►
From YouTube: Kubernetes SIG Windows 20220301
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
march
1st
2021
2022
iteration
of
the
kubernetes
sig
windows
community
meeting.
As
always,
these
meetings
are
recorded
and
uploaded
to
youtube
so
be
sure
to
adhere
to
the
cncf
code
of
conduct
start
off
with
some
announcements
james.
I
believe
this
is
your
announcement.
If
you
want
to.
B
A
couple
weeks
ago
I
did
a
demo
for
the
windows,
exporter
and
the
image
because
of
some
ci
issues.
The
image
didn't
get
published
properly,
but
the
folks
over
there
fixed
that
and
so
now
there's
an
image
available,
and
it's
just
the
latest,
but
you
can
test
with
it
and
the
next
release
of
windows.
Exporter
should
have
a
versioned
release
of
the
image,
so
give
it
a
try
and
let
us
know.
A
A
Does
anybody
have
anything
else
they'd
like
to
announce
here,
you
can
just
throw
it
out.
If
not,
we
can
go
into
some
of
the
agenda.
C
We
have
one
question
regarding
windows:
2022
the
container
image
support.
I
don't
know
whether
we
can
talk
about
it
later.
A
C
Good
question
muhammad:
can
you
shoot
the
question
muhammad's
from
our
team
muhammad?
Can
you
hear
us
sure.
C
Okay,
so
I'm
not
sure
basically,
when
we
try
to
test
a
windows,
2022
server
looks
like
docker
runtime.
By
default
we
used
to
have
docker
runtime
by
default.
Install
looks
like
we
are
not
able
to
find
that
image
in
the
azure,
but
we
are
able
to
find
that
in
aws.
C
So
wondering
is
intentional
or
is
we
are
missing
it
out?
That's
a
question
mohammed.
Please
add.
If
you
have
any
question.
A
Okay,
so
I
I
think
I
have
enough
context
for
that.
That
was
intentional
on
microsoft's
edge.
I
think
brandon
had
a
couple
of
blog
posts
that
went
out
a
while
ago.
We
can
maybe
somebody
can
find
the
link
to
them
while
we're
continuing
the
meeting,
and
I
can
put
them
in
the
notes.
But
yes,
azure
historically
had
released
different
vm
images.
Some
were
ending
with
dash
hyphen
with
container
images
and
those
came
with
usually
docker
enterprise
edition
installed
on
the
nodes.
A
C
And
we
also
have
azure
images
with
container
d
instead
of
docker
as
runtime.
D
In
there
sorry
you're
you're,
looking
for
the
images
with
docker
pre-installed,
like
the
vm
host
images.
D
D
No
because
the
runtime
is
now
owned
by
mirantis
and
that's
part
of
the
mirantis
container
runtime
and
it's
licensed
and
installed
by
them.
So
that's
why
those
those
images
are
being
deprecated,
because
it's
no
longer
under
the
jurisdiction
of
microsoft,
and
it's
now
the
licensing
jurisdiction
of
mirantis
so
to
install
that
runtime,
you
have
to
go
through
them
and
I
would
expect
aws
to
follow
suit
with
that
soon
as
well.
So
that's
kind
of
the
situation.
D
There's
no
plan
at
the
moment
to
install
or
to
provide
a
windows,
container
image
containing
container
d
installed.
Pre-Installed
because
there's
you
know
more
than
just
container
d
needed
usually,
and
our
focus
will
be
to
provide
and
better
a
an
installation
experience
for
container
d
that
makes
it
easy
to
get
it
set
up
on
on
windows,
server
images,
so.
D
C
C
D
Yes,
I
mean
so
that's
you
know,
we've
considered
that
as
an
option
and
we're
still
looking
into
that
and
deciding
whether
or
not
it's
it's
there
are
asks
from
folks
to
do
that,
and
it's
good
to
know
that
you're
asking
these
questions
so
yeah
yeah.
