►
From YouTube: Kubernetes SIG Windows 20180206
Description
Kubernetes SIG Windows 20180206
A
A
So
so
the
first
PR
was
a
hyper-v
isolation,
I
believe
I,
don't
know
if
that
converged
or
not,
but
I
know
for
sure
he
could
have
proved
all
the
way
to
the
end,
and
the
second
was
a
PR
for
the
CRI,
so
that
one
is
under
review
right
now.
I
believe
Bob
is
reviewing
a
change.
He
said
on
the
surface,
it
looks
good,
but
he
has
to
use
a
deep
dive.
A
B
Clarify
on
those
real
quick,
so
last
week
in
a
sig
note
meeting,
they
agreed
to
review
the
future
proposal
for
the
resource
cutter
governance,
and
so
they
approved
that
proposal,
and
so
and
so
now
the
peg
fee
has
the
implementation
of
that
awaiting
review.
So
I'd
say
that
one's
on
track
and
the
hyper-v
isolation
one
was
merged.
A
C
B
To
clarify
on
that
a
bit,
so
the
CB
means
the
update
that
comes
out
in
the
second
Tuesday
of
each
month.
Those
ones
are
marked
political
and
that
contains
all
the
security
fixes
and
then
there's
an
optional
update
that
comes
out
each
month.
Then
I'll
be
denoted
with
this
see
that
that
one
has
feature
changes
in
it,
but
those
feature
changes
are
automatically
rolled
into
the
next
month's
B.
B
B
B
So
there
were
two
fixes:
one
of
them
was
the
service
IP.
One
I
think
we
were
talking
about
that
one
on
the
slack
Channel
a
few
days
ago.
The
windows
node
cannot
access
services
using
the
service
IP
and
the
other
one
was
the
the
other
DNS
failure
where
sometimes
the
DNS
server
would
stopped
getting
set
in
the
container
after
had
been
running
for
a
while
and
all
DNS
queries
were
failing,
cool.
D
E
B
So
those
might
work
in
the
current
windows.
Insider
preview
builds
so
we
changed
the
file
system
filter.
That's
there
in
the
end
for
the
next
release
of
Windows
Server
and
that
fixes
symlinks.
So
it
should
be
testable
there.
But
I
haven't
been
able
to
confirm
that
that
specific
feature
yet
I've
been
trying
to
get
some
better
instructions
on
how
you
can
get
those
builds
up
and
running
on
reserve
because
nobody's
published
how
to
do
that,
and
so
I
can
post
a
gist
in
the
channel,
but
I'm
still
having
a
a
couple
issues
with
it.
B
E
A
A
B
A
A
E
A
A
couple
of
other
things
that
happened:
I
the
folks
from
from
Ticketmaster.
They
made
some
updates
to
our
getting
started
guide.
Those
were
approved
on
Friday
I,
don't
know
if
they
commercial,
not
but
I,
know
you
know
we're
continuing
to
improve
our
documentation
and
make
updates
to
that.
So
thank
you
guys
for
for
working
on
that
and
I
know.
Just
you
were
also
working
on
that
as
well.
So
I
didn't
mean
to
exclude
you
from
that
list.
I.
A
Excellent
I
think
I
think
that's.
Is
there
anything
else?
Anybody
wants
to
bring
up
any
questions.
Any
concerns
I
think
it's
great
that
you
got
the
go-ahead
from
signaled,
so
we
can
continue
moving
forward
with
our
with
the
CRI
changes,
passing
continuing
resources.
I
think
you
know
the
hyper-v
isolation.
Guanyin
was
excellent.
I
know
there's
more
happening
on
CNI
plugins
around
the
X
lon,
but
maybe
looking
at
an
uptick
about
next
week
and
I
know
the
end-to-end
tests
that
Patrick
and
his
team
are
talking
to
cloud-based
forward.
E
So
down
the
same
line
on
the
instructions
that
those
are
very
helpful
I
wonder
if
we
could
add
more
details
like,
for
example,
the
add
route
or
the
star
couplet
I,
think
for
the
even
for
the
start,
kakashi
all
those
they're
bunch
of
PowerShell
scripts.
Then
there
are
a
lot
of
details
about
HMS
and
well
I'm,
trying
out
intended
like
hit
issues
and
don't
know
how
to
troubleshoot
I.
D
E
D
E
I've
seen
like
pardon
my
mapping
for
the
CI
configuration
or,
like
all
the
arguments
around
the
policies,
how
they
route
or
they
are
the
things
really
like
well
I'm,
trying
to
troubleshooting
I,
don't
know
whether
I'm
having
the
cracked
setting
like
say
for
Linux.
You
have
IP
table
given
a
service.
I
know
whether
it
has
the
crack
and
point
or
not,
but
for
HSS
I'm
right
now,
I'm
hating
on
usually
I,
don't
even
know.
E
F
F
E
So
the
usual
I'm
heating
is
now
I
have
a
cluster
with
the
Linux
master
and
two
windows
nodes
and
then
the
particle
part
of
a
build,
our
own
scenic
plugin.
Just
following
the
community's
approach
and
then
the
note
parts
to
other
cross
nodes
can
communicate
correctly.
The
ischium
hating
is,
after
I
deploy
a
service
with
node
port.
E
A
Yeah
I
think
we
resolving
this
in
their
slight
channels.
That'd
be
great,
you
know,
make
sure
you
create
a
thread
that
way
it
doesn't
get
polluted.
Actually,
one
of
the
things
I've
seen
in
the
selection
on
that
makes
it
it'll
be
harder
is
if
we
are
starting
a
conversation,
use
the
threading
in
slack.
If
you
can't
to
just
keep
all.
A
About
a
single
conversation
in
the
same
same
place,
my
making
easier
for
people
to
vote,
contribute
and
help
and
as
well
as
kind
of
finding
answers
on
a
single
topic,
the
I
think
that
there's
a
common
trail
on
the
chunks
on
the
zoom
chat
right
now
around
the
five
seven
one,
two
four
change,
and
that
was
a
change
for
for
VMware
fusion,
I
approved
it
a
couple
of
times
and
I.
Think
teams
can
wear
how
to
obtain
it
to
make
it
happen.
A
G
A
F
A
D
Around
the
put
it
in
slack
beforehand,
but
the
couple
issues
that
we
have
around
metrics
and
missing
resources,
either
generally
or
for
Hawes,
horizontal
pot,
autoscaler
I,
think
the
Bob's,
the
best
like
the
most
knowledgeable
or
appropriate
person
to
talk
to,
but
I
wanted
to
get
details
more
details
about
that
flushed
out
of
like
what
are
we
missing,
etc
and
want
to
check
if
there's
more
people
besides
Bob,
that
would
be
good
for
that
discussion.
I
think.
A
Bob's,
the
right
guy
so
I
feel
free
to
ping
him,
and
you
guys
can
start
that
conversation
another
in
the
past.
We've
had
some
difficulty
getting
the
horizontal
pod
autoscaler
to
work.
I,
don't
know
the
new
changes
in
CRI
is
that's
gonna,
make
it
easier
or
not?
I
certainly
hope
so,
but
let's
flush
it
out
and
see.
If
we
can,
if
you
can
identify
any
issues
or
happening
I
know,
Bob
was
also
talking
to
what
wasn't
jungly.
That
was
hopping
in
that.