►
From YouTube: Kubernetes SIG-Windows 20211109
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
Hello,
everybody
and
welcome
to
the
november
9th
2021
iteration
of
the
kubernetes
windows
community
meeting.
As
always,
these
meetings
are
recorded
and
uploaded
to
youtube
so
be
sure
to
adhere
to
the
cncf
code
of
conduct
got
some
announcements
today,
mainly
around
the
123
release
cycle.
I'd
just
like
everybody
to
be
aware
of
the
first
one
is
code.
Freeze
is
november
16th
that
is
next
tuesday.
A
Usually
they
start
the
kind
of
they
actually
start
to
enforce
that
at
midnight
of
whatever
time
zone.
Some
of
the
release
folks
are
in
so
it's
best
to
get
those
in
as
early
like
before
that,
but
usually
kind
of
end
up
business
hours.
This
is
a
safe
bet
later
on.
A
I
want
to
do
a
collection
of
prs
that
we
do
think
we
want
to
get
in,
so
we
can
discuss
some
of
those
later,
just
to
make
sure
that
we
can
get
reviews
and
all
of
those
merged
next
is
docs,
so
for
any
of
the
enhancements
that
need
documentation
updates,
which
I
believe
is
all
three
of
the
enhancements
that
we're
tracking.
These
are
the
some
of
the
deadlines
set
by
the
release
team.
A
A
The
next
deadline
is
they
want
the
any
of
the
docs
pr's
to
be
out
of
draft
state
and
reviewable
by
both
the
subject
matter,
experts
and
the
docs
team
by
november
23rd.
That
is,
I
believe,
the
tuesday
before
thanksgiving
for
folks
in
the
u.s.
So
please
be
aware
of
that,
and
then
the
last
deadline
is
the
they
want
the
docs
to
be
reviewed
and
approved
by
november
30th.
A
Now
that
doesn't
necessarily
mean
that
they
need
to
be
merged,
because
the
release
team
likes
to
kind
of
control
the
flow
of
some
of
the
merging,
but
they
would
like
reviews,
especially
from
the
content,
owners
and
subject
matter,
experts
for
that.
I
believe
that
is
the
following
tuesday,
so
last
tuesday
in
november.
A
So
it's
better
to
kind
of
open
up
some
of
the
draft
pr's
early.
Just
to
make
sure
that
these
deadlines
all
right
meet
these
deadlines.
B
I
just
have
one
question:
maybe
I
can
take
it
later.
This
is
regarding
container
d,
where,
like
when
we
try
to
pull
windows
image
in
docker,
we
increases
timeout
to
30
minutes
at
20
minutes,
because
windows
images
are
pretty
huge
all
right.
So
that
support,
I
think,
the
pr
I
saw
it
is
still
not
there
in
continuity,
1.6
beta
version-
I
don't
know
like
do
we
say
windows,
people
do
we
work
with
them
and
push
it
or
how
does
that
flow
works?.
A
Yeah,
we
can
quickly
cover
this.
There
are
there's
some
overlapping
maintainers
for
the
two
projects
and
we
generally
kind
of
work
well
with
each
other,
but
most
of
the
time
for
container
dprs.
We
don't
really
track
the
status
in
in
these
meetings.
A
B
Just
type
those
names-
probably
I
can
in
the
slack
I
will
try
to
talk
to
them.
That
will
be
really
good.
Okay,.
A
No
problem,
okay,
I
guess
next
is
the
the
prs
and
work
that
we
want
merged
for
123..
A
I'd
like
to
do
a
quick,
just
kind
of
pulse
on
all
of
the
the
ongoing
work
to
see
if
like
what
we
still
need
to
merge
as
a
sig
and
where
we
are
with
all
of
the
enhancements.
So
I
can
start.
I
think
I
saw
that
for
the
host
process.
Containers
enhancements.
All
of
the
pr's
I
wanted
to
make
in
kubernetes
are
done,
so
we
are
planning
on
going
to
to
beta
that
merged
yeah
yeah
jay
also
reviewed
the
all
the
criteria
and
merged
that.
A
B
Yeah,
so
both
the
pr's
and
the
qubriod
side
got
merged.
Like
remember,
we
had
a
discussion
last
week
around
the
n,
minus
two
version
of
cubelet
may
be
having
some
issues
say.
If
we
introduce
the
feature
gate
in
the
cubelet
and
cubelet
is
going
to
reject
certain
parts
of
the
os
does
not
match
the
the
operating
system
on
which
the
cube
rate
is
running.
B
So
we
were
jordan,
initially
proposed
that
this
should
be
merged
after
n,
minus
2
is
125
or
126
depending
on
when
this
feature
ga
is
so.
B
So
I'm
I
I
told
him,
I'm
fine
with
it,
so
I
am
working
on
making
those
changes.
