►
From YouTube: Kubernetes SIG Windows 20220315
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
A
The
only
announcement
that
I
have
added
is
that
the
kubecon
eu
schedule
is
live
if
you're
interested
in
attending
it's
going
to
be
a
hybrid
event.
Again,
you
know,
take
a
look
at
the
schedule,
see
if
you
can
attend
virtually
or
in
person
if
you
want
and
it's
a
great
way
to
kind
of
connect
and
network
and
learn
about.
What's
going
on
in
the
project,
there's
the
other
and
there's
a
lot
more
information
about
logistics
and
everything
up
here.
A
So
take
a
look
at
that
and
we're
sig
windows
is
doing
a
maintainer
track
session
like
we
do
at
all,
qcon.
That's
a
good
way
to
see
like
a
brief
overview
of
kind
of
everything
that
we've
been
doing
for
the
past
six
months
or
so,
and
just
stay
up
to
date.
But
does
anybody
have
anything
else
announcement-wise
to
to
call
out
here.
A
I
will
reiterate:
code
freeze
for
124
is
in
two
weeks
two
weeks
from
today.
So
let's
just
let
us
know
if
you
have
any
you're
having
trouble
getting
any
prs
merged
or
if
there's
any
issues
that
come
up.
A
Okay,
next
one,
I
wanted
to
start
start
this
up
again,
but
each
meeting,
let's
I'd
like
to
open
it
up,
see
if
there's
any
new
contributors
who
want
to
introduce
themselves
and
just
say,
hi
or
ask
questions
or
just
give
a
quick
intro.
B
Hey
everybody,
I
think
I
can
say
hi,
I'm
not
a
regular
on
this
meeting,
I'm
christian
glombeck
I
work
at
red
hat
and
I've
been
working
on
the
no
blog
viewer
feature
together
with
arvind
and
that's
what
I'm
I'm
gonna
talk
about
today.
Again,
all
right,
you
know
so
just
saying
hi
from
edward
germany.
A
Oh
hi
cool,
hi
yeah,
I
you've
joined
a
couple
of
times,
but
it's
always
good
to
introduce
yourself
again
cool
anybody
else
wants
to
say
hi.
If
not,
we
can
go
into
the
agenda.
C
My
name
is
wendy
just
joining
in
from
nairobi.
This
is
the
first
time
here
I'm
working
out
of
the
microsoft
adc
center
in
kenya,
so
everyone.
A
If,
if
you'd
like
you
can,
are
you
familiar
with
how
to
get
access
to
the
meeting
notes?
You
can
you
can
add
your
name
to
the
agenda
too.
It's
a
good
exercise
too.
In
order
to
get
right.
Access
to
the
dock,
you'll
need
to
join
the
sig
windows,
mailing
list
and
there's
instructions
for
that
somewhere
here,
where's
the.
A
D
A
So
please
join
this
google
groups
to
get
access
to
the
doc.
A
A
Thanks
for
coming
to
say,
hi,
okay,
if
there's
anybody
else
feel
free
to
interrupt
me,
but
if
not
I'm
gonna
start
just
going
to
the
agenda
first.
One
is
one
that
I
added.
I
think
I
need
danny
cantor
in
the
call
so
I'll
reach
out
to
him
afterwards,
but
this
pr
was
opened.
D
But
the
one
thing
one
thing
I
had
there
was
I
I
think
this
particular
feature
if
extended
could
help
out
with
some
some
various
things,
like
the
quality
of
service
and
system
reserve
and
potentially
even
enforcing
node
allocatable
in
the
future,
and
so
I
think
it's
okay
as
it
is,
but
eventually
I'd
like
to
dive
deeper
into
how
do
we
use
this
type
of
thing
across
multiple
components
and
and
deliver
some
of
these
features
that
have
been
missing
from
windows
for
a
little
while.
A
A
A
A
A
Yeah
christian,
do
you
want
to
do
you
want
to
share
a
screen,
or
I
can?
I
think
I
don't.
B
Have
anything
really
to
demo?
I
think
if
you
just
yeah
show
the
pr
here.
I
really
just
wanted
to
plug
the
pr
again.
We
need
some
more
reviews
from
a
couple
of
six.
Actually,
I'm
not
sure
somebody
from
sick
windows
has
has
reviewed
it.
Yet
I
really
just
wanted
to
raise
awareness.
We
we
are
planning
to
get
that
in
for
124
and
I
think
we're
we're
on
a
good
path
there.
But
yeah.
B
There
have
been
some
some
minor
comments.
I've
just
rebased
it
today
again,
so
if
any
of
you
could
take
a
look,
that'd
be
great.
