►
From YouTube: Kubernetes SIG Windows 20190319
Description
Kubernetes SIG Windows 20190319
A
A
The
main
CN
CF
blog
post
for
1.15
is
also
completed
and
looks
good
to
me,
and
the
only
activity
that
you
have
left
from
CN
CF
perspective
is
the
blog
post
for
for
Windows.
That's
going
to
be
the
block
first
blog
post
that
will
go
out
in
April
talking
about
the
Wonder
14
release.
That
one
is
something
that
we're
gonna
work
on
likely
starting
next
week.
So.
A
The
rest
of
the
stuff
around
114,
we
got
our
last
PR
in
that
was
completed
and
indeed
the
work
and
the
shepherding
there.
There
was
a
storage
related
PR
to
basically
prevent
some
create
a
critical
section
around
the
SMB
shares
and
a
month
you
know
mountain
of
volumes,
so
that
went
in
yesterday.
So
he
made
the
thought
late.
What
which
is
today?
A
So
we
don't
need
to
cherry-pick
that
so
from
a
sick
window,
suspect
to
all
our
peers
are
completed
and
although
Adelina
and
when
we
get
to
them
and
claudio
they've,
been
creating
a
lot
of
PRS
to
improve
our
test
and
instability
so
I'll.
Let
them
to
chat
about
that
in
a
little
bit.
The
the
last
item
around
114
is
the
documentation,
so
we
had
three
main
themes
of
PRS
going
on
the
first
one
was
the
contributor
guide
that
would
merge
yesterday.
A
Well,
multiple
versions
of
it
have
merged.
But
yes,
there
was
the
last
one,
then
the
GMS
a
work
when
in
last
week
I
think
you
deep
and
Jeremy
and
and
John
and
then
the
main
PR
for
the
dogs
for
sick
windows
is
waiting
for
a
final
look.
Good
to
me
from
from
Jim-
and
that
will
go
in
hopefully
today
that
work
is
not
done,
there's
still
a
little
bit
of
cleanup
to
do,
but
it's
it's
in
very
good
shape
and
most
people
were
ok
to
approve
it
even
in
previous
versions
of
it.
A
B
B
B
Now,
the
way
our
clusters
work,
we
have
to
apply
a
paint
on
the
master
node,
so
we
do
that
they're,
an
extension,
long
story
short
to
the
cluster,
is
slow
enough.
The
pausing
cube
system
will
not
spawn
at
the
moment.
We
apply
that
thing,
so
we
may
end
up
in
a
state
where
the
node
is
tainted
and
the
pods
cannot
be
searched
scheduled
anymore
and
the
mas,
not
because
it's
it's
tainted
now
that
should
be
fixed
at
operations,
but
due
to
a
bug
in
a
case
engine
on
it
doesn't
have
the
right
colorations.
B
So
what
we
did
to
to
fix
this
in
the
meantime
is
we
made
a
in
the
extension
we
added
awake
so
basically
we're
writing
from
the
pollen
tube
system
to
be
up,
and
then
we
applied
the
thing
that's
gonna
be
temporary.
In
there
I
mean
that
will
be
in
the
extension
until
we
we
fix
it
in.
In
a
guy,
so
it's
a
workaround
if
you
will,
but
it
seems
to
have
fixed
the
the
classification
issues
we've
seen
over
the
weekend
other
than
that,
so
it
did.
A
C
B
B
Another
thing
we
did
they
would
experimenting
with
on
the
staging
job.
We
added
parallel
test
runs
with
a
new
version
of
a
Jersey
and
I.
This
will
hopefully,
if
it's
stable
enough
in
the
state
in
the
staging
job.
This
will
hopefully
give
us
a
chance
to
have
more
runs
because
it's
quicker
it's
much
weaker
than
the
sequential
runs
we're
doing
now,
sequential
tests.
B
So
that's
that's
to
keep
in
mind.
Oh
we've
had
a
lot
of
pr's
going
in
going
out,
but
the
they're,
not
major
they're
they've,
been
tweaks
to
the
testing
config,
so
staging
job,
adding
more
parallel
like
that.
So
not
nothing
major!
All.
A
D
So
this
is
like,
because
we
realize
right
now,
at
the
pod
security
context,
fields
and
as
part
of
the
pods
back
like
allow
privilege
is
the
most
common
one.
But
you
know
all
the
stuff
around
like
setting
Linux
security
policies,
app
armor
stuff
in
the
cap.
It's
mentioned
that
at
least
for
right
now,
with
the
pod
security
context,
we
don't
support
any
of
the
fields,
so
it's
just
kind
of
like
a
way
to
air
out.
If
the
users,
if
users
submit
anything
in
those
fields,
yeah.
