►
From YouTube: Kubernetes SIG Windows 20200630
Description
Kubernetes SIG Windows 20200630
A
Hello,
everybody
and
welcome
to
another
sick
windows
meeting
it's
the
30th
of
june
last
meeting
for
june,
as
always,
is
a
recorded
meeting.
So
please
adhere
to
the
code
of
conduct.
If
there's
any
questions,
concerns
that
you
don't
want
to
be
on
the
recorded
call.
Just
please
find
us
in
private
on
slack
or
the
mailing
list,
all
right.
Let's
get
started.
A
Our
agenda
today
guess
we'll
start
with
james.
He
added
some
stuff
around
the
windows,
testing,
dashboard.
B
Cool
yeah,
so
I've
been
working
on
adding
the
2004
support
for
the
aks
engine
tests
and
when
I
was
doing
that,
I
noticed
that
our
dashboard
for
sig
windows
was
had
a
lot
of
stuff
in
it,
and
so
I
kind
of
poked
around
and
started
to
figure
out
how
we
could
potentially
reorganize
that.
So
I'm
kind
of
one
of
the
people
that
used
to
kind
of
play
with
things.
So
I
started
to
just
mess
around
with
the
dashboard
code
and
opened
up
kind
of
a
potential
pr
for
that.
B
A
B
This
up
as
an
official
pr
and
get
some
leave
some
comments
and
things
like
that.
But
right
now,
there's
no
like
tabs
for
us
to
be
able
to
bucket
things
into
okay,
and
so,
if
we
scroll
down
a
little
bit
further,
there's
one
file
in
particular.
That
shows
the
the
way
that
it's
bucketed
together.
B
Yeah
so
the
green
part
on
the
right
yeah
there.
It
is
so
you
can
build
it
into
different
types
of
dashboard
names,
and
so
that
way
we
can
kind
of
organize
the
various
tests
that
we
have
across
various
areas,
and
I
just
kind
of
threw
some
in
there
to
get
the
conversation
started
but
wanted
to
know
what
folks
thought.
A
C
Yeah
definitely
our
our
tester
is
already
scattered
around
the
various
dashboards.
So
a
little
bit
of
organization
is
excellent.
D
C
E
Just
a
small
question:
will
some
of
the
dashboards
be
multiple
different
locations
as
well?
For
example,
let's
say
see:
windows,
asia,
container
d:
will
that
be
also
on
see,
windows,
azure
and
c
windows,
container
d
or
yes,.
B
B
There's
like
sigma,
does
gce,
but
then
there's
also
sig
windows,
pre-submit,
there's
a
few
other
examples
throughout
the
code
base
that
a
few
of
the
changes
so
so
essentially.
B
B
A
It'd
be
useful
to
also
see
what
the
dashboards
look
like
once
you.
Actually,
I
guess
there's
no
way
to
pre-stage
some
of
that
right.
B
Okay,
so
we
did
this
for
azure's
dashboards
about
three
or
four
months
ago,
so
you
can
go
kind
of
look
at
those.
I
think
a
lot
of
the
other
groups
are
also
organized
in
this
way,
some
of
the
other
sigs
as
well.
So
you
can
kind
of
go.
Look
at
those
to
see
what
it
might
look
like.
A
A
All
right
james
here
you
have
the
next
one
as
well.
2004
testing,
update.
B
Yeah
last
week
we
mentioned
potentially
adding
it
for
119,
so
I
started
to
add
at
least
the
test
for
aks
for
the
azure
side,
which
has
our
1
18
119,
and
so
I've
got
the
it
all
in
place.
We
were
waiting
for
a
release
of
aks
engine
which
happened
last
night
so
and
then
the
2004
images
needed
to
be
pushed
to
the
upstream,
and
I
think
claudio
was
going
to
do
that
today
or
sometime
soon.
B
Yeah
I
didn't
do
gce.
I
think
I'm
not
super
familiar
with
the
gce
side
of
things,
and
I
think,
let's
see
peter
on
today
had
opened
up
an
issue
for
that.
Okay,.
