►
From YouTube: Kubernetes SIG Windows 20230815
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
Okay
welcome
everyone.
This
is
August
15th
2023.,
and
this
is
the
Sig
Windows
community.
A
Every
Tuesday
at
9,
30
Pacific
and
we
are
cncf
project.
So
we
follow
the
scenes.
You
have
guidelines
if
you
have
any
concerns
or
questions
about
those
please
reach
out
and
go
ahead
and
get
started.
A
So
announcements
I
think
it's
not
on
here,
but
128
comes
out
today,
so
it
is
at
least
today
at
some
point.
So
that's
exciting.
Thank
you
to
everyone
who
contributed
to
that
and
does
anybody
else
have
any
other
announcements
that
they
want
to
bring
up.
A
All
right
so
then,
the
other
thing
we
do
is
just
give
a
moment
for
anybody.
That's
new
to
the
community
to
take
a
moment
and
say
hello,
and
let
us
know
what
you're
interested
in
and
how
do
you
want
to
get
involved.
A
Okay,
cool,
since
no
one
said
that
we'll
continue
on
with
the
agenda,
so
we
have
a
few
things
here.
The
first
one
is
from
Jay
and
I.
Don't
see
him
today
and
I,
don't
know
exactly
what
this
means.
I
think
we'll
skip
that
one
and
then
Jay
shows
up
and
put
it
back
on
unless
anybody
that
has
anything
they
wanted
to
add
here.
A
Muncie's
also
not
here
I
know
she
started
a
thread
in
see.
Windows.
A
Did
you
want
to?
We
started
the
agenda,
but
you
had
the
first
one
here,
this
agents,
what
balance
did
you
want
to
quickly
talk
about
that.
D
I
think
it
was
a
question
about
what
so
so
there's
these
hns
has
this
load
balancer
parameter
that
you
can
send
it
and
we're
digging
into
this,
and
it's
like
distribute
for
connection
zero,
distribute
traffic
of
the
same
protocol.
One
distribute,
client,
ip2
and
I
was
just
wondering
if
that
is
like.
If
those,
if
those
are,
are
we
actively
like
changing
or
modifying
those
or
do
those
map,
or
has
anybody
documented
how
those
are
supposed
to
map
over
to
Coupe
proxy
Behavior
at
all?
For
you
know
what
I
mean
like?
D
A
That
I,
don't
I,
don't
know
the
answer.
I,
don't
know
that
setting
do
you
know
how
it's
being
used
in
the
current
computer
proxy.
D
So
yeah
I
didn't
I
didn't
get
a
chance
to
look
much
into
it.
I
didn't
yeah
I,
don't
I,
don't
really
have
I,
don't
really
have
anything
there.
I
figured,
maybe
somebody
would
just
know
but
but
yeah.
A
Yeah,
but
we
can
start
at
thread
and
see
windows
and
try
to
get
David
to
take
a
look
at
it.
Yeah,
okay,
you
would
know
off
without
his
head,
probably
yeah.
D
D
B
D
But
it's
like
just
one
line
now:
I.
Remember
this
because
I
looked
at
this
yeah,
so
it's
in
there
there
it's
like
there's
some
code
that
says
like
if
it's
session
Affinity
is
on
then
distribute
by
The,
Source
IP,
but
like
I,
don't
know
whether
that
really
works
like
or
whether
we're
working
on
that
or
testing
it.
You
know
so.
A
There,
it
is
I
think
there
is
a
test
that
covers
Affinity
I,
think
it
was
added
in
like
121.
A
B
A
So
somebody
so
is
there
any
other
questions
or
comments
on
that?
One.
A
All
right,
so
someone
opened
up
a.
A
Issue
here
for
supporting
pot
eviction
in
Windows,
our
documents
say
that
in
fiction
hard
and
eviction,
software
not
implemented
so
I
I
started
looking
into
this
I
know
that,
like
we
do
have
some
working
stuff
around
pod
disruption,
budget
and
then
also
Mark
and
I-
didn't
link
it
here.
But
Mark
did
some
work
to
verify
that
if
there's
disk
pressure,
then
there
is
cleanup
done
for
that.
A
But
I
don't
know
if
there's
any
pod
evictions
that
happen
because
of
that
so
and
I
don't
know.
If
anybody
else
has
looked
into
this
or
is
aware
of
this
but
or
requires
this.
But
I
was
just
kind
of
collecting
some
information
here
on
odd
eviction.
A
Anybody
aware
Brandon,
if
you're
speaking,
you're
on
mute.
A
Okay,
I'll
continue
to
do
some
research
on
this
and
figure
out
what
what
works
and
what
doesn't
mean.
Last
time,
I
read
the
docs
on
this.
It
they've
changed
quite
a
bit,
so
I
think
even
some
of
the
implementation
might
have
changed
so
probably
need
to
revisit
this.
A
Cool
and
then
the
last
thing
on
the
agenda
here
is
vagrant.
If
you
haven't
seen,
hashicorp
updated
their
business.