E
We
didn't
see
enough
demand
for
docker
ce
based
images,
because,
because
since
kubernetes
has
moved
to
container
d
and
kubernetes
is,
is
the
popular
you
know,
it
is
a
popular
orchestrator
that
everybody's
looking
for
trying
to
use
windows
containers.
E
We
know
there's
some
demand
for
having
container
d
similar
images
that
we're
looking
into
can
we
can.
Can
I
ask
like
what
is
it
specifically
you're
looking
for
docker
ce
to
do
or,
like
you
know,
docker
based
images.
C
You
know,
basically
it
is
kind
of
like
oh,
we
have
been
using
2019
with
our
test
suite,
even
with
our
customer
base.
Oh,
we
are
all
moving
to
2022
or
we
are
in
the
middle
of
continent,
d,
validation
and
I
I'm
not
sure.
Like
myself
and
james,
we
had
a
discussion.
We
we
are
having
some
issue
with
the
cmi.
We
are
trying
to
address
that
so
container
d
is
still
not
available
for
us,
so
the
only
option
right
now
with
the
2022.
C
If
you
want
to
move
away
from
2019
like
2022
bundle
our
ca.
Everything
fails
because
we
expect
a
via
a
docker,
runtime
pre-installed,
which
was
the
case
in
2019.
But
now
until
we
move
to
container
d,
we
want
to
make
sure
we
continue
to
support
a
duck
right.
So
that's
when,
like
in
2022,
it's
not
there.
So
that
is
why
we
are
asking
this
question.
A
Going
to
and
cut
this
time
box
this
I'll,
let
mothers
respond
and
then
ten
boxes
to
get
so
that
we
have
a
chance
for
the
people
who
added
to
the
agenda
to
go.
E
Let's
take
this
offline,
do
you
wanna?
Just
I
mean
I'm
putting
my
email
address
and
a
brandon's
email.
Let
us
do
it
just
you
know
email
it
to
us,
and
then
we
can
have
this
chat
more
offline
and
whatever
the
result
will
we'll
share
with
the
community.
Does
that
work?
It.
A
A
Now
those
are
not
available
on
december
2020..
Okay,
thanks
for
for
asking
that,
though,
we
will
try
and
help
navigate
all
that
as
best
as
we
can
james.
Do
you
want
to
talk
about
the
windows
20
or
the
e
test,
with
kept
c
I'll?
Try
to
keep
this.
B
Pretty
quick,
I'm
going
to
share
my
screen.
Okay,
let
me
just
stop
sharing.
B
So
one
of
the
things
we've
been
working
on
is:
can
everybody
see
the
screen
with
the
test
grid
dashboard?
So
the
one
that's
being
shared
I
can,
but
you
might
want
to
zoom
in
a
little
bit,
it's
kind
of
yeah,
so
one
of
the
things
we've
been
working
on
is
sig.
B
Windows
has
been
using
aks
engines,
cluster,
bring
up
scripts
to
create
clusters
for
all
of
the
performance
tests
and
aks
engine
we've
been
investing
more
in
cluster
api
for
azure
to
be
a
part
of
the
kubernetes
sigs,
and
so
with
124.
B
We've
now
cut
over
our
release,
jobs
that
the
sig
release
team
looks
at
to
use
cluster
api
for
azure,
and
so,
if
you're
not
familiar
with
testgrid,
this
is
the
url
up.
Here
we
have
it
broken
down
into
a
sigma,
sig
window
signal,
which
is
the
one
that
we
monitor
and
we
have
a
triage
meeting
every
tuesday,
actually
right
before
this
meeting.
If
you're
interested
in
joining
that
you
can
please
we
put
the
meeting
in
slack.
B
We
we
go
through
here
and
see
if
there's
any
flakes
and
try
to
track
those
down,
and
then
we
break
it
down
per
release
that
we
have
so
we
have
120,
121,
123
and
then
release
and
then
there's
a
bunch
of
other
related
tests
out
there
that
and
anybody
can
create
tests
out
here.