E
Yeah
those
images
are
already
built
currently
pushing
them
from
my
registry
to
the
e3
team,
one
it
should
be
finished
in
two
or
three
minutes.
I
think
you
can
go
ahead
with
that.
A
A
D
Been
a
couple
of
months,
I
think,
since
we
discussed
this
last
time
so
remember
last
time
you
talked
about,
I
think,
deep,
the
poor
you
and
we
all
agreed
that
we
should
have
a
separate
me.
We
should
continue
to
review
the
backlog
in
some
way
of
our
farm,
not
necessarily
in
this
meeting,
but
it's
good
to
keep
an
eye
on
it
and
it's
been
quite
a
while,
since
we
like
organized
the
backlog
again.
So
thanks
to
james
and
mark,
we
looked
at
it
and
we
added
a
new
column
v120.
D
The
new
way
we're
going
to
do
this
is,
if
you
look
at
it
in
review
119,
we
have
put
the
enhancements
on
top
and
then
the
bugs
on
the
bottom,
which
kind
of
shows
you,
like
you
know
somebody
joining
the
team,
the
big
bucket
items
more
like
a
road
map,
so
enhancement.
I
can
consider
that
and
then
you
know
the
the
bottom
we
have
done.
A
couple
of
you
know
clean
up
here
and
there
as
well.
So
today,
just
this
is
fyi
for
the
community.
D
I'm
gonna
be
sending
out
a
doodle
on
a
bi-weekly.
Like
you
know,
every
two
weeks
they're,
we
are
proposing
a
meeting
I'll
send
out
the
agenda.
Basically,
we
just
go
through
the
backlog
on
a
regular
basis,
I'll
send
it
out
in
slack,
please
respond
to
it
and
propose
a
time
that
works
for
all
of
you,
we're
hoping
that
most
of
you
can
join
again.
The
purpose
is
to
keep
an
eye
on
like
where
we
are
where
we
are
going,
keep
things
a
little
bit
organized.
D
This
is
some
of
the
feedback.
Michael
probably
you've
heard
we
were
talking
with
a
couple
of
folks
from
atlanta
and
a
couple
of
other
folks
who
wanted
to
join
sig
windows.
They
were
like,
we
don't
know
where
what
is
happening
like
what
are
you
guys
working
on
today?
I
think
these
kind
of
efforts
are
going
to
help
them.
A
D
A
D
So
so
that's
it
that's
the
update.
The
rest
I
was
thinking
of
whatever,
like
I
I'm
not
like
we're
not
doing
too
much
we'll
do
it
in
the
meeting
in
the
first
meeting,
we're
gonna
go
through
the
board
and
like
take
all
the
session.
D
A
Yeah,
you
know,
obviously,
would
I
I
personally
don't
do
a
great
job
of
looking
into
it.
I
have
a
couple
of
suggestions
when
you
have
the
doodle.
Let's
add
it
to
this
see
windows,
meeting,
notes
and
don't
choose
the
date
yet
until
everybody
has
a
chance,
because
most
people
don't
pay
attention
to
slack,
but
they'll
see
this
meeting
some
of
them
viewed
offline.
So
you
can
add
the
link
here.
So
when
you
have
it,
then,
maybe
after
next
tuesday,
like
the
wednesday
next
wednesday,
you
pick
the
date
that
way.
C
A
Give
everybody
a
day
to
to
do
that.
The
second
thing
is
one
of
the
things
I've
seen
that
works
very
well
with
other
communities
is
actually
tagging
issues
as
a
good
first
issue
for
someone
that
wants
to
get
started.
What
are
some
issues
that
are,
you
know,
like
usually
contributors
start
by
first
attending
meetings,
then
they
play
with
the
product
and
they
look
at
dogs.
Maybe
they
make
a
pr
to
dogs
and
then,
after
that,
they
say.
A
Okay
now
it's
time
for
me
to
fix
a
bug,
but
they
don't
have
the
domain
knowledge
yet
to
figure
out
which
bugs
are
easier
for
them
to
address
versus
harder.
So
one
thing
that
worked
well
is
basically
tag
issues
as
good.
