►
From YouTube: Kubernetes SIG Multicluster 2022 Mar 8
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
B
B
B
A
B
I'll
take
the
opportunity
to
do
paul's
normal
welcome
to
today,
tuesday
march
8th
episode
of
sig
multi-cluster.
I
have
taken
over
the
floor
because
I
have
also
taken
over
the
agenda
but
happy
to
sorry.
My
zoom
just
did
something
weird.
Well,
I'm!
Hopefully
you
all
can
still
hear
me.
B
Okay,
great,
I
can't
see
any
of
you
then,
but
that's
okay,
okay,
so
yeah
I've
taken
over
the
agenda,
but
it
should
be
kind
of
short
if
people
have
other
topics
later
okay,
so
I
wanted
to
give
some
kept
updates.
From
my
perspective,
I
know
jeremy
talked
last
time
about
how
we're
you
know
trying
to
move
everything
towards
the
next
steps
for
graduation
feels
like
the
constant
uphill
battle.
B
But
there
are
some
recent
updates,
so
I
just
want
to
make
sure
everybody
knows
so
for
cluster
id
me
and
tim
hawkin
are
still
negotiating
on
the
api
review,
but
I
think
we're
very
close.
I
did
want
to
highlight
that
the
main
discussion
on
this
point
has
been
about
max
length
and
everyone's
favorite
unicode
and
character
encoding.
B
If
you
would
like
to
read
a
long
and
involved
story
about
that,
there's
a
sort
of
comments
here
and
like
a
bunch
of
like
gists,
linked
to
because
we
wanted
to
test
some
things
empirically,
especially
if
you
have
you
know
any
like
prior
experience
with
open
api,
v3
specification
or
crd
development
for
something
for
for
fields
that
were
strings.
B
B
If
anybody
has
that
perspective
just
so
that
we
can
get
some
more
points
of
view,
and
I
want
to
shout
out
to
ishmeed-
I
don't
know
if
she's
in
the
call
right
now
but
she's
been
helping
with
the
actual
crd
implementation,
so
we've
been
working
with
on
the
cue
builder
implementation
of
the
actual
cluster
id,
which
is
also
what
enabled
me
to
easily
test
all
this
stuff
that
me
and
tim
are
talking
about
so
highlighting
that
for
multi-cluster
dns.
B
B
If
you
want
to
have
sort
of
the
shortest
version
of
what
the
latest
change
list,
I
guess
is
in
this
long
process.
This
comment
represents
the
most
material
things
besides
other
like
formatting
knits
that
were
handled
so
yeah,
that's
all
informational
and
then
in
terms
of
actual
next
steps.
Again,
both
of
these
are
between
me
and
tim
right
now.
B
B
Okay,
so
two
kind
of
just
random,
maybe
not
super
random
curveballs,
but
just
two
thoughts.
I've
been
having,
especially
while
talking
with
other
people
about
various
things.
So
one
is
there's
this
prr
questionnaire,
which
here
this
part,
which
some
people
here
may
have
interacted
with,
which
is
a
questionnaire
that
any
cap
is
required
to
fill
out
and
get
a
pr
reviewer,
which
is
like
a
subset
of
people
similar
to
api
review.
B
Who
can
give
the
lgtm
on
this,
and
I'm
just
wondering
whoops
sorry
over
here,
I'm
just
wondering
if
now
or
in
the
future,
there's
any
interest
or
thought
about
putting
in
any
questions
about
how
future
enhancements
to
kubernetes
as
a
whole?
What
it's
thought
about?
What
that
means
for
multi-cluster
deployments
and
yeah?
I
kind
of
just
want
to
open
the
floor
of
if
this
has
been.
Maybe
ask
the
lore
people
if
this
has
been
discussed
before.
B
A
B
A
Certainly,
if
it
was,
it
was
before
my
time
in
the
sig
right,
but
I
think
we
should
think
about
what
what
that
would
actually
mean
yeah,
I
know
it's,
it
was
purposefully
an
open-ended
question,
but
what
kind
of
question
might
we
might
we
ask
there
like
just
is
it
like?
Is
this
a
multi-cluster
feature?
