►
From YouTube: Kubernetes SIG Windows 20170711
Description
Kubernetes SIG Windows 20170711
A
We
have
three
folks
right
now:
they're
working
on
the
sig
I,
don't
see
any
of
them
online
right
now,
but
hopefully
they
are
actually
working
right
now.
So
that's
that's
a
good
thing,
but
to
have
a
pub
and
generally
from
aprender
they're,
both
they're,
both
working
on
getting
us
to
beta.
More
specifically,
generally
right
now
is
getting
her
feet
wet
on
the
seek,
and
then
she
sees
a
new
resource
that
got
added
and
she's.
A
Looking
into
hoping
well
put
some
of
the
networking
specific
artifacts,
as
well
as
she's,
going
to
look
into
some
secret
config
maps
and
volumes
on
Windows.
There
were
a
couple
of
folks
that
complained:
the
sig
on
the
slide
channel,
that
certain
items
like
volumes
were
not
very
well
tested
and-
and
that's
right
since
there
wasn't
a
supported
item
so
far,
so
we're
working
on
on
on
fixing
that
and
then
Microsoft
has
added
a
resource
to
seek
Windows
from
Anthony
house
team
and
they're,
essentially
working
right
now
on
on
CRI.
A
So
there
are
responsibilities
looking
into
all
the
work
that
they
did.
You
see
advisor
on
the
Linux
world
and
how
you
capture
metrics
from
docker
and
replicating
the
same
experience
for
sig
windows,
so
that
auto
scaling
and
other
tools
that
depend
on
perf
metrics
coming
out
of
containers
are
able
to
leverage
that
so
it's
an
intern,
that's
working
on
this
and
obviously
oh,
very
excited
to
to
have
more
contributions
to
the
project.
A
A
A
Not
just
questions
but
any
other
studies.
Anybody
else
wants
to
bring
and
talk
to
the
team
I
forgot
to
mention
that
also
Bob
is
talking
to
Alan
and
Eileen
as
well,
who,
both
of
them
from
cloud-based
into
some
of
the
networking
aspects.
So
there
are
they're
all
brainstorming
and
looking
into
what
solutions
can
I
do
to
implement
the
pod
abstraction
in
Windows?
There
were
some
concerns
around
functionalism
may
not
exist
on
Windows,
yet
at
least
Windows
Server
containers
and
and
Eileen
and
his
team
are
following
up
with
Microsoft.
Is
there
been
a
Piercy
on
that.
B
So
I
was
like,
as
you
can
cover
for
what
this
in
Windows
slot
channel
like
I've,
been
trying
to
I
guess
duplicate
the
steps
in
terms
of
getting
cluster
up
and
running.
So
what
I
see
is
that
I
can
get
the
cluster
up
and
running.
My
annoyance
to
me
is
that
a
service
IPS
looks
a
while
on
the
different
host
I
can't
connect
to
the
pod.
Ip
I
can't
seem
to
connect
the
server's
IP
from
the
host.
B
Is
this
unexpected
behavior
I
think
they
work
within
the
pasta
seems
like
DNS
I'm
able
with
like
my
standard
example,
is
trying
to
use
the
DNS
service
IP
like
DNS
works
within
de
pod,
so
that
point
in
the
K
pad
of
our
server
site
keys
are
working
but
I
can't
test
it
by
trying
to
make
my
DNS
server
outside
of
the
pod,
be
the
server's
IP.
Additionally,
maybe
I'm
looking
at
it
wrong
as
well.
Like
I,
don't
see
the
Gateway
node
like
showing
my
node
ports,
but
maybe
like
again,
maybe
I'm
using
it
wrong.
B
How
how
am
I
supposed
to
use
the
Gateway
node
I
would
assume
I
just
connect
to
or
I
get
I
would
assume.
I
can
connect
the
local
host
node
port
in
a
work
I
tried
connecting
to
the
internal
IP
on
GCP
and
node
for
OSA
didn't
work
next.
That
didn't
show
anything
listening,
so
I
can
queue
proxy,
which
I
was
always
able
to
sort
of
nest
that
and
see
the
load
ports
I'm
just
wondering
if
anyone
can
give
me
insight
into
how
sort
of
these
two
things
are
supposed
to
work.
So.
C
For
the
beauty,
node
huge
is
solidified
like
internal
connectivity,
for
example,
within
a
pod,
we
communicate
Google
DNS
servers.
Okay,
if
that
works
on
the
Gateway
node
should
work
with
no
additional
configuring.
The
DNS
I
mean
the
service
IP.
If
you
say
it
was
from
within
the
board,
but
not
from
the
hole.
It's
maybe
you're
missing
a
role,
so
you
can
indicate
you
want
to
get
a
product
for
this
I
think
so.
B
C
B
B
Like
at
first
like
I
at
first
I,
try
I
did
a
part
of
the
whole
cluster
with
like
Linux
windows,
gateway
worker
nodes,
and
it
wasn't
working
well,
so
I,
a
scaled-down
Owen,
all
just
right.