First
issue,
someone,
that's
just
starting
something,
that's
simple,
maybe
something
that's
just
as
as
a
small,
a
small
ui
change
or
or
something
that,
like
that's,
basically
gonna,
nothing
too
complicated.
A
I'm
trying
to
figure
out
an
example
and
for
an
extra,
for
example,
gene
talked
to
us
either
last
week
or
two
weeks
ago
about
changing
how
volume
mounting
works.
That's
a
complicated
change.
No,
no
newbie
will
be
able
to
pick
it
up,
but,
for
example,
exchange
that
said,
the
statistics
are
coming
from
container
are
aggregated
on
the
wrong
interval
like
per
minute
versus
hourly.
That's
simple
someone
that's
coming
up
could
figure
out
where
the
code
is
figure
out
the
algorithm
for
the
aggregation
and
fix
it
right.
A
So
so
I
think
I
think
all
of
us
from
the
community
should
start
tagging
issues
as
as
good,
first
issue.
When
you
see
them.
D
Yeah,
I
know
I
think,
that's
a
great
idea
so
and
that's
the
meeting,
so
I
will
post
the
doodle
and
wait
for
it.
So
once
that
meeting
starts
in
during
that
meeting,
you
know
these
kind
of
things
we
can
like
collectively
look
at.
I
think
everybody
should
do
it,
but
you
know
that's
purpose
of
the
meeting
to
kind
of
like
reinforce
these
things
and
you're
absolutely
right.
Actually,
my
first
pr
in
the
kubernetes
was
the
first
good
first
issue,
and
it
really
helped
me
like
when
I
was
looking
to
you
know,
do
a
pr.
D
Yeah-
and
that
may
mean
like
splitting
up
the
bigger
issues
into
smaller
issues
and
making
that
I
mean
that's,
also
fair
and
I'm
like
willing
to
work.
You
know
on
this
and
keeping
it
maintained
and,
like
you
know,
a
little
bit
up
to
date.
I
think
it's
it's
going
to
be
a
little
bit
of
help
to
the
community.
A
Anything
else
on
this,
no
at
all
I'll
post,
something
and
then
we'll
go
over
it
next
week.
Absolutely
this
is
awesome.
Thank
you
for
driving
some
of
that
all
right.
The
last
one
I'll
sit
deep
on.
We
have
a.
We
have
this
long-standing
pr.
It's
been
like
months
now,
this
device
manager
supporting
windows,
containers
that
lots
of
users
and
customers
have
asked
for
or
not
is
basically
kind
of
shepherd.
A
One
of
the
requests
that
you
have
made
from
him
in
the
community
was
hey.
Can
you
add
tests
for
this?
I
think
the
importance
of
adding
tests
is
super
super
critical
right,
that's
how
we
verify
that's,
how
make
sure
things
don't
break
catch
regressions.
A
On
the
other
hand,
someone
that's
new
and
trying
to
contribute.
I
think
it's
high
mountain
for
them
to
tell
them
to
go
ahead
and
add
both
a
test
as
well
as
a
pr.
I
want
to
get
everybody's
thoughts
on
this.
Like
you
know
this,
this
change
with
the
device
manager
seems
to
be
needed
by
many
folks,
I'm
wondering
if
we
should
just
let
it
in
now
before
tests
are
ready,
and
then
you
know
put
it
in
our
test
backlog
whether
are
not
created
or
someone
from
our
team
does
that's
kind
of
the
next
step.
A
Think-
and
this
I
mean
this
is
a
pr
here-
that
he
he
just
rebased
it
and
trying
to
get
it
through
the
through
the
motions
to
actually
get
it
approved.
A
In
no
thoughts,
okay,
I'll
work
with
deep
since
deep-
and
I
have
been
talking
to
are
not
before
I
mean
the-
has
anybody.
Let
me
ask
a
different
question:
this
is
more
for
the
folks
on
microsoft
and
amaz.
Have
you
guys
heard
anybody
asking
for
device
manager,
support
for
windows,
container
images.
D
I
think
I've
heard
about
it,
the
the
device
manager
I
wanted
to
about
sorry,
I
missed
it.