Could
this
be
a
multi-cluster
feature,
or
how
would
this
be
a
multi-cluster
feature
if
we
want
to
start
spurring
thought
in
that
direction?
B
Yeah
and
one
thing
that
I
feel
both
was:
how
does
how
do
I
say
so,
for
example,
kind
of
the
flagship,
current
flagship
project
of
sig
multi-cluster
is
mcs
and
it's
optional
and
I
think
a
lot
of
other
multi-cluster
projects
too
historically
have
been
optional
or
add-ons,
but
they
are
still.
You
know,
like
native
upstream,
so
there's
just
an
interesting
interaction
compared
to
the
other
types
of
things
that
are
asked
on
this
specific
questionnaire.
B
B
So
I'm
just
wondering
if
that
has
an
impact
on
any
decision
we
as
sig
multi-cluster,
might
make
about
what?
If,
if
we
put
a
question,
that's
like
have
you
thought
about
multi-cluster?
B
A
Yeah,
I
I
actually,
I
really
like
this
in
terms
of
helping
projects
think
about
like
mcs
as
a
good
example.
If
we
look
at
other
changes,
you
know
I
think
networking
is
most
likely
to
cause
immediate
issues,
but
right
you
know,
we
there's
the
network
policy
working
group
right
now
and
they're
coming
up
with
new
plans.
Is
this
gonna?
A
Is
this
going
to
step
on
multi-cluster
like?
Yes,
I
think
multi-cluster
will,
for
the
foreseeable
future,
be
made
up
of
add-on
features
in
the
community,
but
they
still
shouldn't
break
because
some
new
entry
feature
rolls
out
and
I
think
that's
a
that
might
be
worth
a
checkpoint.
B
It
might
be
a
good
like
fundamentally,
this
prr
questionnaire
is
about
like
the
process,
and
so
this
may
might
just
be
a
good
place
to
throw
that
in
there
and
I
don't
or
how.
What
am
I
trying
to
say
that.
B
All
of
us
members
of
sigma
multi-cluster,
we
don't
have
the
temperature
on
everything
going
on
in
every
other
sig
right
and
we
might
get
lucky
if
we,
like,
you
know,
know
enough
in
a
in
a
given
space
or
if
things
move
slow
enough,
that
we
can
catch
up,
but
this
might
help
prevent
some
things
falling
through
the
cracks
solely
just
in
the
form
of
while
someone's
filling
this
out,
because
they
have
to
anyways
they're.
Like
oh,
say,
multicluster
cares
about
stuff
or
we
could
be
like.
B
A
Yeah
I
like
it.
Maybe
we
can
get
a
brainstorming
dot
going
with
like
a
sample
question
or
yeah
or
a
few
and
share
it
with
the
list.
B
Make
a
brainstorming,
doc
yeah,
I
happen
to
have
recently
filled
this
out,
so
I
have
an
idea
of
what
types
of
things
get
asked
on
it,
and
also
that's
why
I
thought
it
was
a
good
spot
for
us
to
be,
but
I'll
also
just
encourage
anybody
to
take
a
look
at
these
questions.
If,
if
you
have,
if
you
want
sort
of
the
backstory
of
what
they
are
like
and
then
I
will
do
that
I'll
send
out
a
brainstorming
doc
where
we
can
combine.
B
You
know
our
goal,
which
I
think
is
make
sure
future
enhancements
to
upstream
kubernetes
single
cluster,
not
break
sig
multi-cluster
supported
multi-cluster
add-ons
and
also
make
it
easier
for
people
to
for
us
to
keep
tabs
on.
What's
going
on,
given
our
application
to
maintain
multi-cluster
compatibility
in
upstream.
B
Okay,
bam
bam,
bam
right
on
to
number
three,
so
this
is
also
kind
of
related,
which
is
just
that.
I
observe
that
you
know
there's
a
lot
of
sig
network.
Sig
mc
overlap
right
now,
particularly
about
things
that,
like
compatibility
with
sig
network
projects
and
again
like
I
had
to
say
like
I
know
that
mcs
itself
required
a
lot
of
interaction,
and
I
came
in
a
little
bit
later
where
that
link
was
very
well
established.
