►
From YouTube: Kubernetes SIG Windows 20181120
Description
Kubernetes SIG Windows 20181120
A
Hello,
everybody
and
welcome
to
another
sig
Windows
meetup
I.
Thank
you
all
for
attending
I
know
that
this
is
franking
week,
so
were
likely
not
going
to
have
too
big
of
an
attendance.
But
thank
you
for
the
guys
that
put
the
time
to
join
us.
Patrick
is
on
vacation
today,
so
he's
not
able
to
join
us.
So
I
want
to
update
everybody
on
the
events
that
will
happen
in
the
last
week
or
so
is
there
anybody
that's
willing
to
take
some
notes
if
not
I'll
just
go
back
and
update
the
meeting
notes
later
on
myself.
A
I'll
take
some
notes.
Oh
thank
you
appreciate.
So
the
first
things
first,
so
we
Patrick
and
I
were
put
through
the
wringer.
Last
week
we've
had
multiple
meetings
with
seek
release,
seek
architecture
as
well
as
a
conformance
team
trying
to
get
approval
to
go
stable
with
kubernetes
1.13,
ultimately,
which
was
short
of
that
goal.
A
Ultimately,
even
though
that
was
a
big
blow
to
our
effort
to
bring
Windows
to
stable.
The
big
thing
here
is
that
it
will
give
us
a
little
bit
more
run
right
now
to
stabilize
both
our
tests
make
a
call
to
action
to
maybe
get
more
contributions
in
to
seek
windows
and
lay
out
a
really
good
foundational
plan
from
going
to
GA,
Patrick
and
I
are
going
to
take
ownership
of
that.
That
includes
dogs
includes
creating
a
cap
that
will
outline
the
clear-cut
details
for
going
to
GA.
A
We
will
surface
that
multiple
seeks,
for
example,
cigar
signal,
cheek
release,
cig
network
and
storage
to
get
up
whoville
on
that
cab.
Well,
ahead
of
time.
Our
goal
is
to
do
that
in
December.
Sometime
that
way,
we
have
the
full
approved
list
for
going
to
GA,
including
the
testing
all
wrapped
up
before
the
end
of
the
year,
and
potentially
the
new
year
would
allow
us
to
make
some
more
contributions
to
the
tu-114.
Maybe
some
more
features
now.
A
Now
I
know
this
is
not
great
news
for
anybody.
One
of
the
cost
of
actions
that
we'll
have
and
will
reiterate
down
the
next
Co.
Hopefully
you
have
more
attendance-
is
that
if
you
have
any
customers
that
have
a
direct
need
for
for
Windows
support
in
kubernetes,
if
some
of
those
customers
are
using
Windows
containers
today
and
they're
looking
for
an
Orchestrator,
we
want
you
guys
to
come
forward.
Peter
I'm.
Definitely
looking
at
you
at
this
point
in
time.
A
A
A
The
last
part
is,
you
know:
if
anybody's
has
time
and
can
teaching
on
the
test
contributions.
That
would
definitely
be
a
welcome
addition
to
our
team
and
I'm
asking
for
resources
from
my
organization
to
to
help
contribute
here,
but
essentially
we
want
to
make
sure
that,
whatever
you
can
do
all
our
tests,
we
need
to
get
them
to
green
as
early
as
possible
so
that
we
don't
get
into
the
same
situation
that
we
call
the
sick
architecture
and
seek
release
last
time.
A
C
I
wanted
to
follow
up
on
what
you
just
said
about
helping
with
the
testing
we're
at
a
point
where
we
have
clusters
we're
funding
clusters.
We
can
bring
them
up
automatically,
but
we're
still
having
some
connectivity
issues
and
we
are
really
struggling
to
understand.
What's
going
on
on
the
windows
nodes
and
and
what
what.
A
C
D
C
E
C
C
F
Documentation
on
how
networking
works
on
Windows,
so
last
week
we
talked
about
there
were
some
nolden
issues
on
both
Wednesday
and
IRL
to
bridge
and
win
bridge
DXE.