So
hopefully
I
can
get
them
merged
before
the
end
of
this
week.
A
Okay,
yeah
ping,
us
when
you,
if
you
have
those
changes
out,
one
thing
I'd
like
to
just
reiterate,
is:
if,
if
there
are
prs
for
enhancements,
please
link
them
back
to
the
enhancements
the
shoes
I
went,
and
I
did
that
for
some
of
the
os
fields
too.
It
just
helps
the
release
team,
know
kind
of
the
where
the
how
things
are
trending
too,
but
sure
yeah.
A
Okay,
so
I
think
there's
that
one
here
and
then
for
the
docs
prs
this,
we
did
say
that
this
enhancement
needed
documentation,
updates,
correct.
B
Yeah,
so
the
docs
for
the
field
in
the
pod
spec,
it
should
be
available
because
of
swagger,
so
that
should
be
good,
and
I
have
already
opened
a
pr
for
the
port
security
standards
based
on
this
field,
but
it's
I
think
I
open
it
against
master.
I
need
to
change
it
to
123
branch.
I
I
will
get
to
it
once
I'm
done
with
the
existing
pl.
I
should
be
able
to
open
it
before
november
18th.
A
B
The
other
thing
is
I'll
be
out
next
week,
so
most
probably
I'll
try
to
get
everything
done
before
the
code.
A
Freeze,
if
you
need
any,
if
you're
going
to
be
out
and
things
aren't
merged,
reach
out
to
folks
and
slack
so
what
I've
done
in
the
past
is,
and
when
I
was
in
that
situation
is
made
other
folks
kind
of
contributors
on
my
fork
and
then
helped
just
had
some
folks
help
get
those
prs
merged.
If
there
was
any
feedback
or
you
know,
merge
conflicts
so
yeah.
B
C
Do
we
need
any
updates
to
kind
of
the
windows
like
page
to
just
talk
about
how
this
might
be
used.
B
Yeah,
so
that's
that's
something
that
I
wanted
to
talk
about
as
part
of
the
docs,
but
I
I
did
not
look
into
the
existing
state
of
documentation
before
I
can
make
recommendations.
So
since
we
have
like
next
two
to
three
days
I'll
go
through
it
and
then
when
I'm
opening
I'll,
let
you
guys
know:
okay.
A
Okay,
the
node
service
log
viewer-
I
saw
some
more
action
on
that
arvind
or
sebastian
do
either,
if
you
want
have
any
updates
so
which
one
is
this.
D
The
node
service,
the
no
log
service
yeah,
so
christian
and
I
have
been
hacking
away
at
it.
I
don't
think
it's
going
to
make
it
this
release.
The
last
update
we
have
is
we
got
all
the
piping
working,
so
the
api
has
been
added.
All
the
piping
is
working
we're
now
having
to
test
this
across
linux
and
windows
and
I'm
at
the
moment
trying
to
add
the
heuristics
to
it.
So
this
is
yeah.
This
is
definitely
at
risk.
D
Okay,
but
I
don't
want
to
say
we're
not
going
to
do
it
because
I
just
want
to
keep
the
pressure
going
so
that
it'll
make
me
and
christian
it'll
push
us
forward
to
try
and
get
this
through.
So
hopefully
what
my
plan
is,
you
know
I
don't
want
it.
I
don't
want
me
and
christian
to
drop
the
pressure
on
this
so
that,
at
least
in
the
next
cycle
we
start
like
reviewing
it
very
early
on
and
just
push
it
through.
So
that's
the
plan
from
us.
A
A
Okay,
we
will
keep
just
keep
up
to
date
on
that,
so.
B
To
be
clear,
arvind
like
you
would
not
need
any
reviews
this
week
right
before
18.
A
A
Okay,
I
saw
that
somebody
like-
I
think
these
are
james's
peers
to
a
couple
of
metrics
related
issues
for
performance
reasons
that
I
think
we
would
like
to
get
merged
in
123
as
well.
Does
anybody
if
anybody
has
any
other
kind
of
critical
pr's
that
want
to
land
in
123?
A
Please
either
bring
them
up
or
add
them
to
the
dock
here
or
just
message
us
on
slack,
so
we
can
stay
on
top
of
them
and
really
push
on
those.
D
A
Can
add
that
here,
just
for
completeness
docs
do
have
more
time.
A
Okay,
I'll
I'll,
add
it
there
did
you
have
anything
you
wanted
to
say
about
either
of
these
james
or
just
adding
them
there
for
tracking?
I
think
that
they're.
A
Okay,
next
thing
is:
I
wanted
to
ask
how
folks
felt
about
having
a
having
or
not
having
a
community
meeting
on
november
23rd.