A
So
so
I
at
the
high
level
I
agreed
it.
It
allows
us
to
fail
fast
if
the
users
are
very
clear
error,
why
this
is
no
working
and
allows
them
not
to
get
into
situation
where
they're,
creating
yeah
Mel
pod
specs.
That
essentially,
would
attempt
to
use
features.
I
know,
Claudio
had
some
concerns
around
that
around
conformance
test,
but
I
kind
of
agree,
I,
don't
know
if
it
was
you
or
someone
else.
I
mentioned
it.
A
A
D
Just
submitted
us
today
so
I
have
it
as
like:
a
whip
PR,
because
I
just
want
to
get
some
feedback
from
people
from
sig,
node
and
and
from
anyone
here
in
sick
windows
about
well.
If,
if
this
is
as
far
as
I
could
figure
out,
this
is
the
only
way
we
can
do
port
forwarding
where
we
have
to
it's
like
we
have
to
spawn
a
container.
D
Ns
enter,
is
a
way
from
the
house
that
you
can
run
a
process
inside
the
network
namespace,
but
I.
Don't
think
you
can
do
that
on
Windows
so
like
this
is
the
closest
that
I
think
we
can
do.
I
mean
it
works,
but
it
will
require
us
to
add
a
new
I
think
the
best
way
for
doing
this
would
be,
if
we
add
a
new
container
that,
like
just
like,
we
build
the
pause
container
on
Windows
and
have
it
always
like
pre
pulled
onto
the
machine.
If
we
have
like
a
port
forward
container.
D
A
E
A
E
The
thing
I'm
working
on
storage,
Google
side,
so
I
think
in
our
windows
dock
we
mentioned
the
sub
pass,
is
not
supported,
but
from
my
testing
it's
actually
passed
most
of
tests
except
a
cup
of
Testament
Y
is
related
to
read-only.
So
seems
like
the
read.
Only
option
is
not
working
well
and
gather
is
the
use
file
as
a
sub
pass,
but
largely
I
think
the
sub
has
featured
it's
like
available
like
we
can
support
the
surpassed
feature.
I,
don't
know
why,
like
someone
tested
before
there's
some
issue
about
it
or
not,.
A
That's
that's
interesting.
We
actually
did
have
someone
that
tested
it
and
I.
Don't
remember
who
he
was
right
now
to
be
honest
with
you,
but
you
had
found
the
volume
subpath
mounts
did
not
work.
You
could
only
mount
entire
volume
and,
more
importantly,
the
usurpers
volume
mounting
for
secrets
that
definitely
did
not
work.
F
A
E
A
F
G
A
A
H
H
F
C
D
E
E
E
F
One
thing
I
have
for
fastest
storage
is
that
flex
volume
is
basically
there
won't
be
any
new
features
and
they'll
be
focusing
on
CSI.
So
in
your
future,
that
means,
if
we
continue
using
flex
following,
we
may
be
losing
a
lot
of
benefits
and
we
won't
be
able
to
have
feature
parity
with
other
things.
So
this
is
something
to
consider.
Maybe
we
should
find
another
time
and
to
invite
all
go
to
the
xxo.
A
All
right,
so
so,
let's
file
the
issue
and
we'll
follow
up.
It
looks
like
there's
gonna,
be
a
long.
Two
discussion
on
this
won't
be
as
simple
as
deciding
that
all
right.
A
couple
of
additional
things.
I
forgot
about
this,
but
we're
gonna,
have
a
cig
face-to-face,
a
meeting
for
Kubik
on
so,
if
anybody's
attending
Barcelona
I
reserve
some
time
on
Monday,
May,
20th
I,
don't
know
the
exact
time
yet,
but
when
the
organizers,
let
us
know
well
I'll
talk
about
it
here
in
this
meeting
and
advertising
as
well.
A
So
if
you
guys
are
attending
who
both
have
a
deep
dive
on
Windows,
it's
that's
more
for
end-users,
but
we
do
want
you
guys
to
attend
as
well
and
we'll
also
have
a
cig
members
face
to
face.
The
last
item
I
have
here
on
the
agenda,
is
there's
a
new
API
review
process
that
cig
architecture
is
following.
It's
in
the
notes.
There's
a
link,
I'm
also
pasting
it
here
in
chat,
so
anybody
that's
working
on
api's
or
that
will
introduce
new
api's.
A
More
specifically,
I
think
the
the
two
three
big
areas
that's
gonna
have
API
is
probably
going
to
be.