So
since
the
last
week
we've
been
working
on
fixing,
those
I
do
have
new
binaries
that
I
can
distribute
today,
I'm
just
validating
I'm,
doing
final
sign-off
right
now.
Actually
I
can
include
the
configuration
file
that
I
used
as
well
like
this,
you,
an
icon
just
for
the
time
being
and
distribute
those
minor
ease.
F
/
plugins
repo,
like
there's,
there's
a
lot
of
back
and
forth,
and
it
took
us
almost
a
year
to
merge
the
PRC,
because
some
features
were
missing
that
are
present
in
Linux,
like
ipv6
support
and
basically
the
bar
was
set
really
high
for
the
initial
PR
to
go
in,
and
there
was
also
some
additional
work
on
integration
tests
that
we
also
have
to
do
so
that
that
kind
of
made
it
difficult
for
us
to
release
plugins
in
a
maintainable
way.
So
yeah.
F
And
after
that,
if
you
still
facing
problems,
we
have
to
collect
logs
the
PSone
that
we
can
use.
If
you
could
run
that
on
any
windows
hosts
that
you
have
can
give
us
a
mapping
of
you
know
all
the
policies
that
are
applied
on
the
end
points
on
the
pods
and
we
can
troubleshoot
it
that
way
to
double-check
what
could
be
going
wrong.
E
G
E
Would
it
be
make
make
sense
to
maybe
do
like
a
session
and
like
record
it
on
youtube
or
something
so
that
we
can
all
kind
of
have
an
understanding?
How
to
do
this?
Like
I,
know
some
stuff,
but
I
don't
know
at
all.
It
sounds
like
we
all
can't
have
a
mix
of
stuff,
but
maybe
if
we
had
like
some
kind
of
video
to
say
hey,
this
is
how
you
I
need
to
go
debug
these
things.
These
are
the
steps,
maybe
be
helpful.
Yeah.
A
F
G
One
of
the
causes
that
we
ran
into
when
I
was
digging
through
this
and
ranch
I
ran
into
it
as
well.
Is
that
one
six
nine
addresses
are
automatically
removed
and
you
have
to
put
the
routes
back,
so
you
can
actually
get
AWS
discovery
and
GCP
discovery,
so
you
can
actually
discover
metadata
and
in
just
every
single
time
you
just
order
the
H&S
or
add
the
H&S.
It
just
disappears
and
you
have
to
click
script,
that
back
in
yeah.
That's
an
actual
HMS.
F
Book
at
the
route
sermon
removed
from
the
physical
adapter
on
top
of
wheat,
on
top
of
which
we
create
the
host
Phoenix,
that's
used
by
that's
HMS
issue.
So
that's
so
that's
on
us
to
troubleshoot
and
we
have
someone
assigned,
but
they
they're
they
can
only
get
to
it.
Late,
December
or
early
January.
That's!
F
F
F
So
yeah
I
mean
we
can
add
additional
documentation,
maybe
on
misty,
on
Docs,
for
this
specific
issue.
But
that
is
definitely
something
that
we
are
tracking
our
consolidated
ugly.
It's
just
that
we
don't
have
time
yet
to
get
to
it,
because
there's
other
issues
like
the
wind
bridge
that
we
want
to
figure
out
first
and
there's
also
overlay
coming
out
so
yeah
in
terms
of
Docs,
so
I
was
gonna,
make
a
blog
post
documenting
all
the
networking
options
and.
B
F
F
That
are
available
on
Windows
and
publish
that
if
there's
additional
troubleshooting,
like
maybe
a
video,
some
James,
you
said
meant,
you
mentioned
that
that
might
help
I
can
try
to
do
that.
I,
don't
know
how
much
time
I'll
have
for
that.
I
know:
I've,
updated,
darks
recent
leave
with
some
FAQ
style
kind
of
troubleshooting
pages
on
our
mr.yun
or
microsoft.com
can.
F
F
Certainly
I
can
expand
it
a
bit
more
I
just
wanted
to
fix
one
bridge
and
get
that
out
I'm
also
having
I'm
also
pushing
a
PR
out
to
remove
some
of
the
assumptions
from
the
scripts
like
we
have
an
assumption
on
the
network.