So
the
one
that
we
switched
over
to
is
this
capsid
container
d
master
job,
and
it's
at
this
point
is
fairly
stable.
B
B
So
all
of
the
test
cue,
the
test.
The
test
configurations
are
in
kubernetes
test
infra
and
we
are
under
the
config
jobs,
kubernetes
sigs,
see,
windows
and
there's
a
bunch
of
all
our
jobs
are
configured
here
so
underneath
the
master
job
here
and
actually
there's
so
much
ammo
that
it's
easier
to
look
at
it
in
the
in
the
github
viewer.
Here.
B
Oh
yeah,
so
there's
a
bunch
of
pre-submits
in
here
and
then
there's
a
bunch
of
periodics
and
so
the
job
that
we
were
just
looking
at.
Is
this
one
here
and
it's
configured
to
use
keuken's
image,
and
then
it
calls
this
script
and
then
there's
some
configurations.
So
this
is
just
telling
it
to
use
those
kubernetes
pre-built
ci
artifacts.
It's
telling
it
to
use
windows
and
container
d
and
the
size
of
the
vms
underneath
the
hood,
but
you
can
find
them
all
in
here.
They're
they're,
all
configured
in
here.
B
This
is
another
example
of
where
we're
using
container
d
on
windows,
server,
2022
and
all
the
configuration
that
is
required
for
that
these.
This
script,
the
ci
conformance
script,
is
documented
over
in
cluster
api
for
azure's
documentation,
the
developer
documentation.
So
there's
a
section
called
conformance
testing
and
there's
a
bunch
of
information
on
how
to
run
this.
How
to
set
the
different
settings?
B
I'm
not
going
to
go
deep
into
this,
but
I
just
want
to
make
it
aware
for
people
that
are
looking
at
debugging
these
things
in
any
capacity
the
cluster
api
for
azure
scripts.
There's
the
ci
conformance
script.
Here,
it's
basically
just
takes
a
bunch
of
those
environment
variables
and
makes
sure
that
everything
is
all
set
up
properly
and
then
it
calls
make
test
windows
upstream
and
eventually
drops
us
into
this
folder
here,
which
is
test
ede
and
there's
a
go.
B
Cluster
eventually
it
it
configures
all
of
the
things
that
are
required
for
running
the
tests,
and
then
it
runs
cubetest,
and
so
this
is
where
all
that's
configured
up.
You
can
go
and
see
that
and
run
it
yourself
in
sig
windows.
We
have
a
dashboard
that
we
track
all
the
flaky
tests.
So
if
you
want
to
go
debug
one
of
these
tests,
you
can
find
a
test
from
this
dashboard.
B
B
You
can
also
find
that
in
kubernetes
by
searching
for
label
sig
windows
and
kind
flake,
and
then
when,
if
you
go
to
the
dashboard
and
you're
on
the
container
d
cap
c
jobs,
you
can
click
into
one
of
these
ones.
That's
failing
you
just
click
on
it.
It
brings
you
in
there's
kind
of
overwhelming
here,
but
up
in
the
top
right,
you
can
see
the
job
history.
You
can
see
the
actual
prowe
yaml.
This
is
very
similar
to
what
we
were
just
looking
at
in
the
test.
B
Infra
job,
so
you
can
see
that
all
those
those
variables
got
passed
through
and
then
you
can
this.
This
is
the
link
to
bring
it
back
to
test
grid.
Here,
we'll
see
that
one
of
these
tests
failed.
We
get
some
information
here.
This
isn't
super
helpful,
but
if
you
go
into
the
standard
out,
there
is
a
whole
output
of
the
test
run,
including
it
collects
all
the
events
for
the
failed
test.
So
you
can
see
that
this
happened
to
be
a
the
lifecycle
hook,
failed
for
some
reason.
B
B
This
is
a
collection
of
all
the
logs
that
were
run
so
the
entire
build
log
is
here
from
the
point
where
it's
actually
bringing
up
the
cluster
to
running
all
the
individual
tests
in
the
artifacts
folder.
