►
From YouTube: Kubernetes SIG Windows 20190507
Description
Kubernetes SIG Windows 20190507
A
A
The
first
one
is:
we
have
three
caps
in
for
1.15
the
kind
of
high
leveled
last
week.
That's
the
cube
ADM
one,
the
CRI
continuity
and
they're,
honest
username,
API
update.
All
three
caps
are
approved
for
the
1.15
cycle.
Two
of
them
were
had
to
be
side
loaded
later
in
through
an
exception
process,
but
they
were
all
approved
fairly
easily
and
we
have
a
game
plan
for
all
our
work.
A
I
think
the
one
that's
gonna
be
a
little
bit
the
most
contentious,
not
necessarily
because
of
the
implementation,
but
because
of
timing,
a
resourcing
is
a
cube.
Adam-16
luster
lifecycle
has
a
tremendous
number
of
responsibilities
and
they
don't
have
a
ton
of
time
allocated
that
they
can
use
for
cube.
Adm
reviews,
I'm,
hoping
that
we'll
be
able
to
snag
some
of
their
time
left
and
right
and
escaliers
working
on
this
with
Lumiere
we'll
be
able
to
get
sufficient
time
for
review,
so
can
Katie
twelve
are
with
one
to
fifteen.
A
If
anything
is
at
reduced
functionality,
so
our
goal
here
is
not
to
support
the
full
cube
ATM
functionality
from
phase
one,
but
allow
us
to
progressively
supporting
more
and
more
features
and
give
us
the
ability
to
provide
immediate
value
to
customers
and
iterate.
On
top
of
that,
is
there
anybody
that
has
any
questions
on
our
cab
or
the
enhancements
so
far.
C
I
said
in
there,
so
all
yours
Windows,
my
name-
is
lucky
I'm,
calling
with
my
release,
lead
hat
on
today.
So
what
we're
doing
as
part
of
the
115
release
team
is
just
going
round
up
every
cig
and
asking
for
a
human
readable
sentence
about
a
theme
that
or
sig
is
specifically
targeting
in
the
115
release.
So
you
don't
have
to
give
it
to
me
off
the
cuff,
like
I'm,
happy
to
follow
up
afterwards
unless
you're
willing
to
give
it
now.
C
But
basically
we
want
to
summarize
all
the
themes
and
then
put
together
the
appropriate
blog.
That
represents
all
the
great
work
that
everybody's
doing.
So
that's
the
ask
I've
kind
of
taken
a
look,
but
how
about
I
just
suggest
that
I
follow
up
with
you
Michael
and
Patrick,
and
up
with
the
sentence,
that's
works.
Yeah.
A
That
works,
I.
Think
the
overall
theme
of
our
release
here
is
continuously
improving
the
capabilities
of
Windows
and
kubernetes.
We
actually
do
have
a
security
focus.
This
release,
especially
with
the
run
as
username
and
and
the
CRI
container
D
and
some
of
the
stuff,
the
kind
of
fault
both
under
improvement,
as
well
as
under
security,
but
yeah,
let's
follow
up
offline,
will
give
you
the
one
sentence:
perfect.
That
was
it
for
me,
thanks
Pete
on
problem,
all
right
next
on
our
agenda
is
seek
face-to-face
during
comic-con.
We
actually
have
a
time
for
that.
A
It's
11:30
a.m.
to
1:00
p.m.
obviously
we,
if
you
don't
need
to
use
a
full
90
minutes,
you
don't
have
to
that's
on
Monday
May
20th.
You
do
have
to
register
for
the
contributor
summit.
That's
free
for
all
when
a
distributor,
so
I,
don't
think
anyone
is
grab
a
problem
and
that
can
I
get
a
quick
acknowledgement.
If
anybody
would
be
able
to
make
the
meeting.
Oh
man,
our
slaw,
our
soon
our
zoom
doesn't
have
the
option
to
raise
your
hand.
A
A
All
right
cool
all
right!
Well,
anybody
that's
coming
to
a
free
tier
to
sign
up
will
give
an
opportunity
chat
if
you
Eve
know,
buddy
shows
up
and
me
and
Patti
would
likely
walk
around
on
our
deck
and
any
demo
for
for
today
for
the
regular
presentation
of
Cuba
con
all
right.
These
are
the
items
that
we
had
on
our
agenda.