A
It
is
the
weekend
or
the
tuesday
before
thanksgiving,
which
is
a
big
holiday
in
the
united
states
at
least,
and
I
wanted
to
kind
of.
If
people
aren't
wanting
to
join,
we
can
give
people
time
off
to
relax
and
spend
time
with
family
or
if
people
are
here,
we
can
have,
we
can
meet
I'm
fine
either
way.
Let
me
know
what
other
people
want,
so
the
majority
of
people
aren't
going
to
be
here.
A
We
can
cancel,
but
if
we
do
need
to
discuss
anything
especially
around
the
docks
you
can
meet,
does
anybody
have
any
thoughts.
E
A
All
right,
that's
good,
so
I
guess
we'll
cancel
if
there's
nothing
on
the
agenda
by
the
monday
evening,
but
yeah.
I
agree.
It's
kind
of
important
to
wind
down
and
just
relax.
Take
a
break.
Okay.
Next
was
the
windows
operational
readiness
kept.
I
mean.
F
What's
this
deciding
again
the
agenda
to
for
us
to
as
a
reminder
of
the
pr,
so
we
are
almost
good
to
go.
We
tackle
the
open
issues
like
created
by
for
the
plugin
that
we're
missing
and
jay
tackled
most
of
the
comments,
so
just
a
reminder
for
a
lesson
review.
A
H
Oh
yeah
me
and
james
and
amin
were
fighting
about
that
last
week
or
I
was
I
wasn't,
gonna
have
to
yeah.
I
I
think
I
think
the
the
conclusion
was
everybody
agreed
on
sono
buoy,
which
is
fine.
I
think
totally
like.
I
think
that's
great.
I
I
didn't
want
to
be
the
advocate
for
that.
Since
you
know
I've
been
in
vmware
for
so
long
and
I've
always
been
a
sono
buoy
advocate.
I
felt
like
it
would
be
weird,
but
if
other
folks
who
aren't
at
vmware
are
like
yeah,
let's
do
it
in
sonaboo.
I
B
Cnc,
it's
that's
how
most
of
the
cnc
distribution
testing
is
happening
right.
H
B
H
H
E
E
H
The
alternative
would
be
to-
and
this
is
the
thing
that
I
I
was
talking
about
last
week
and
is
the
alternative-
would
be
that
we
build
our
own
tool
that
in
our
own,
like
set
of
like
definitions
and
stuff
inside
that
tool,
you
know
for
reporting
and
all
that
stuff
and
then
we'd,
like
version
the
tool
and
all
that,
but
like
sony
kind
of,
does
that
some
a
lot
of
that
for
you
like
it,
it
sort
of
has
a
this.
Is
the
kubernetes
version
that
this
tool
works
with?
H
Yeah
yeah
suno
buoy
is
just
a
wrapper
to
the
e2e
tests,
ultimately,
at
the
end
of
the
day
right.
So
what
james
said
right
it
packages
the
logs
for
you,
it
allows
you
to
build
plugins
that
can
wrap
all
the
command
line
options
to
ginkgo.
So
you
can
add.
You
know
more
behavior
that
we
care
about
that's
specific
to
windows
without
embedding
that
in
the
test
code
and
so
on.
E
H
A
H
A
A
All
right
yeah,
so
for
this
sorry
I
mean
and
folks
that
I
think
this
has
been
a
little
bit
slow,
like
slaughter
review,
hopefully
with
code
freeze
coming,
but
folks,
like
I
know
I'll
have
more
time
to
you,
know
dedicate
to
this
and
you
can
get
this
merged.
So
thank
you
for
working
on
this.
I
think
this
is
important.
H
A
A
For
this
well,
I
like
getting
it
merged
and
getting
it
into
the
124
release
are
separate
like
we
can
always
merge
it
in
a
provisional.
We
can
always
merge
it.
B
A
A
I
did
want
to
spend
a
little
bit
of
time
too,
since
we
are
kind
of
the
123
releases
entering
code
freeze
pretty
soon
to
at
least
start
planting
some
seeds
about
what
some
of
the
sig
windows
priorities
are
for
124,
so
that
we
can
start
like
looking
at
them
figuring
out
resources
for
for
working
on
those
different
generation
chips
and
everything.
I
threw
out
a
couple
of
ideas
here
that
we
can
expand
on
a
little
bit,
but
everybody
feel
free
to
either
add
to
them.
Think
about
these.
A
We're
probably
discussing
this
in
more
community
meetings,
especially
later
in
the
124
really
early
on
the
124
release
and
later
in
the
123
release,
we're
going
to
continue
the
enhancements
that
we
have
here.
So
I
don't
think
we
need
to
talk
about
a
lot
of
that,
but
outside
of
the
existing
enhancements,
which
are
going
on
like
the
os
field
host
process
containers
the
nodes
log
viewer.
