,
if
you've
used
docker
a
lot
you
you
recognize
that
right,
that's
the
I
get
rights
of
the
docker
container
right.
So
if
I
do
docker
ps
you'll
see
here
that
I
have
these
four
containers
running
here
right.
F
Each
one
of
these
is
like
a
kubernetes
node
in
and
of
itself,
as
as
I
mentioned
earlier
right,
so
we've
got
like
so
so
so
we've
got
four
ip
addresses
1.2.5.4.3.2
right,
two,
two,
two
three
four
five
right,
four
different
ip
addresses,
172.18.0.3.172.18.
F
Right,
okay,
so
we
got
four
four
four
different
ip
addresses,
four
different
nodes
and
every
other
container
we
make
is
going
to
be
on
the
pod
network,
which
is
in
the
192
subnet.
Okay.
So
now
I
can
do
something
like
this.
I
can
say
group
ctl
scale
deployment
core
dns
right.
I
can
scale
this
deployment
and
now
friedrich.
Do
you
want
to
guess
what
subnet
that
core
dns
is
going
to
be
in
the
one
one
on
two
192.
see
you've
got
it
you're
ready
to
go
one
day,
30
minutes?
F
Okay?
Here
we
go
dash
dash,
replicas,
equal
equal,
like
say
four,
let's
say:
let's
say
like
five:
what
did
I
do
wrong?
Cubes
scale
deployment,
coordinate
oops.
I
did
the
pod
scale
deployment
core
dns
cool,
okay.
So
now,
if
I
say,
kubectl
get
pods
dash
a
show
wide
right
now,
you're
going
to
see
a
lot
of
these
core
dns
containers
of
see
how
we
have
all
these
core
dns
containers
now
right,
so
nc
and
they're
all
in
the
192.
F
F
If
I'm
going
too
fast,
coop
ctl
edit
deployment
or
if
you
have
a
question
cord
core
dns
dash
n
system,
but
I
can
go
in
here
and
I
could
do
something
weird
like,
for
example,
I
can
say
you
know
here's
the
pod
spec
right,
I
could
go
in
here
and
I
could
say
something
weird
like
I
could
say:
host
network
equals.
True.
I
have
no
idea
what
this
is
going
to
do
now:
right,
okay,
so
now,
coupe
ctl
get
pods
dash
a
right
dash
shell,
wide
okay.
F
Now,
let's
see
what
what
what
kind
of
nonsense
we
did
here
so
now
now
you
can
see
that
let's
run
it
one
more
time.
F
So
now
we
can
see
that
nothing
is
running
well.
Actually,
no
we've
got
one
two,
so
these
are
all
running
and
they're
all
attached.
So
we
have
four
of
these
running.
We
wanted
to
have
five,
but
we
only
have
four
running.
The
reason
is
because
it's
on
the
host
network-
and
so
because
it's
on
the
host
network,
you
know,
whatever
port,
it's
binding
to
there's
only
one
of
those
free
ports
on
each
node
right.
So
you
only
have
now
we
can
only
have
like
you
know
we
used
to
have
like.
F
Maybe
we
could
have
10
core
dns
containers
running
if
we
wanted,
because
the
pod
network
has
an
infinite
number
of
ip
addresses
right
or
I
mean
you
know,
it's
got
hundreds
of
ip
addresses,
but
the
host
network
only
has
one
ip
address,
so
that
means
you
can
only
run
one
core
dns
container
per
node
here.
Okay,
so
that's
so
that's
like
that.
F
F
So
now
it's
gonna
go
back
to
running
these
core
dns
containers
in
the
pod
network.
Okay,
so
anytime,
you
want
to
do
a
networking
experiment.
Core
dns
containers
are
the
easiest
to
do
it
with
because
they're
already
sitting
there
on
your
cluster,
because
every
kubernetes
cluster
typically
comes
with
the
these.