Is
there
anybody
that
wants
to
talk
about
the
agenda
items
or
anything
else
that
they
wanna
bring
up?
I.
B
A
B
So
the
way
it
works
is
first,
first
changes
go
into
the
so
that
they
do
to
there's
basically
two
rollouts
per
month.
One
of
them
is
a
cumulative
update
and
that's
list,
as
recommended
for
everybody,
and
then
the
other
one
is
an
optional
update.
So
it's
going
to
be
in
the
optional
update
in
May
and
it
will
be
in
the
cumulative
update
in
June
and
so
you'll
have
to
opt
in
into
it,
but
they
started
doing
that
to
give
people
a
chance
to
sort
of
pretest
some
of
the
changes
before
they
were
mandatory.
E
A
D
A
In
a
video
okay,
cool
yeah,
when
I
reach
a
resource
contention
across
kubernetes
between
to
become
US,
Europe
and
Shanghai
and
I
wanna
make
sure
if
anything
they'll
confront
load.
Let's
do
it
all
right.
Oh
I,
see
Dinesh
on
Dinesh
I
know
that
in
the
past
we've
talked
about
creating
a
troubleshooting
guides
for
networking
that
we
can
post
in
our
public
kubernetes
Doc's.
A
F
A
A
As
excellent
awesome,
you
know
the
place
we
should
put.
It
is
the
country.
Well,
let
me
ask
you:
did
you
guys,
second,
that
into
kind
of
two
sections,
if
I'm
a
developer,
developing
on
Windows
and
I'm,
kidding
to
some
networking
issue
or
I'm
an
end-user
and
according
to
some
networking
issues,
I'm
an
end-user.
A
Okay
got
it,
so
what
do
you
probably
wanna
do?
Is
you
want
to
put
that
on
kubernetes
dot,
io
so
needs
to
go
to
the
public
facing
documentation
so,
rather
than
actually
sending
the
dog
and
asking
for
reviews,
why
don't
you
just
create
the
PR
yeah,
let's
just
create
the
PR
I
think
that
we,
that
will
give
everybody
an
opportunity
to
provide
feedback
and
on
the
on
the
document.
A
F
A
So
you
guys
consumer
screen,
there's
three
or
four
issues
that
I
have
identified,
that
I
want
to
actually
talk
about
and
see
who
can
either
assign
an
order
for
them
or
some
of
the
contribution
up
further
the
first
one
is
addition:
I
wanna
show
the
in
fact
continued
image
between
a
private
repository
that
requires
auth
cannot
be
pulled.
Is
there
someone
here
that
I
can
look
into
these,
and
potentially
this
might
be
a
easy
fix?
It
might
be
hard
to
fix.
We
don't
know
it'd
be
great.
If
someone
can
take
a
look
at
this.
G
A
If
you
have
a
test
on
this,
the
unit
just
kind
of
talk
to
the
customer,
make
sure
that
we
understand
how
he's
doing
that.
I
know
that
and
I
assume
there's
already
off
enabled
on
the
on
the
repository
and
the
customer.
Not
the
customer
know
who
that
is.
He
basically
said
it
requires
basic
authentication,
so
is
doing.
A
G
A
A
H
A
F
F
B
F
F
A
H
A
B
A
B
A
A
A
A
A
A
Know
we
we
up
to
this
some
of
the
stats
right,
so
we
updated
how
we
were
not
completing
them,
but
how
we
were,
but
remember
that
back
that
we
had
that
were
basically
getting
the
stars
all
the
time
and
were
for
these
partial
commands
that
were
executing
to
both
create
and
populate.
The
stats
were
basically
bogging
down
the
system
so
made
those
updates.
But
there
were
those
before
before
been
deposited,
which
was
on
March
5th
or,
if,
if
not
before,
it
was
right
around
that
time
frame.
B
Go
ahead
and
sign
this
one
to
me
because
there's
another:
this
is
a
duplicate
of
some
another
issue
and
the
the
problem
here
is
that
some
of
these
stats
are
calling
PowerShell,
and
so
what
I
want
to
do
is
change
it
to
where
we
basically
keep
a
captive
PowerShell
instance
running,
because
it
takes
more
time
to
start
up
PowerShell
than
it
does
to
actually
run
the
command
that
returns
the
stats.
So.
A
I
was
under
the
impression
that
we
fixed
that.