D
A
And
so
there's
a
thread
in
Slack
around
identifying
usages
of
any
of
the
hashing
Corp
libraries
and
products
within
any
of
the
kubernetes.
Some
projects-
Sig
Windows
Dev
tools
came
up.
Jay
I
saw
you
commented
on
that
issue.
A
We
don't
actually
package
anything
or
all
in
a
libraries
as
far
as
I'm
aware,
but
I
thought
I'd
bring
it
up
here.
Just
in
case
somebody
knew
differently.
I
think
we
were
even.
D
Yeah
yeah
I
mean
we
want
to
get
away
from
vagrant.
We
haven't
found
anything
better,
though
I
assume
it
doesn't
affect
us
at
all,
though,
because
we
just
consume
it
as
a
as
an
executable
people
download
it
they
install
it
on
their
own.
We
don't
vendor
any
of
its
libraries.
We
don't
ship
anything,
we
don't
even
have
releases
right,
so
I
don't
think
it
affects
us
at
all.
It's.
D
A
Send
out
some
guidance
this
morning
on
it:
okay,
which
I
can
link
somewhere,
but.
C
A
Yeah
we
use
Packer
over
in
the
image
Builder
project
and
so
we're
trying
to
figure
that
out
right
now.
That's.
A
A
Nope,
okay,
cool:
is
there
any
other
agenda
items
or
anything?
Anybody
wants
to
talk
about
today.
A
If
not,
we
did,
by
the
end,
I
opened
up
a
PR
for
some
windows,
documentation
I
think
we
could
potentially
review
that
as
a
group
I
think
that
was
kind
of
you
know
one
of
the
things
that
we
have
a
light
agenda
and
we
have
a
few
extra
minutes.
We
can
just
go
over
stuff
and
discuss
it.
So.
A
See
y'all,
do
you
want
to
just
share
your
your
PR
and
we
could
quickly
review
it
as
a
group
and
get
some
early
feedback.
E
Sure
thing:
can
you
hear
me
yeah
yep.
E
In
this
team,
I'm
doing
a
message,
I
can't
share
my
screen.
Let
me
drop
something
I
know.
A
E
Okay,
perfect,
so
this
is
the
pr
it's
the
pr
to
specify
like
the
hundred
requirements
that
we
want
for
Windows,
so
I
think
the
best
would
be
to
open
preview.
E
B
E
So
I've
added
it
in
this
section.
This
is
Windows
containers
and
kubernetes
I,
don't
know
if
you
want
it
to
have
its
own
section.
Maybe
that's
the
first
thing
you
can
want
to
discuss
and
bring
it
down
here.
So
I
guess
the
challenge
with
specifying
the
requirements
from
what
I've
seen
is.
The
requirements
will
differ
based
on
what
the
person
has
for
their
workflow
requirements
and
what
they
might
have
installed.
But
right
now
most
of
what
I've
specified
is
relying
on
the
documentation
from
Microsoft
for
Windows,
server
and
I.
E
E
F
A
D
E
A
Think
windows
and
containers
and
kubernetes
seems
like
the
perfect
place
for
this.
We
can
link
the
hardware
requirements.
I
think
the
free
space
of
disk
is
needs
to
be
much
higher.
A
I,
don't
know
if
we
have
actual
guidance
anywhere
in
Brandon
I,
don't
know
if
you
know
if
we
have
actual
guidance
somewhere
on
like
the
windows
side
of
things
but
I
think
for
AKs.
We
we
at
least
bumped
that
to
100
gigabytes
as
soon
as
you
download
one
or
two
images.
You're
gonna
hit
that
free
disk
space
pretty
quickly.
That's.
B
Yeah
I'm
not
sure
if
we
have
it
on
the
Windows
containers,
docs
page
but
yeah
20
gigs
might
be
a
bit
small.
F
Should
not
processor,
it's
only
one
core
or
what
always.
E
That's
a
good
point:
I
think
it's
supposed
to
be
four
four
cores
I,
think
that's
why
I
remember
seeing
I
think
I
should
specify
that
that's
a
good
point.
E
A
It
looks
pretty
good
here,
I
think
so,
thanks
for
taking.
E
F
As
for
other
requirements,
there's
one
thing
that
jumps
to
mind
if
this
is
still
needed,
I
think
before
I,
don't
know
if
it's
still
true,
but
we
kind
of
needed
the
virtualization
enabled
in
order
to
create
the
V
switch,
requires
to
tie
the
Sienna
to
to
pause
themselves
only
once
basically,
if
I,
remember
correctly
after
which
is
no
longer
necessary.
Is
that
still
the
case.
A
So
we
have
a
workaround
that
allows.
B
A
To
to
get
that,
and
maybe
we
should
call
that
out
because.
B
F
F
Because
I
think
I
remember
recently
someone
tried
to
deploy
Windows
worker
nodes
for
kubernetes
and
all
of
the
issues
that
they
encountered
and
I
I.
Don't
remember
which
club
provider
they
basically
had
to
switch
the
flavor
to
a
higher
one
which
had
necessary
transition
enabled
just
to
actually,
you
know,
configure
the
the
worker
nodes
so.