F
Well,
normally
they
come
with
these
internal
dns
containers
right
or
if
you're
on
a
cloud
they
may
provide
dns
through
some
other
means,
but
if
in
kind
at
least
you're
running
these
core
dns
containers
they're
running
in
the
container
network,
you
can
scale
them
up
and
scale
them
down
and
play
with
things
that
way.
F
I
think
this
is
like
a
decent
start
introduction
to
getting
client
up
and
running.
One
other
thing
is,
if
you
saw
I'll,
show
you
how
to
get
into
these
vms.
F
So
if
I
do
docker
ps
and
you
use
you've
used,
I
assume
docker
to
some
to
some
extent
right
yep.
So
if
I
do
docker
exec
dash
t
dash,
I
here
bin
bash,
I
can
create
a
bash
shell
on
here
right
now
that
I've
got
a
bash
shell.
I
can
do
a
cube
kube
adm.
F
What
is
it
print
join
token?
I
think
kubay.
So
kube
adm
is
the
thing
that
you
use
to
install
these
things.
These
clusters,
right
so
kube
adm,
is
like
a
binary
that
it's
like
a
swiss
army
knife
for
creating
a
coupe
cluster
right,
so
kind
under
the
hood
is
just
running
these
cube
adm
commands.
So
I
can
say,
like
cube
adm
I
can.
I
can
run
this
command
and
this
is
the
trick
for
windows.
Okay,
so
I
can
do
kubernetes.
F
How
do
you
print
a
joint
token?
Does
anybody
know
how
to
print
a
joint
to
cover
dm
print
join
token?
I'm
gonna,
google,
it
yeah
print,
oh
yeah,
here
it
is
print,
join
command.
F
Adm
token
help
generate
maybe
print
join
command.
F
Cube
adm
token
create
dash
print
join
command.
That's
it.
F
F
F
Oh
here,
it
is
yeah
see
that
you
see
this
command
here,
frederick,
okay,
so
this
is
it
so
this
is
kind
of
what
you're
going
to
be
doing
right,
so
you're
going
to
jump
in
here
and
then
you're
going
to
get
this
create
join
token
thing
I
and
the
tricky
thing
is:
I
don't
know
how
well
this
is
going
to
work
because
see
you're
joined
into
this
calico
control
plane,
but
that
calico
control
plane
is
not
like
an
ip
address.
It's
not
addressable
right.
F
It's
like
a
so
you're
gonna
have
to
like
somehow
when
you
run
this
you're
gonna
have
to
like
look
dig
into
whether
or
not
I
don't
think
this
works
with
an
ip
address,
because
it's
like
done
over
tls
or
something
you
know
what
I
mean,
but
you'll
have
to
like
see
how
to
like,
replace
that
like
you'd
have
to
like
do
something
like
you
know,
get
the
ip
address
of
the
of
that
container
right.
So,
but
that's
not
a
hard
thing
to
do.
F
You
just
do
ipa
grep
one,
seven,
okay,
so
my
ip
address
is.
Is
that
right?
Oh
my
gosh.
What
did
I
just
do?
I
just
joined
the
notes
itself.
I
have
no
idea
what
I
just
did
so
I
just
ran
like
I
just
ran.
I
joined
that
note
to
itself
okay,
so
it's
172
dot,
okay,
get
out
of
here
right
and
then
I
can
get
that
join
token
here,
which
I
didn't
mean
to
do
that
either
so,
okay.
So
if
I
do
this
so
you're
gonna
have
to
do
something
like
this
right.
F
F
Okay,
but
I
don't
think
you
can
run
kuba
dm
join
against
an
ip
address.
I
think
you
actually
have
to
run
it
against
a
host
name,
so
you
may
have
to
modify
windows
probably
has
some
concept
of
etsy
hosts
and
I'll
I'll.
Add
this
note
to
that
issue
that
we
created
just
now,
but
which
is
here.
F
And
then
that
command
we
ran
was
print
join
token
join
command
right
that
yeah,
okay,.