There
is
the
main
folder
for
cluster
api
for
azure.
Is
this
clusters
folder
the
bootstrap
folder?
Is
the
bootstrap
cluster
if
you're
familiar
with
cluster
api
for
azure
there's
a
bunch
of
information
here
we
also
collect
all
the
cappy
controller
manager
logs.
B
We
collect
what
was
actually
deployed
onto
the
cluster
and
the
the
clusters
cni
select,
but
the
the
interesting
one
is
actually
probably
this
capsi
conf.
This
is
the
the
workload
cluster
that
was
actually
created
that
the
tests
are
being
run
against.
Here
we
collect
the
azure
activity
logs,
so
anything
that
happened
in
azure
like
creating
the
vms,
creating
the
networks.
All
that
stuff
is
all
collected
there,
and
then
we
collect
the
cube
system
things.
B
B
For
that
thing,
and
then
also
q,
proxy
log,
so
in
here
is
the
q
proxy
windows
logs
and
we'll
see
that
we've
got
the
entire
q
proxy
for
the
entire
run.
So
you
can,
if,
if
something
happened,
q
proxy
you'll
be
able
to
look
that
up
and
then,
if
we
go
back
over
to
machines,
we'll
see
we
have
the
control
plane
logs.
There's
things
like
the
cubelet
log
for
those
on
the
cloud
init
log
and
then
for
windows.
B
We've
got
a
whole
bunch
of
stuff,
so
we
collect
all
the
services
we
collect
the
networking.
We
collect
what
interfaces
we're
running
any
containers
that
we're
running
at
the
end
of
the
job.
So
if
some
container
got
stuck,
we
collect
the
cni.
This
is
the
actual
cni
executable
binary
log,
and
you
can
see
there's
quite
a
bit
of
details
in
here
and
the
cubelet
log
is
also
here,
which
is
probably
where
you'll
find
be
looking
for
most
errors
and
flakes
of
that
sort.
B
So
I
know
that
was
a
really
really
fast
overview,
but
I
wanted
to
get
it
just
share
it
and
have
it
recorded
so
that
we
could
point
folks
back
at
this,
and
with
that
I
will
stop
sharing
my
screen
and
answer
any
questions
and
then
hand
it
back
to
march.
A
Yeah
links
in
the
notes,
I'm
sorry
go
ahead.
If
somebody's
asking
a
question.
F
Yeah,
sorry,
I
didn't
quite
see
the
container
deal.
It
might
have
skipped
for
me
in
the
screen
at
least,
but
it
is
connected.
B
It
is
and
underneath
that
cube
system
logs,
there's
a
container
d
log
system
collector
so
the
way
we
collect
those
is
via
a
host
process,
container
log
that
listens
to
the
etw
events
and
then
it
out
outputs
that
to
standard
out,
and
so
we
collect
it
all
there
and
then
we
can
filter
and
turn
it
on
higher
verbosity
and
things
like
that.
If
we
need
to.
F
And
in
addition
that,
in
addition
to
that,
I
might
be
misremembering
things.
But
when
wasn't
there
a
discussion
at
one
point
to
also
collect
logs
from
from
him
as
well,
because
it
might
be
useful
for
a
more
deep
dive
that
we've
been
seeing
and
couldn't
really.
B
Yeah,
you
cut
up
a
little
bit,
but
I
think
you
asked
for
the
hcs
shim
logs
they're,
also
in
they're,
also
in
that
container
d
glogger.
So
it
listens
to
the
hcs
shim
etw
events
as
well,
so
we
collect
all
of
that
within
and
it's
all
interlaced
so
that
you
can
you
get
the
whole
context
of
what's
happening.
A
Awesome
thanks
yeah.
I
think
james
did
a
demo
a
couple
of
meetings
ago
about
that
event,
flow
log
or
container
image
that
he
built
up.
It's
it
lives
in
sig
windows
tools.