I've
also
tagged
a
few
folks,
specifically
already
and
yeah,
for
those
of
you
who
don't
know
what
this
is
about.
B
This
is
a
feature
that
lets
you
access
the
the
nodes
logs
the
service
logs
on
both
linux,
as
well
as
windows,
nodes
through
the
from
a
cube
ctl
command
through
cubelet
and
qubit,
will
act
as
a
as
a
streaming
server
for
those
locks
and
read
the
the
logs
from
the
what
is
it
the
application
log
provider
on
windows?
I
think
the.
B
Yeah,
exactly
that
could
also
be
extended
to
the
system
lock
probe,
there's
like
log
providers,
I'm
not
too
familiar
with
windows,
even
though
I
I
implemented
that
part
in
the
pr
it
does
work
for
our
purposes
in
openshift.
We
already
have
that
in
our
openshift
giant
and
tubelet.org
and
yeah.
We.
B
A
Yeah
a
couple
things
arvind
did
demo
this
on
the
january
18th
community.
B
A
There's
the
recording
is
available
on
our
youtube
playlist,
if
anybody's
interested
in
checking
that
out
good
overview
of
the
dem
of
the
feature.
I
did
review
this.
I
had
the
one
comment
about
making
about
the
duplicate
logs
and,
as
I
see
that
the
I
was
out
of
office
yesterday,
so
I
didn't
have
a
chance
to
look
at
this.
That
was
my
really
only
feedback
was.
B
I've
actually
changed
it,
so
it
looks
at
at
both.
First,
it
looks
at
I
think,
info
dot.
Anything
so
it'll
only
get
the
past
logs
if
it
finds
them
and
then,
if
it
doesn't
find
any
any
like
time
stamp
blocks,
it'll
look
for
the
the
one
that
ends
with
info
and
there's
no
no
dot
and
nothing
afterwards.
B
The
idea
I
was
discussing
this
was
with
aaron
yesterday
and
we're
thinking
it
might
suffice
to
just
kind
of
get
the
latest
if
it's,
if
it's
always
named
that
way,
if
all
the
services
put
the
latest
one
in
this
without
the
timestamp
just
with
the
info
at
the
end
of
the
file
name,
otherwise
we
could-
and
this
is
the
way
I've
implemented
it
now.
I've
changed
that.
B
I
will
first
look
at
all
the
timestamp
blocks
and
if
there
are
no
timestamp
blocks,
then
we'll
try
to
get
the
latest,
but
I'm
not
I'm
not
sure
how
the
behavior
is
there,
whether
all
of
these
services
actually
do.
Is
it
the
hard
link
or
assembling
for
the
latest
one.
A
A
That's
any
component
that
uses
the
k
log
package
so
most
of
the
kubernetes
components,
but
other
things
like
container
d
use
logarithms,
which
has
slightly
different
behavior
so
might
I'll
take
another
look
too,
might
be
good
to
do
the
behavior
that
you
mentioned,
where
we'll
check
to
see
what,
if,
if
there's
the
the
files
there
and
then,
if
not,
we'll,
look
for
the
dot
info.
B
At
least
that
will
get
us
out
of
the
situation
where
we
would
see
the
latest
vlogs
twice
the
way
it
is
now
so
yeah.
If
you
could
check
that
part
of
the
code,
that'd
be
awesome
and
yeah
I'll
definitely
be
addressing
all
those
comments
that
are
gonna
be
coming
in
over
the
next
couple
of
days
and
rebasing.
This
as
well
yeah.
A
Okay,
yeah,
I
haven't,
I
haven't
checked
again:
how
does
ryan
phelps
he
was
he
the
primary
reviewer
from
node.
A
A
A
B
A
E
A
Feel
free
to
ping
me
on
slack,
if
I'm
not
responsive
enough
here.
Sometimes
I
miss
github
comments,
just
because
I
get
so
many
don't
feel
bad
about
pinging
me
on
slack
and
we'll
figure
out
what.
A
Get
who
all
needs
to
review
to
get
this
done
so,
okay,
yep
and
thanks
for
coming
in
well
for
working
on
this.
Firstly
and
then,
oh
yeah,
all
the
work
in
come
here.
B
Yes,
it
is,
we
still
have
to
open
a
docs
pr.
I
think
we
haven't
done
that
yet
I
think
that
has
to
just
be
open
before
march
29th
yeah.
We
do
have
that
on
our
on
our
list.
A
Yeah
and
the
the
docs
peer
that
gets
opened,
they
it
doesn't
need
to
have
any
content
in
it,
yet
they
just
want
to
they.