A
Yeah
we
do
have
a
link,
and
this
come
up
several
times
so
I
think
it
probably
makes
sense
to
link
it
in
here.
So
I
will
find
that
and
give
it
to
you
once
I.
Have
it
good.
E
Call
okay,
thanks,
I
think
James
I'll
probably
reach
out
to
you
to
see
how
we
can
phrase
that
list,
but
thanks
for
that,
Claudia
or
the
five
being
I'm,
not
sure.
C
So
the
other
thing
I
just
want
to
call
out
is
I
think
we
should
need
make
it
clear
I
that
these
minimum
requirements
that
we
are
specifying
could
okay.
So
how
do
I?
How
do
I
erase
this?
So
we
are
saying
that
these
are
the
minimum
requirements.
C
But
then
we
are
also
listing
a
bunch
of
other
considerations
that
need
to
be
taken
into
account,
which
are
you
know,
container
on
time
requirements
kubernetes
core
components
that
run
I
feel
like
we
should
at
least
take
into
effect
or
take
into
like,
maybe
take
into
consideration
the
core
components
and
the
runtime
when
we
come
out
with
a
minimum
Hardware
recommendation,
because
it's
going
to
be
hard
for
a
user
to
figure
that
out
on
their
own,
in
addition
to
like
them
having
to
run
an
antivirus
or
something
like
that
right.
E
I,
don't
know
if
it
addresses
what
you're
mentioning,
but
what
I
did
do
is
I
spun
up
a
node
and
I
monitored
the
resource
usage
with
just
the
core
components
that
we
did
I'm
like
checking
about
the
CPU
utilization
and
the
RAM
usage
I
can
try
and
find
my
notes,
but
that's
also
something
I
took
into
consideration
to
put
these
matrixes
here,
but
also
something
you
said
that
I
liked
is.
We
need
to
really
make
it
clear
about
what
these
minimum
requirements
are
based
on.
C
Yeah
yeah,
because
you
know
you
have
said
that
these
are
the
minimum
requirements
and
you're
going
to
bump
some
of
them
up,
but
but
you're
going
to
add
the
kubernetes
core
and
the
container
runtime.
C
And
then
the
other
thing
you
really
need
to
call
out
is
what
workload
is
the
customer
going
to
run?
If
the
customer
is
going
to
run
some
I've
seen
some
customers
trying
to
just
take
a
VM
and
convert
it,
convert
it
into
a
container
and
then
they
need,
like
I,
don't
know,
48
gigs
of
memory
and
eight
cores.
C
So
at
that
point
the
requirement
for
the
node
just
becomes
what
you
call
it
becomes
dependent
on
what
the
customer
customer
workload
is.
So
that
is
the
other
thing
we
need
to
call
out
here
is
that
this
is
the
minimum
requirement,
including
the
kubernetes
core
components.
C
A
E
Okay,
I'll
find
how
I
can
rephrase
that
I
get
the
point
and
then
I'll
definitely
reach
out
to
you
to
Alvin
to
just
give
some
feedback
of.
If
you
think,
IT
addresses
it
nicely.
Yeah
sounds.
C
Good
yeah
tag
me
on
the
pr
and
I'll
take
a
look,
but
this
is
great.
It's
a
great
start,
so
thank
you
for
doing
it.
E
Maybe
the
last
question
I
have
before
I
close
this
office.
Is
the
like
I
try
to
think
of
situations
that
maybe
we
want
to
call
out
something?
Maybe
I
mean
that's
specific
to
Windows.
A
good
example
is
with
with
Linux
they
discourage
enabling
swap
memory
on
the
nodes,
because
it
might
impact
the
performance
I'm,
not
sure
if
it
still
applies,
because
it
was
something
that
seemed
to
apply
to
all
of
it
to
the
kubernetes.
F
Okay,
I
mean
cubeadm
explicitly,
gives
you
an
error
if
you
try
to
initialize
it
or
join
it.
No,
no,
which
has
swap
but
not
on
Windows,
so
such
check
doesn't
exist
on
windows
at
this
moment
might
still
be
interesting
to
have
in
the
future
Maybe.
F
But
it's
not
a
requirement
that
this
time,
but
another
thing
that,
if
you're
mentioning
other
requirements,
not
sure
if
we
support
any
other
container
runtimes
at
this
moment,
but
it
might
be
useful
to
mention
that
if
continuity
is
being
used,
they
are
going
to
need
at
least
compare
the
1.6.
In
order
to
have
phosphorus
containers,
which
we
use
for
the
cnis
and
qproxy
Demon
sets.
A
Yeah
sort
of
out
of
scope
for
this
particular
section
of
the
document,
but
yeah
I
think
we
call
that
type
of
thing
out
in
some
other
areas
around
community.
You
know.
E
I'll
screw
that
for
now
I'm
making
a
little
bit
either
way,
but
I'll
exclude
it
from
my
change.
But
thanks
as
always
told
you
yeah.
A
Yeah
yeah
sorry
I
was
on
mute
and
thank
you
for
for
doing
that
work
and
we
will
see
everyone
next
week.