The
continuity
support
coming
from
Microsoft
the
GMs,
a
support
coming
from
deep
and
Jeremy
and
John,
so
you
guys
need
to
definitely
look
at
the
new,
the
new
review
process
and
if
you
do
have
any
questions,
Sagarika
teacher
is
available
for
answering
that
and,
and
and
and
hope
you
navigate.
That
also
me
and
Patrick
can
be
available
to
service
go-between
Savini
if
needed.
A
A
I
would
like
everyone
to
kind
of
look
at
the
issues
on
these
items
and
see
if
anybody
is
able
to
contribute
in
a
specific
area
so
that
we
can
start
slotting
folks
to
to
the
different
contributions
that
we
want
to
make
and
start
tracking
those
on
a
on
a
weekly
basis.
You
don't
have
to
decide
right
now,
but
start
thinking
about
where
you
may
be
able
to
put
in
a
little
bit
of
time.
A
F
A
So
the
release
notes,
don't
the
things
of
interest.
I
didn't
mention
that
when
GA,
but
in
there
PR
articles
for
for
Windows,
we
are
going
GA
I,
we
hashed
it
out
at
seek
release
and
cigar
kotecha.
Last
week,
SiC
architecture
indicated
no
concerns
on
us.
Moving
to
stable,
we've
got
nod
of
approval
from
the
two
stakeholders
at
Erin
and
Brian,
with
two
kind
of
follow-up
items.
A
Number
one
is
and
I
created
an
issue
for
this
is
we
want
to
make
sure
that
we
have
create
the
images
that
you
create
both
for
our
testing
as
well
as
the
post
image
already?
Has
it,
but
the
rest
of
the
test
I
want
to
make
sure
that
they're
built
from
source
code.
That's
in
kubernetes,
repo
and
they're,
hosted
in
like
something
like
GTR,
for
example.
So
they
want
us
to
follow
that
process.
A
So
we
don't
have
to
work
on
that
and
improve
that
down
the
line,
and
the
second
thing
was
a
documentation
which
we've
already
covered
very
very
well.
Actually,
the
docs
PR
I
got
merge,
while
we're
talking
in
this
call
so
good
stuff
there
so
and
then
seek
release
was
happy
with
the
sea
of
green
that
we
had
on
our
tests
across
the
last
10
days
or
so
like
the
Lena
mention.
A
A
Willa
Craig
is
unit,
so
so
that's
kind
of
the
team
feedback
and
like
I
mentioned
earlier
since
EF,
the
first
blog
post
for
1.14,
will
be
dedicated.
100%
Windows
we've
already
been
contacted
for
PR,
so
we
will
have
PR
responsibilities
to
be
Khan
and
that's
something
that
myself
and
Patrick
will
have
to
deal
with.
A
Any
other
questions
or
concerns
guys.
You
have
no
idea
how
important
this
is
by
the
way.
It's
huge
for
the
kubernetes
release
for
the
folks
that
have
been
here
for
a
long
time,
props
to
you,
I
actually
did
I
am
presenting
at
the
community
meeting
on
Thursday
and
I
have
a
slide
deck
on
Windows
and
one
of
the
big
things
I
have
there.
Is
that
the
folks
that
helped
us
get
here
right?
Recognition
for
everybody,
including
what
I
call
my
oh
geez,
the
original
gangsters?
A
There
are
some
of
the
folks
that
worked
on
cig
windows
from
long
long
time
ago
and
no
longer
part
of
this
community
either
because
of
change
in
company
priorities
or
other
things.
But
those
are
the
guys
that
brought
us
to
alpha
and
two
beta
and
without
them,
wouldn't
be
where
we
are
right
now
so
attend
a
community
meeting.
If
you
can.
I
A
G
Hopefully
he
won't
have
too
many
issues,
but
a
one
open
question
was:
he
was
looking
for
solid,
defying
like
how
did
you
always
specific
security
context
feels
like
should
like
SELinux
settings,
go
into
Linux,
specific
fields
and
Windows
stuff
like
GM
s
it
going
to
another,
and
that's
like
one
area
where
we
might
need
to
get
some
more
guidance
from
the
architecture.
Okay,.
A
Yeah
I
mean
Jordan
will
continue
being
are
like
Cynthia's
familiarity
the
feature
I,
don't
see
any
reason
why
he
would
not
continue
working
with
us
on
this.
It's
just
him.
You
know
what
the
process
would
look
like
and
then
good
thing
that
he
sports
both
on
API
reviewer
and
Asik
architecture
member.
So
it's
kind
of
killing
two
birds
with
one
stone
in
that
area
with
him
think
all
right,
excellent!
Well,
everybody
have
a
great
day.
We've
accomplished
a
lot,
the
last
three
months
or
last
three
years.
Sick
winters
actually
was
started
exactly
three
years
ago.