B
So
maybe
I
maybe
this
is
already
the
way
that
that
process
goes
well,
but
I
have
just
talked
a
little
bit
to
people
about.
Is
there
a
need,
as
since
we
have
a
larger
number
of
related
proposals
going
on
right
now,
to
have
like
a
more
formalized
check-in
between
sig
network
and
sigmc
right
now?
I
think
that
there's
basically
representatives
from
both
who
go
to
both
meetings.
So
on
the
one
hand,
I
don't
want
to
make
like
another
meeting
for
just
those
people,
I
guess
and
take
up
time
of
their
day
but
yeah.
B
A
So
working
groups
tend
to
be
kind
of.
I
believe
it's
actually
like
in
the
charter
that
they
are
right
time
bounded
and
for
a
specific
purpose.
So
in
general
I
definitely
like
the
idea,
but
I
wonder
if
it
should
be,
if
we're
gonna.
If
we
wanted
a
like
working
group
between
signet
and
sigmc,
would
it
be
better
to
pile
into
like
network
policy,
for
example
right.
C
A
People
who
attend
both
already-
I
don't
know
how
much
that
helps,
but
you
know
obviously
love
to
hear
anybody
who
who
has
other
opinions,
but
if.
A
If,
if
we
have
like
a
specific
goal
here,
like
like
figuring
out
network
policy
or
figuring
out
multi-network,
I
think
those
are
those
are
pretty
good
candidates,
yeah.
D
B
Yeah
yeah
and
I
feel
like
the
first
sort
of
like
you
know.
I
don't
know
how
to
say,
like
flags
of
that
have
kind
of
occurred
where
like
when
we
were
talking
about
multi-cluster
network
policy
or
multi-network
like
we
would
discuss
the
documents
here
and
we'd,
be
like
oh
we're
we're
having
this
conversation
over
in
sig
network,
or
this
is
pending
this
conversation
in
sig
network
or
I
went
to
network
this
week
and
blah
blah.
You
know
that
type
of
thing,
so
that's
where
it
like
started
to
feel
like
it
was
trending.
B
B
A
And
and-
and
I
think
policy
well,
both
multi-network
and
network
policy
are
kind
of
exactly
the
types
of
problems
that
working
groups
are
generally
formed
to
solve.
Like
we
have
a
specific
goal
brings
together,
the
the
sigs
seems
like
a
pretty
good
candidate.
B
B
Yeah
yeah-
and
I
guess
like
we
don't
need
to
like
make
a
grid
of
everybody's
attendance,
but
I
don't
know
like
ever.
I
personally
come
to
this
meeting
regularly
and
not
the
other
one.
So
I
feel
a
lot
less
aware
of
what's
going
on
or
I
feel
like
it's
disjointed,
but
if
a
like
strong
enough
group,
you
know,
is
already
in
all
those
places,
then
you
know,
maybe
I
just
need
to
show
up
to
all
of
those
if
that's
how
how
it
is
going
forward.
B
B
B
Cool
any
other
people
on
the
call
want
to.
C
B
Okay,
okay,
okay,
cool
yeah.
I
am
not
a
good
mc
right
now,
because
my
zoom
is
freaking
out,
so
that's
cool,
but
I
will
yield
the
floor
and
to
the
question
and
chat
from
I'm
in
who
is
new
to
the
sig
and
has
a
question
for
the
end
feel
free
to
come
off
me.
If
you
want
to
ask
yourself,
but
it's
is
there
any
development
roadmap
for
cube,
fed.
A
Yeah,
I
am
not
actually
currently
aware
of
one.
There
were
some
projects
I
think
about
six
months
ago
in
flight,
but
it
does
seem
like
work
has,
has
stalled
a
bit.
So
I'm
not
currently
aware
of
of
an
updated
roadmap
from
keep
fed
paul.
Have
you
heard
anything
different.
D
No,
I
haven't
heard
anything
different
and
I'm
I
have
started
asking
myself
the
question:
is
it
time
to
archive
that
project?
D
It's
been
in
a
state
that,
like
full
disclosure,
I
had
my
own
hand
in
like
at
one
point
a
state
of
like
it
will
be
beta
real
soon
now
for
more
than
longer
than
the
pandemic,
at
least
so
I
think
it's
it's
better
if
it's
not
being
actively
developed
and
I'm
not
sure
that
that's
the
case
but
like
if
there's
really
no
road
map,
I
think
it.