They
track
to
make
sure
that
for
new
features,
especially
ones
that
expose
user
facing
scenarios
that
they
know
who
the
primary
point
of
contact
who's
going
to
be
making
the
docs
vr
is
so
they
just
like
a
draft
vr
open
for
that
cool.
Thanks
for
that.
B
B
A
I
added
some
other
comments
on
the
to
the
meeting
too.
I
was
hoping
danny
would
be
here
to
discuss.
I
can
go
over
them
real
quickly
and
if
anybody
else
has
any
comments,
do
this
so
in
for
for
docker
shim
on
on
windows,
there's
support
for
device
assignment
on
windows
which
allows
us
to
map
gpus
into
containers
and
then
run.
You
know,
gpu
accelerated
workloads.
A
That
work
was
only
currently
works
for
docker,
shim
and
so
there's
a
couple
of
pr
are
there's
a
couple
of
issues
that
are
open
discussing
behaviors
around
how
to
get
this
to
work
for
container
d,
including
a
I
think,
a
pr
into
container
d
that
map
or
that
that
adds
this
support
to.
So
I
was
wanting
to
just
get
agreement
with
everybody
from
sig
windows
and
how
this
works.
A
The
biggest
I
think
point
of
discussion
here
is
captured
in
this
issue,
where
the
way
that
linux
and
windows
specify
devices
in
the
oci
spec
are
different.
I've
highlighted
that
here
on
linux
there's,
you
know
a
host
path,
a
container
path
and
then
permissions
that
get
assigned
on
windows.
The
devices
are
assigned
or
are
on
windows,
there's
just
a
device
id
or
there's
a
device
class
and
a
device
id.
A
So
those
work
a
little
bit
differently
and
how
the
docker
shim
implementation
for
getting
device
assignment
to
work,
kind
of
worked
as
it
just
concatenated
the
device
class
and
device
id
and
stuck
that
into
the
host
path
on
the
machine
and
docker
shim
knew
how
to
parse
that
out
or
docker
shim
and
actually
docker
knew
how
to
parse
that
out,
and
so
it's
kind
of
overriding
or
kind
of
forcing
the
linux
behaviors
or
the
windows
behaviors
to
comply
with
the
linux
kind
of
class
definitions
or
here
so
there's
a
big
discussion
here,
and
I
think
that
yeah
so
here
you'll
see
here's
an
example.
A
A
Unpack
that
id
type
or
device
class
and
then
id
to
do
the
same
behavior
as
the
as
docker
was
doing
too
so
since
that
already
merged
it
might
make
sense
to
just
do
the
same
thing
in
the
cubelet
to
map
that
in
there.
But
I
was
hoping
to
have
a
discussion
with
some
folks
here.
If,
if
we
don't
do
that,
we
need
to
update
the
fields
in
the
cri
api
to
to
do
that,
which
could
you
know,
take
potentially
a
couple
of
releases
to
propagate
to
get
the
cri
changes.
A
D
Liked
I
do
have
quick
questions,
so
we
did
have
a
test
that
was
running
against
the
docker
shim
side
of
things.
With
this
prs
that
went
in.
Does
that
mean
we
can
now
potentially
enable
that
test
for
continuity
as
well.
A
I
don't
think
so
because-
and
I
need
to
double
check
this-
I
think
that
the
continuity
changes
just
merged.
We
still
need
to
make
changes
in
the
cubelet
to
map
the
to
map
like
tough
to
populate
those
fields
and
to
do
the
same
change.
I
think
that
the
changes
that
put
the
device
class
and
id
into
the
the
the
host
path
field
of
the
devices
on
the
container
create
calls
in
docker
shim
was
it
was
it
was
all
in
docker
shim.
A
We
need
to
go
and
wire
up
like
all
of
that
into
the
cr
api
fields.
So
I
was
if
we
do
want
to
go
that
route.
I
was
hoping
that
we
could
also
get
that
those
changes
merged
in
124
and
then,
after
that,
after
the
cubelet
knows
how
to
like
request
those
devices
get
assigned
to
containers,
then
we
can
open
up
or
then
we
can
start
running
those
pr
jobs
or
those
are
they.
I
think
those
periodic
jobs
to
test
this.
A
I
I
need
to
double
check.
I
think
I
I
thought
that
the
code
that
was
adding
that
was
in
a
linux
specific
file
somewhere
in
the
cubelet,
but
we'll
check
that
so
yeah
I
I
hope
I
hope
we
can
just
light
it
up
and
like
wire
everything
up
and
then
start
performing
some
validation,
but
I
think
we
may
need
to
make
a
couple
changes
like
that.
A
All
right
that
is
all
that
I
see
on
the
agenda.
Does
anybody
have
anything
to
to
add?