Then
we
started
using
the
WMI
providers
to
do
not
I
think
that's
what
Pantheon
was
working
on
remember.
He
was
working
on
getting
there.
Basically
that
do
my
performance
providers
so
that
we
didn't
have
to
encounter
the
PowerShell
I,
see
on
the
stats
yeah.
B
B
A
B
A
J
B
A
Ok
cool
the
last
item
that
I
have
is
was
this
big
discussion
that
we
started
right
around
when
we
were
supposed
to
ship
114,
which
was
around
better
logging
options
for
Windows
cube
led
into
proxy
and
in
the
dogs
we
kind
of
talked
a
little
bit
about
being
able
to
use
the
SCCM
as
a
service
manager
and
basically
being
able
to
log
to
a
file
and
utilizing
both
the
log
rotation
and
a
log
file
size
from
SCCM
as
our
option.
However,
a
lot
of
folks
have
indicated
that
hey.
A
Why
don't
we
actually
push
these
logs
directly
to
directly
into
Windows
system
and
lock
as
an
alternative?
That
would
be
possibly
more
scalable
than
going
to
a
file
and
can
be
consumed.
Other
tools,
I'm
wondering
anybody,
can
take
kind
of
ownership
on
this
and
keep
creative
proposal
that
we
couldn't
review
in
the
future,
say:
windows
meeting
with
our
two
or
three
options,
as
well
as
a
recommendation
of
the
why
you
should
choose
a
or
b
and
give
us
opportunities
a
commune
to
discuss
it
and
and
I.
Don't
think
we
need
a
cap
for
this.
B
J
A
And
the
other
question
on
that
which
Madden
talked
about
here
is
that
container
D
they're
gonna
use
it
TW
tres
in
the
window.
So
if
this
becomes
an
obsolete
issue,
when
we
actually
move
to
container
D,
then
maybe
we
don't
need
to
worry
about
this
at
the
colony.
Just
for
continuity,
you
go
down
the
path
of
HW
tracing.
A
J
K
J
G
B
I
think
that's
I
mean
this
issue
is
a
very
tactical
thing
and
it
is
implemented
you
so
that
way
you
can
do
whatever
you
want
underneath,
and
so
we
could
install
something
on
the
host
they'll
convert
etw
to
a
text
file.
But
the
key
point
is
that
we
need
etw
implemented
if
you
want
to
be
able
to
collect
logs
the
same
way
that
you
collect
them
for
container
d,
h,
NS,
that's
right
and.
B
K
Mentioned
on
that
issue,
I
was
supposed
to
take
a
look
into
it,
but
I
don't
have
the
bandwidth
currently,
but
I
can
talk
to
Dinesh
about
having
someone
else
on
our
team.
Look
into
it
because
they'll
know
what
we
need
from
the
HMS
perspective,
which
is
usually
the
logs
that
are
the
most
relevant
when
debugging
acute
proxy
enemies.
Well,.
K
I'd
have
to
talk
to
Nathan
more
about
what
cubelet
logs
would
need
to
look
like
I.
Just
like
we
need
to
consider
both
and
Nathan
did
some
work
for
another
project
where
he
added
logging
to
it
or
enhanced
it
so
I,
just
personally,
don't
have
the
time
to
look
into
it
anymore,
but
I
can
find
someone
on
our
team
who
can
at
least
give
input
on
it.
Okay,.
A
J
A
J
A
So
it'd
be
great
for
us
to
kind
of
come
up
with
a
proposal
here
and
and
and
that
we
can
review
right.
So
so
I'm
gonna
put
here
clear
to
work
with
her
team
to
identify
donor
that
will
create
a
proposal
and
let
me
actually
put
some
requirements
based
on
this
discussion.
Right
requirements,
pic
logging
framework
with
etw
facing
being
being
a
likely
candidate.
A
A
Want
to
figure
out
what
do
you
want
to
do
first
and
then,
let's
figure
out
how
we're
gonna
do
it
or
the
cap
process
I
the
cap
process?
Is
there
so
can
provide
structure
to
our
ability
to
deliver
features
and
kubernetes
right
now?
What
I
want
to
do
is:
let's
figure
out
what
the
best
option
is
like
how
we
should
do
this
and
identify
the
right
way
to
do
this,
and
then,
if
we
need
to
file
a
cab
and
get
that
approved
very
late.
Who
can
do
that?