F
So
yeah,
so
that
would
be
roughly
the
first
steps
right
and
then
so
like
this.
So
then
you
would
run
this
join.
You
would
run
this
join
command
like
in
in
a
different
vm.
Now
right,
so
you
spin
up
some
other
vm
and
then
you
try
to
join
that
vm
to
this
cluster
and
that
vm
would
be
like
to
start
just
a
linux
vm
just
to
see
if
this
works
from
linux
to
linux.
G
F
Yeah
yeah
re-watch
the
video
look
at
the
issue.
The
issue
is
sick
windows,
test
sig
windows
tools,
145
and
I'm
going
to
ccu.
Frid
friedrich.
Don't
know,
I
guess
you
can
I'll
just
put
it
in
the
upstream
slack
kubernetes
slack.
Okay,.
F
Free
drake,
okay,
there
we
go
so
I
ccg
you
in
there
and
for
anybody
else
that
happens
to
be
watching
this
video.
If
you're
interested
in
participating
in
these
these
little
sessions
with
us,
you
know
we'll
be
doing
these
every
week,
so
friedrich
yeah
get
started
and
then,
if
you
wanted
to
like
pair
again
like
one-on-one
with
me
sometime
this
week
or
something
I
might
be
able
to
do
it
like
possibly
friday
or
something
or
I
may
I
can.
F
I
can
do
a
quick
sync
up
like
tomorrow
or
the
day
after,
if
you're
blocked
on
something.
But
I
think
the
best
thing
to
do
is
try
to
ask
questions
in
the
upstream
k-8
windows,
slack
room
that
way
me
or
perry,
or
anybody
else
can
also
help
you
and
but
let's,
let's
definitely
get
a
hop
on
a
zoom
call
again
in
a
in
a
few
days
whenever
we
need
to.
G
Okay,
so
my
problem
at
the
moment
is,
I
have
this
super
old
notebook
and
I
have
no
idea
if
I
can
make
this
work,
but
my
new
notebook
is
ordered
already
and
then
cool
cool
yeah.
As
soon
as
it
arrives,
I
will
definitely
be
ready
to
start.
F
You
also,
if
you
yeah,
I
mean
I'm
not
saying
you
have
to
start
tomorrow,
but
I'm
just
saying
that
if
you,
if
you
wanted,
you
could
get
a
free
cloud
account
like
in
google
or
azure,
or
something
potentially
to
hack
on
a
vm
there.
If
you
wanted
to
just
do
it
on
linux,
vm,
just
just
fyi,
okay,
because
they
you've
never
like
if
you've
never
spun
up
a
vm
there.
Before
that,
I
think
they'll.
F
J
Basically,
there's
a
website
katakura.
Basically
it
spins.
J
F
J
It's
like
a
mini
guide
series
to
get
you
knowledgeable
about
how
to
use
kubernetes
and
so
on
and
so
forth,
starting
from
a
small
deployment.
F
J
F
F
J
And
regarding
you
know,
you
had
edited
the
contributing
document
that
it
was
missing
the
running
test.
Part,
if
you
remember
I'm
wondering
if
that
was
actually
intentional,
because
you
know
you
don't
you
you
cannot
have.
You
cannot
have
failing
tests
if
you
do
not
run
any
tests
right.
F
I
I
feel,
like
just
people
need
to
know
that
there's
an
end-to-end
binary
that
comes
with
that
you
can
build
to
test
kube.
I
feel,
like
that's
like
a
first
class
piece
of
information
that
you
need
to
know
if
you're
getting
involved
in
development,
but
I
don't.
I
agree.
J
J
F
F
Contributing
sig
windows,
contributor
guide
I'll
make
a
thread
in
slack
called
you
know
for
you,
frederick,
where
we
talk
about
that.
F
Actually,
you
know
what
we'll
do
it
all
at
it
in
that
issue.
What
was
that
issue
that
we
have.