I
was
I
was
dealing
with
something
at
home.
I
I
missed
so
is
this
like
when,
when
we're
gonna
move
to
container
d
or
like
you
know,
when
continuity
is
going
to
be
supported,
like
are
you
going
to
pass
it
to
the
container
decry
interface
like
this
device
assignment
about
this
device.
A
A
Head
okay,
going
back
months
now,.
A
Yeah
I
got,
I
can't
find
it
right
now.
There
was
a
specific
thing
that
someone
needed.
Maybe
it
was
in
a
discussion
that
this
happened,
but
there
was
a
specific
capability
that
someone
needed
that
this
enabled
and
yeah.
I
can't
find
it
right
now.
I
think
oh
there's
140
loaded
items,
that's
right.
D
I
definitely
have
have
heard
about
this
before
from
multiple.
I
I
you
know
I
I
can't
remember
the
context,
but
I
can
get
back
to
you
on
the
on
this.
We
we
certainly
have
seen
you
know
some
requirements
around
this,
and
it
may
be
a
good
discussion
with
mark
like
around
how
the
container
b
is
gonna
support.
This.
D
A
A
All
right
cool
all
right,
that's
all
the
stuff
that
we
had
today
in
our
agenda.
Any
other
questions
concerns
from
anybody
and
it
will
help
if
we
had
the
right
date.
6
30
here.
F
So
this
is
jing.
I
added.
I
asked
a
question
on
the
slack
about
nfs
in
windows,
so
I
was
trying
to
see
whether
it
is
easy
to
support
nfs
in
windows
so
that
both
linux
and
windows
can
use
the
same
nfs
share
for
my
like
very
simple
testing
manually
testing.
When
I
mount
nfs
on
windows
is
only
readable
and
there
is
article
I
search
online.
F
There
seems
some
complicated
configuration
in
windows
needed
so
that
it
can
access
an
fs
share.
So
I
want
to
see
anyone
have
some
experience,
knowledge
about
it.
A
So
so
we,
I
don't
remember
if
it
was
david
that
had
done
the
initial
work
here
david,
I
may
be
misremember.
It
may
have
been
james
or
someone
else
from
microsoft,
but
when
we
were
looking
into
supporting
nfs,
it
looked
like
things
were
just
not
working,
so
we
made
that
decision
to
say
nfs,
based
storage
and
volume.
Support
is
not
supported.
A
F
G
Yeah
so
nf,
the
nfs
client
in
windows
supports
v3
and
in
v2
nfs
server
supports
v4,
but
not
the
client
side.
A
If
you
have
an
nfs
server
v3
would
that
work
nick.
G
F
F
Okay,
so
then
I
would
try
again
to
to
see
whether
I
can
use
v3
and
right.
You
know
yeah.
F
F
So
then,
I
want
to
ask
like
at
that
time
when
people
investigate
nfs,
why?
Like
they
stopped?
So
is
that
because
we
may.
F
A
G
F
A
F
It
is
actually
persistent
volume
coming,
so
you
just
create
a
pvc
pve
right
to
represent
the
nfs,
like
put
your
nfs
information
there
and
use
that
pvc
in
the
past.
Okay,.
A
F
So
right
now
we
are
like
moving
to
csi
and
the
andy
from
microsoft.
Just
had
a
csi
smb
driver
pushed
like
in
the
ripple,
so
we
are
testing
the
csi
smb
driver
and
plan
to
use
that
instead
of
flex
volume,
okay,
because.
A
F
F
Well,
so
for
csi
we
have
css
windows
meeting
every
monday
and
we
also
have
csi.
F
Dash
windows,
slack
and
yeah:
we
have
a
slack
channel
there,
it's
not
just
google,
it
also
microsoft
and
docker
right.
We
have
some
people
working
on
that
csi
project.
G
So
can
you
post
the
time
for
that
and
you
already
get
an
invite
for
that
meeting.
A
A
All
right,
everybody
have
a
great
rest
of
your
day
and
week
and
enjoy
the
long
weekend.
If
you're
in
the
states,
bye.