It's
under
the
host
process,
folder
that,
instead
of
just
collecting
text
logs
for
these
we're
able
to
hook
into
like
wire
up
an
etw
event,
listener
and
I'll
put
those
to
the
container
images
text
and
collect
those.
That's
definitely
worth
checking
out.
A
I
think
we
highlighted
some
of
them
here,
but
if
anybody
has
any
questions,
don't
hesitate
to
reach
out
to
us
for
in
sync
testing,
sig
testing,
the
sig
has
done
a
couple
of
cube
contacts
about
how
all
of
this
works.
So
that's
also
interesting
how
how
proud
actually
works.
We
just
kind
of
touched
the
tip
of
the
iceberg
about
how
the
configs
and
prowe
are
set
up
too.
So
a
lot
of
moving
pieces
and
a
lot
to
learn
here
but
yeah.
Thank
you,
james
for
the
overview.
A
A
We
do
have
a
meeting
every
other
thursday
at
9
30
pst
around
this
for
folks
who
want
to
join
it's
on
the
should
be
on
the
same
calendar,
but
I
can
give
a
quick
overview
of
the
status
here.
So
there's
one
big
issue
that
lives
in
the
kubernetes
website
repository
that
kind
of
highlights
a
lot
of
kind
of
just
thoughts
right
now
about
all
the
different
things
that
we
want
to
do
to
update
the
windows
documentation.
A
The
big
ask
from
a
lot
of
people,
including
the
six
docs
team,
is
to
break
up
the
big
intro
to
windows,
kubernetes
page,
which
most
people
land
on
it's
been
kind
of
some
content's
been
slowly
getting
removed,
but
there's
a
plan
highlighted
in
that
issue
about
how
to
how
to
break
it
up.
So
we've
got
like
bullet
items
for
the
individual
topics
that
we
want
to
move
out,
where
we
want
to
move
them
and
pull
requests
open
for
a
lot
of
them
too.
A
So,
and
another
thing
I'll
highlight,
is
that,
where
possible,
as
we
peel
information
off
of
that
that
page,
we
are
trying
to
put
it
next
to
the
equivalent
next
to
like
the
discussion
topics
that
are
not
focused
on
the
specific
operating
system.
Networking
had
a
lot
of
examples
of
that.
So
I
can
just
highlight
that
real,
quick,
for
example,
when
we
talk
about
you,
know
dns
resolution
behaviors
on
windows.
A
We
are
just
adding
that
to
the
dns
pod
service
page
in
the
kind
of
in
the
already
existing
networking
same
with
dual
stack
support,
and
this
is
fully
supported
by
the
the
say,
docs
and
the
enhancement
team,
to
do
that,
and
we
think
it's
really
going
to
help
make
windows.
Look
more
like
a
appear
to
make
windows
like
appear
to
linux
in
in
the
documentation,
rather
than
having
to
caveat
on
a
special
page.
A
All
the
differences
in
windows
just
have
people
say
I
want
to
learn
about
dns
and
then
say
like
see
that
information
there
there's
still
a
lot
of
work
that
needs
to
be
done,
requests
that
need
to
be
opened
for
even
moving
for
both
moving
content
and
for
just
going
and
reviewing
the
pages
that
we
do
have
if
anybody's
interested
feel
free
to.
You
know,
comment
on
the
issue
open
up,
open
up
a
pr
and
get
involved
with
that
too.
A
So
I
will
ask
if
there's
any
questions
quickly
and
then
hand
it
over
to
the
for
the
next
demo.
A
Oh,
I
will
add
to
that
we're
targeting
all
of
this
work
in
the
dev
124
branch
of
the
kubernetes
website.
So
we
are
hoping
to
have
all
of
this
done
by
the
time
124
releases
and
then
that
will
just
kind
of
be
when
all
those
changes
go
live.
A
Okay
shinji,
it
looks
like
you
wanted
to
present
something,
but
you
said
10
minutes.
I
think
there's
four
or
five
minutes
left.
I
would
do
you
want
to
present
now
or
push
it
and
go
first
next
week.