The
limited
cobia,
no
BS
with
1x
worker
node,
oh
the
other
thing,
I
notice,
is
that
I
reboot
the
gateway
node.
It
seems
like
he
can
no
longer.
Oh
the
one
service
yeah.
A
So
so
that
we
fix
so
the
issue
were
there's
some
lack
of
persistence
after
you
start
booting.
Certain
things
is
something
that
Bob
has
fixed
I,
don't
know
if
you've
published
the
latest
code,
that
might
not
be
stable
yet
so
that's
probably
why
he
didn't
publish
it
but
I'm
for
certain
that
the
issue
will
reboot
certain
things
will
not
function.
A
In
the
same
way,
we
will
fix
that
of
the
other,
should
note
toward
not
being
thinkable
sure
I,
don't
know
if
that's
something
we
fixed
or
not,
when
you
start
a
thread
with
a
lean,
go
ahead
and
add
Bob
stead
check
to
this
wall,
it's
a
bob
BST
ECI
UK.
This
is
Elias
and
he
might
reach
out
to
you
as
well.
B
To
say
that
obtainable,
the
node
weight,
it's
like
and
I,
don't
like
I
guess
I
might
be
doing
something
wrong.
It's
like
I,
just
don't
like
I
was
just
trying
to
access
it
from
within
the
cluster
and
okay
I
would
have
sir
I
guess
expected
to
see
the
node
for
visible
and
next
step
again
doing
it.
That's
that,
like
listen,
a
mic,
PPP
listening
a
call
and
I
didn't
see,
I,
don't
a
kind
of
like
so
I
guess
but,
like
I,
think
I
think
I
can
ping.
C
B
Like
in
digging
into
the
gateway
stuff,
take
it
if
I
dump
the
flows
like
I,
see
a
flow
like
with
the
port
on
the
that
would
be
the
node
port.
I'm,
not
sure
is
like
everything
is
exactly
there
like,
like
I.
Don't
have
enough
at
this
point,
understanding
of
sort
of
OBS
Oviatt
and
how
it
all
works
to
my
face
or
things
for
sure,
but
it
come
up
very
much
a
newbie
and
trying
to
wrap
my
head
around.
They
know
there's
the
nomenclature
with
it.
We.
C
A
C
Good
food-
oh
no
ordinal,
help
I
think
it's
doable,
probably
some
word
on
Friday.
If
your
own
would.
A
Let
me
let
me
coordinate
with
Bob
and
and
find
the
time
I'm
actually
living
on
vacation
this
Friday,
so
so
I
might
not
be
the
best
person
to
host
this
so
but
I'll
have
Bob
kind
of
taken
initiative
or
for
the
month
of
them
off
to
drive
the
meeting
this
meeting
recorded
as
well
as
be
able
to
to
to
code
a
one-hour
primer
on
ovn
OVS.
So
let
me
talk
to
him
and
we'll
sing
together
all
feelers
from
spike.
If
you
don't
mind.
C
C
Coding
gawd
progress
with
hello,
vs,
an
OB
n
general.
We
got
two
points
of
one
release
certified
by
Microsoft
and
we
will
make
uninstaller
with
the
driver
signing
and
yet
probably
early
BC.
Probably
this
week,
I'm
also
working
on
a
few
blog
posts
on
how
to
create
a
quick
environment
on
obvious
Nov
M
with
OpenStack
and
kubernetes,
but
I
don't
know
when
those
will
be
published.
A
I
pestered
Alessandra
a
little
bit
about
it,
but
whenever
you
get
a
chance
that
that'd
be
great,
you
can
get
the
ball
rolling
earlier,
rather
than
later,
just
given
the
long
product
cycles
that
that
that
agrees
so
yeah
as
early
as
possible,
if
you
can
send
it
to
my
email
rather
than
slightly,
will
be
easier
for
you
to
convince
me
an
email
when
I'm
on
vacation
but
I
will
definitely
respond
and
provide
feedback.
And
then
you
can
talk
to
to
the
Redmond
folks
in
networking
about
that.
So.
C
D
Hello,
okay,
so
Kevin
we
okay,
some
progress
with
us
yennai
on
kubernetes
within
some
manual
tests,
and
we
managed
to
create
multiple
containers
in
the
sample.
We
are
now
looking
how
to
edit
the
code
for
it
in
kubernetes
and
to
finish
the
CNI
stuff,
probably
in
a
few
weeks,
we'll
have
the
full
concept.
This.
A
A
A
A
A
Cool
alright,
well,
everybody
have
a
great
rest
of
the
day,
we're
depending
on
your
time
zone
and,
like
I,
said
I'll,
be
gone
on
vacation,
but
we'll
have
someone
else
that
will
run
the
meetings.
Someone
else
from
aprender
unless
someone
from
here
wants
to
volunteer
they'll
record
them
again.
Like
always,
and
we
will
you'll
see
back
in
about
a
month,
I
will
still
be
active
on
on
slack
or
emails,
so
feel
free
to
reach
out
to
me.
If
you
need
anything,
cool
all
right
have
a
great
rest
of
your
day.
Thank
you
guys.