It
doesn't
make
a
lot
of
sense
to
keep
to
keep
the
project
unarchived.
A
Yeah
yeah,
I'm
yeah.
We
should
do
that.
I
think
we
we
need
to
figure
this
out
and
then
archive
it
and
then,
if
somebody
were
to
want
to
run
with
it,
you
know
reach
out.
We
would
we
would
let
the
the
mailing
list
know
as
well
but
yeah
if
nothing's
happening
with
it.
That's
probably
a
better
home.
B
B
One
more
chat
comment,
michael
plus
one.
The
specific
project
working
groups
gave
an
example.
So
that's
good
to
know.
I
will
incorporate
that
into
the
calculus.
E
Hey
this
is
lithium
from
azure.
I
have
a
question
about
following
up
the
topic
on
gateway.
E
So
now
we
have
gateway
plus
mcs,
so,
like
I
know,
gateway
is
for
north
south
traffic,
then
mcs
is
for
east
west
traffic
and
based
on
what
I
deserve,
that
gke
actually
put
a
put
mcs
as
backhand
for
for
gateway.
So
I'm
wondering
what's
the
community's
recommendation
in
terms
of
how
we
put
them
together
and
like
how,
like
the
north,
south
and
stress
networking
should
work
in
the
multi-cluster
world.
A
So
I
think
the
gateway
api
working
group
speaking
working
groups
basically
came
to
that
conclusion.
We
were,
we
were
pretty
involved
with
them.
Well
well,
specking
that
out,
but
the
thinking
is
that
you
know
gateway
gateway.
Is
the
right
api
for
north
south
and
if
you
want
multi-cluster
back-ends
pointing
at
a
service
import
instead
of
a
service
is
a
is
a
pretty
reasonable,
straightforward
way
to
do
that,
because
multi-cluster
services
already
are
the
multi-cluster
equivalent
of
a
service.
A
So
if
the,
if
the
yeah,
basically,
if
your
back
into
a
multi-buster
service,
it
should
use
the
multi-cluster
service
equivalent,
and
so
I
think
the
community
was
kind
of
going
that
way
and
that's
why
gke
did
it
folks
from
gk
were
pretty
involved
in
in
that
community
as
well,
and
I
think
there
was
a
lot
of
uptake.
A
E
I
see
got
it
so.
Basically,
the
recommendation
for
connecting
multi-cluster
service
at
the
back
end
for
gateway
is
to
use
service
imports
at
the
back
end.
That's.
A
A
But
I
don't
know
if
it
I
don't
know
its
current
state
in
the
api.
There
were
some
good
discussions
in
the
gateway
working
group
about
it,
but.
A
This
gets
into
the
kind
of
interesting
case
of
the
sig
projects,
because,
like
mcs
gateway,
api
is
a
is
an
optional
add-on
right
right,
and
so
that's,
why
kind
of
it
has
its
own
website
we're.
A
Right,
maybe
we
are
developing
a
we're
starting
to
see
that
maybe
a
better
solution
as
crds
are
the
path
forward
for.
B
E
So
I
also
have
another
question
about
qbfas,
so
we
talked
about
that.
We
don't
have
any
development
roadmap
for
it
and
basically
we
are
some
setting
it
then,
as
a
community.
What
do
we
provide
as
an
alternative,
or
we
just
gave
up
right.
D
D
So
what
I
mean
by
that
is
like-
and
I
I
haven't-
I
may
have
spoken
like
too
soon-
that's
possible
right
in
the
sense
that
I
haven't
gone
and
recently
looked
at
like
the
state
of
the
cube
fed
repo.
D
D
But
to
answer
your
question
I
about
what?
What
would
we
provide?
Instead
of
that?
The
answer
might
be
nothing
right.
The
one
of
one
of
the
things
that
in
the
history
of
the
sig
was
a
process
that
the
sig
went
through
is
refining
the
concept
of
what
the
sig
did
from
being
very
closely
tied
to
the
concept
of
like
kubernetes
federation.