It's
totally
up
to
you.
G
I
think
I
can
like
do
that.
Do
that
now
and
make
a
quick
demo
to
try
to
finish
that
in
five
to
ten
minutes.
Yeah.
G
Okay,
so
basically,
what
I'm
doing
is
that
now
we
provide
a
way
to
run
the
windows
like
test
cases
in
the
operational
redness
folder,
so
we
created
two
local
object:
operational
redness
to
run
that
and
basically
what
I'm
doing
here
is
that
I
add
another
like
test
case
into
the
gmsa
full
and
we
call
this
function
from
the
operandi
of
redness
and
in
order
to
do
that
so
here's
my
environment.
I
have
my
active
directory
server
running
here
and
I
have
this
node.
G
It's
a
windows
worker
node-
and
I
add
this
to
the
domain
of
this
id
server.
So
in
this.
So
in
addition
right
here.
G
We
can
see
that
I
add
so
so.
This
is
the
node
that
I
mentioned
previously,
and
this
is
the
member
of
this
service
group
and
this
services
group
is
linked
with
this
gmsa
account,
so
they
are
linked
together
and
and
and
then
I
create
another
folder
here,
and
I
should
let
me
do
like
this
first,
so
this
folder
is
shared
with
the
group.
G
So
you
can
see
that
here,
so
this
security
group
has
the
permission
to
read
and
write
files
into
this
folder
and
I
create
a
pod.
The
part
I
created
here
is
linked
to
this
default
tool
is
jointed
to
this.
G
Let
me
find
it
here,
so
I
have
like
this
setup
here
so
so
after
I
create
this
part,
I
should
have
the
ability
to
read
and
write
files
into
this
into
this
right
into
this
folder,
and
so
in
order
to
run
this
test
from
the
uprightness.
We
have
like
some
input
like
this.
So
now
we
have
two
tasks
in
the
gmsa
tests.
G
One
is
called
a
work
end
to
end,
which
is
same
as
here
so
that
so
work
end
to
end
and
another
one
which
is
I
added,
is
called-
can
write
right
file
into
the
remote
folder,
so
we
can
run
both
of
them
here
and
in
order
to
trigger
this
test.
We
need
some
command
like
this,
so
the
category
is
active
directory
and
both
of
these
test
cases
have
this.
It's
in
this
step
it's
in
this
category,
so
they
can
run
at
the
same,
so
the
both
of
them
will
run.
G
So
I
click
so
and
now
this
test
is
going
to
run.
So
what
it's
going
to
do
is
that
when
we
create
this
part,
let
me
find
this
so
I'm
going
to
create
this
part
with
gmse
and
if
we're
going
to
test
the
storage.
I
read
this.
G
I
should
be
able
to
like
cut
this
file,
cut
this
file
and
get
the
content
so
this
so
that
this
string
should
be
inside
the
output.
So
that
is
my
test
and
we
can
see
that
it's
finished
and
there
are
two
tasks
running
here.
The
first
one
is:
is
this
test
case
to
test
the
to
test
the
file,
and
we
can
see
that
so
this
is
the
contents,
the
the
contents
of
the
file
and
also
it
shows
up
here.
G
G
E
A
E
A
A
E
You
see
because
mark
I
was
more
thinking
about,
like
being
you
know,
remember
how
we
used
to
do
gms
and
then
how
you
know
it
has
evolved
kind
of
things.
So
so,
basically,
that
expectation
is,
however,
you
do
it
at
the
end
of
the
day.
As
long
as
you
have
the
credentials,
this
should
be
able
to
test
right.
Okay,.
G
A
G
A
No
worries,
I'm
sure,
we'll
figure
that
out
too.
Thank
you
for
demoing
this
no
problem
here.
Okay,
so
I
think
that's
going
to
conclude
today's
community
meeting
we'll
see
everybody
in
the
future,
hopefully
next
week,
thanks
all.
Thank
you
bye.
Thank
you,
james.
I
think
you
need
to
stop
the
recording.