Adapter
like
it
has
to
be
called
Ethernet
or
otherwise.
It
won't
work.
So
those
kinds
of
assumptions
and
improvements
to
describe
symbols.
F
So
there's
there's
a
lot
happening
right
now.
I
want
to
close
up
on
the
initial
wind
bridge.
Release
maybe
make
a
repo
for
that
and
determine
how
we
can
push
out.
Binary
is
easier
and
document
that
better
and
then
after
that's
done,
I
can
push
push
out
review
the
FAQ
style
troubleshooting
section
there,
and
then
we
can
look
at
documenting,
see
and
I
can
fake,
maybe
as
part
of
the
a
new
repo
that
we
create
for
pushing
out
these
plugins
having
a
better
documentation
around
it.
A
Also
be
great
eventually,
if
we
can
migrate
these
dogs
out
of
MSDN
to
kubernetes
report.
That
way
we
can
I
I
know
that
some
of
them
are
not
only
specific
to
kubernetes.
Maybe
this
one
is,
though,
that
way,
other
contributors
can
keep
adding
on
them
over
time,
I
mean
that's
the
only
way
you
can
actually
scale
sure.
A
E
D
Actually
calls
that
that's
one
of
the
things
it
does
so
the
so
the
log
slurp
script
that
I
have
connects
to
a
Windows
node
gets
all
the
docker
logs
gets
all
the
container
logs
and
then
calls
the
networking
collect
log
scripts
and
puts
it
all
into
a
single
zip
and
so
typically
after
a
test
pass.
I
run
that
and
then
I
go
and
look
at
the
test
case.
That
failed
and
then
I
review
the
logs
around
the
same
time
frame
from
the
node
that
the
test
case
was
running
on.
D
D
D
So
another
topic
related
to
this-
that's
I
think
it's
actually
time
we
to
bring
it
up.
I
opened
a
PR
for
a
Windows
testing
repo,
so
there's
a
github
organization,
that's
kubernetes,
SIG's
and
they'll.
Let
SIG's
create
repos
for
whatever
projects
they
want
there
and
so
I
have
that
linked
in
the
notes
here.
But
one
of
the
things
that
we
need
to
do
is
basically
get
a
lazy
consensus
from
this
from
sig
windows
to
create
this
on
behalf
of
sig
windows.
D
So
what
I'm
proposing
is
creating
a
Windows
testing
repo
that
we
can
use
to
store
the
code
for
the
test
images
that
Adelina
steam
has
been
working
on.
Those
are
currently
in
a
non
kubernetes
repo
today,
but
we've
prepared
proposing
to
put
those
in
there
and
then
have
the
docs
on
how
to
run
a
test
pass
there.
So
that
way
everything
somebody
needs
to
set
up,
build
containers
to
validate
a
new
Windows
version,
or
you
know,
whatever
kind
of
test
pass.
D
A
A
But
by
the
way,
I
do
also
have
one
question
related
to
that.
I
know
that
we
had
some
test
cases
that
failed,
because
one
of
our
images
was
no
conformally.
Remember
that
run
SH
will
something
like
this
approach
alleviate
problems
like
that
when
we
have
like
a
solid
repo
that
can
service
our
images
as
well
as
execute
our
test
cases,
yeah.
D
Then
that's
one
of
the
goals
for
this
repo
is
to
make
sure
that
we've
got
all
that
code
there.
Where
people
on
this
sake
can
find
it,
then
we
can
use
the
same
review
and
approval
process,
because
prowl
will
be
hooked
up
on
that
repo
for
us,
so
we
can
maintain
all
of
those
images
just
the
same
way.
We
maintain
the
other
code.
That's
in
that
mini
kubernetes,
kubernetes
repo.
A
D
A
You
guys
will
get
emails
on
your
list.
One
of
it
is
under
testing
report,
as
well
as
the
customer
story
document
that
I'm
gonna
send
out
for
people
to
fill
in
with
their
customer
testimonials
or
customer
voice
cool.
That's
all
the
stuff
that
I
had
on
the
agenda.
You
partook
anything
else
or
a
delay.
Hi.
Sorry,
sorry.