That
was
was
sort
of
conceived
originally
as
something
that
was
had
a
lot
of
different
jobs.
D
D
I
am
not
personally
of
the
opinion
that
it
would
necessarily
be
good
to
try
to
build
another
thing
that
did
all
of
the
things
that
cute
fed
tried
to
do.
I'm
not
sure,
but
I
I
think
that's
the
best
answer
I
could
provide
right
now.
I
don't
know
if
that
makes
sense
to
you,
but
that's
what
came
out.
E
E
What's
the
real
status
of
the
project,
without
I
mean
watching
all
the
community
videos,
and
what
I'm
thinking
about
is
that
it
seems
that
it's
very
hard
to
agree
on
a
solution,
but
is
it
possible
that
we
start
to
think
about
whether
we
can
agree
on
additional
apis
that
can
help
use?
I
mean
different
providers
or
different
social
club
committee
to
provide
solutions
because,
like
how
do
we
represent
a
a
member
of
a
multi-cluster?
E
A
This
to
me
is
exactly
the
right
question
like
what
what
what
specific
problems
do
we
need
to
solve?
I
think
you
just
threw
out
like
a
couple
there
that
are
much
more
targeted
than
cube
fed
or
another
one
that
paul,
and
I
and
a
few
others
have
talked
about
a
few
times
now
is
like
helping
with
multi-cluster
leader
election.
A
If
you
want
to
develop
a
multi-cluster
controller,
that
seems
like
a
really
interesting
problem
super
useful
and
if
you
want
to
use
multi-cluster
for
aha,
it
kind
of
seems
like
a
problem
everybody's
going
to
have.
So
if
we
can,
I
think
the
right
path
forward
in
the
way
that
we've
been
taking
the
sig
is
to
figure
out
that
list
like
what
are.
What
are
the
specific
problems
that
that
we're
seeing
pop
up
on
a
bunch
of
on
a
bunch
of
different
use
cases,
and
can
we
just
solve
those.
D
Yeah
in
the
neighborhood
I'll,
just
plus
one
to
everything,
jeremy
jeremy
said
I
will
throw
out
that,
like
in
the
functional
neighborhood
of
one
of
the
things
that
is
like
in
the
scope
of
what
q
fed
attempted
to
do
is
spreading
resources
around
and
one
one
one
of
the
other
things
that
the
is
like
in
the
neighborhood
of
things
that
the
sig
has
invested
some
time
in
is
the
work
api.
D
I
would
say
that's
probably
the
closest
point
of
continuity
like
as
far
as
other
work
within
the
sig
to
to
what
cube
fed
mostly
does,
but
that
one
also,
I
think,
there's
a
question
of
is:
is
there
an?
Are
there
enough
people
interested
in
the
work
api
to
continue
developing
that?
So
that's
when
you
may
want
to
take
a
look
at
if
you're
interested
in
apis
that
can
be
vendor
neutral.
D
That's
certainly
fits
checks.
Checks
that
box
checks,
the
functional
neighborhood
of
cube,
fed
box
might
be
might
be
worth
taking
a
look
at
I
did.
I
did
reach
out
to
trojan
some
time
back
and
and
asked
if
that
was
still
under
under
development
or
if
there
were
any
updates,
and
he
he
indicated
to
me
that
it
was.
But
I
would
I'd
really
like
to
get
an
update
in
a
sig
meeting
about
work
api
so
that
we
can
at
least
understand
what
the
current
state
of
that
stuff
is.
E
C
Okay,
I
do
have
a
update
on
the
work
api.
We
we
just
recently
made
a
enhancement
on
it.
If
anyone
have
some
feedback,
please
please
go
ahead
and
thank
you
honkai
for
the
feedback
so
far.
Thank
you.
B
I
may
be
overdoing
it
on
the
notes,
but
I'm
wildly
trying
to
capture
all
this.
So
if
I'm
missing
anything
laura.
A
A
All
right:
well,
then,
I
think
we
will
call
it
a
day.
Thank
you,
everyone
and
thank
you
laura.
That
was
some
great
discussion
I'll.
I
will
see
you
all
in
a
couple
weeks.
B
Yeah
and
feel
free
to
update
the
notes,
if
I
missed
anything.