If
not,
we
can
end
the
meeting.
Seven
minutes
early.
F
Going
once
so
one
question
regarding
2022
support:
will
you
guys
notify
when
it
officially
gets
released
here
in
this
forum.
F
A
F
Okay,
because
we
are
trying
to
test
everything
with
the
private
bill,
whatever
was
given
by
david
right.
So
we
just
want
to
make
sure
that
we
replace
that
with
the
official
one.
A
H
H
F
For
c
is
like
april
4th,
or
that
is
I
remember
david-
the.
H
H
A
Okay,
we'll
once
it's
released,
two
we'll
make
an
amount.
We
can
make
an
announcement
or
inform
everybody
here.
G
F
Oh
yeah,
I
know
I
I
think
we
are
good
like
what
we
did
like
the
basic
functionality,
whatever
issue
we
run
into
it
all
worked.
But
what
to
for
us
to
make
our
ca
run
against
this?
We
try
to
build
a
private
golden
image,
the
vsphere
and
we
are
trying
to
run
it
in
our
ci
to
make
sure
all
other
traffic
bases
are
passing.
That's
what
we
are
trying
to
do
so
for
that
we
were
able
to
build
last
week
with
the
patch.
So
we
will
let
you
know
like
in
case.
I
Cool
sounds
good.
I
had
a
quick
idea
that
I
was
thinking
of.
I
didn't
write
it
in
the
agenda
feel
free
to.
A
I
I
I
was
thinking
what,
if
you
know,
we've
been
doing
the
pairing
sessions
and
I
think
those
were
going
kind
of
well
for
a
while,
and
then
I
I
feel
like
the
people
who
sort
of
wanted
to
get
involved
have
gotten
involved,
and
I
was
thinking
maybe
to
sort
of
put
some
effort
into
windows,
networking
like
sort
of
migrate
away
from
the
pairing
sessions,
unless
people
really
want
to
keep
doing
them
and
I'm
totally
down
for
to
keep
doing
them
or
whatever.
I
But
I
was
thinking
to
sort
of
invest,
sort
of
in
a
windows,
sort
of
networking
subproject
or
something
of
that
sort
right
and
and
then
sort
of
instead
of
doing
the
two
hour
thing
that
we
do
nowadays,
we
just
do
the
regular
sig
windows
meeting
we
sort
of.
We
don't
do
the
pairing
anymore
unless
people
specifically
want
to
stay
after
and
they
could
just
ask,
and
we
could
always
do
that
if
they
want
to
and
then
what
we
do.
I
Is
we
sort
of
what
I
would
do
at
least
is
sort
of
spend
that
time
running
a
sig,
sig
windows,
networking
sub
group
or
something
like
that,
and
I
maybe
do
it
at
a
time.
That's
more
asia
friendly,
I'm
not
sure.
Yet,
just
a
thought.
I
I
don't.
I
I
don't
know
because
I
do
feel
like
that's
something
I
work
on
every
week
and
I'd
be
happy
to
just
work
on
that
in
public
with
people,
and
I
think
it
would
give
people
a
forum
to
come
and
join
and
collaborate
as
opposed
to
us
having
to
schedule
meetings
and
so
on
and
so
forth,
and
I
feel
like
there's
a
lot
of
networking
work
going
on
in
the
windows
space
lately.
I
Cool,
so
so,
if
there's
no
major
objections,
I'll
sort
of
get
the
ball
rolling
on
it,
but
then
I
guess
the
question
I'd
have
is:
do
people
want
to
keep
doing
the
pairing
and
if
so,
would
somebody
else
want
to
like
run
that
or
do
we
want
to
just
stop
doing
pairing
and
it's
just
an
as
needed
thing
if
people
want
to
stay
after
they
can
and-
and
we
always
give
them
the
opportunity
to
do
that.
A
I
Okay,
I'm
gonna
fish
around
and
see.
Okay,
I'm
gonna
ask
some
people
and
see
who
might
be
willing
to
come.
Obviously,
I'd
be
willing
to
come.
I
I
think
we
could
probably
get
some
of
our
antria
counterparts
to
come
and
I
think
maybe
we
have
some
friends
over
at
microsoft
that
are
that
are
working
on
this
stuff,
so
maybe
they'd
be
down
to
come
also,
and
I'm
gonna
ping
people
this
week
and
I'll
start
looking
into
this
and
then
if
anybody
wants
to
come
and
be
a
part
of
that
group
and
help
me
run
the
group
just
reach
out
to
me
on
the
sig
windows
slack
channel-
and
you
know
I
think,
if
we
can
get
say
four
people
that
are
interested
in
coming.