A
I
think
that
now
is
getting
to
a
point
where
we
need
to
start
focusing
and
improving
some
of
the
capacity
story
for
windows,
and
things
like
there's
been,
there's
been
a
kind
of
a
trickle
of
pr's.
Like
the
peers
that
james
has
to
improve
performance
when
nodes
are
under
high
density
and
everything
like
that,
but
windows
still
doesn't
support
eviction.
A
A
That
all
makes
sense
to
me-
and
I
guess
also
for
like
folks
who
interact
with
either
customers
or
yourself
more
are.
Are
you
guys
seeing
issues
with
kind
of
capacity
and
density
on
windows?
I
know
in
kind
of
in
azure
we
occasionally
do
hence
some
of
the
like
the
issues
and
pr's
and
the
the
kind
of
motivation
for
this
work.
I'm
curious.
You
all
see
this
as
well.
H
A
I
think
yeah
it
depends
on
like
the
node
size
too,
we're
seeing
that
things
definitely
are
not
as
happy
when
people
try
and
over
provision
like
greatly
over
provision
machines
on
linux.
I
think
it's
kind
of
expected
that
you
can.
You
know
over
provision,
and
hopefully
like
workloads,
won't
get
starved
out
if,
if
necessary,
or
look
at
move
to
other
nodes-
or
you
know
scaled
appropriately
and
that
doesn't
work
as
smoothly
on
windows.
A
A
Yeah,
okay:
we
can
continue
to
kind
of
talk
about
this
and
see
some
specifics.
Another
area
of
feedback
that
I
think
we've
touched
on
a
little
bit
and
there's
been
a
couple
of
attempts
to
welcome
some
worker
on
this
but
docs.
A
Today.
First
for
windows,
we
have
kind
of
one
really
big
page
on
the
kubernetes
website.
That
has
a
whole
lot
of
information
about
everything
windows
related
and
I
think
it's
kind
of
hard
to
consume
or
find
in
like
individual
pieces
of
information
and
then
there's
a
couple
of
other
and
then
some
of
the
windows
documentation
is
kind
of
thrown
in
with
the
other
docs.
Here
I'm
wondering
if
people
like
how
people
are
people
satisfied
with
the
current
documentation
for
windows,
should
we
work
on
breaking
that
up?
A
E
Yeah
there
were
a
couple
of
questions
recently,
I
think,
both
at
kubecon
and
from
customers
who
were
trying
to
do
their
own
kubernetes
for
windows,
and
they
were
asking
if
netapp
has
a
csi
proxy
or
not.
And
then
there
were
a
couple
of
questions
like
that,
which
I
felt
like
the
infrastructure
providers
have
it
covered
pretty
well
for
their
own
thing.
But
if
somebody's
looking
just
at
kubernetes
windows,
they
don't
get
a
full
picture
of
like
what's
available.
What
what's
not
available.
E
H
I
have
a
quick,
I
gotta
drop
for
10
minutes
or
15
minutes,
and
then
I
can
come
back
for
the
pairing,
because
I
have
a
15
minute
scrum
meeting
thing
I
have
to
do
today.
Does
anybody
want
to
prepare
after
this
or
do
we
do
we
want
to
skip
today?
I'm
definitely
down
to
do
it.
I
just
can't
do
the
you
know.
A
A
Problem
I
see
james
added
or
somebody
is
that
you
james
added
that
yeah
cuba
dm
and
two
proxy
images.
I
guess
we
can
briefly
talk
about
that
and
then
in
the
meeting.
C
Yeah
I
just
we've:
had
the
cube
adm
kept
open
for
several
releases
now
I
think
we
got
it
to
beta
or
something
and
now
that
we
have
host
process,
we
can
get
this
working
fairly
cleanly.
I
think-
and
so
I
was
thinking
that
that
should
be
a
good
priority
for
our
next
one.
I
think
it'll
reduce
a
lot
of
the
noise
that
we
get.
C
A
D
A
A
But
things
like
running
you
know:
crew
proxy,
as
a
demon
set,
I
think,
are
things
that
we
can
definitely
pursue
at
running
as
host
process
containers,
and
I
think
that
there's
also
a
lot
of
opportunity
to
find
other
projects
that
could
benefit
from
or
that,
like
don't
have
windows
support,
because
there
wasn't
a
privileged
containers
kind
of
equivalent
that
we
can
start
integrating
with
like
prometheus
the
node
exporter,
and
things
like
that.
So.
A
C
I
Yeah,
I
think
we
can
do
something
similar
with
it,
with
the
pause
image
and
the
other
images
as
well.
We
can
use
the
same
pipelines.
C
Yeah,
I
have
the
q
proxy
stuff
that
I
did
in
the
sig
windows
tools
project,
so
you
can
probably
take
that
and
clean
it
up
and
improve
it
and.