►
From YouTube: Kubernetes SIG Multicluster 20180102
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
So
this
this
actually
this
discussion,
what
you're
talking
about
that
sort
of
happened
over
here
and
if
you
see,
for
example,
Klein,
go
kubernetes,
client
or
some
other
reports
which
are
maintained
separately.
They
publish
a
metrics
metrics
of
support
occur
like
save,
and
auto
version
corresponds
to
kubernetes.
1.4
version
of
client
goes
something
like
that
which
is
published
on
the
repo
page
and
the
release
page
I.
Don't.
C
A
B
And
do
that
see
my
I
have
no
problem
waiting
on
this
and
then
everybody
comes
back
and
then
do
it,
whichever
way
take
shape.
But
you
see
that
some
people
have
already
raised
issues
and
all
about
kuba
fed
not
being
found
in
the
release
and
all
that
so
I
think
some
stopgap.
At
least
we
need
to
do
right,
kinda.
C
I
just
think
that
there
needs
to
be
more
messaging
as
to
you
know
the
support
ability
of
what
they're
trying
to
use
anyway
like
previously,
there
is
kind
of
oh
yeah.
This
is
part
of
kubernetes
and
you
can
use
it
and
it's
like
it's
I
mean
how
is
someone
to
differentiate
the
quality
of
Cupid
versus
a
cute
kettle?
B
C
You
have
people
continuing
to
use
Federation
as
if
it
was
something
that
they
should
put
into
production
in
current
form
and
I.
Think
that's
a
problem
so
to
me
the
fact
that
it's
not
part
of
kubernetes
and
people
are
like
well.
What
do
I
do
is
an
opportunity
to
message
the
real
state
of
the
world
and
say
where
actively
working
on
developing
a
solution.
That's
supportable!
Until
that
time
you
can
get.
C
You
know
the
pre-alpha
saying
that
we're
working
on
in
the
tree
today,
but
make
it
clear
that
you
know
we're
gonna,
see
something
that
looks
a
little
bit
different
from
an
API
perspective
and
maybe
a
usability
perspective,
and
you
might
want
to
wait
until
that's
available
before
you
think
about
production.
Izing,
your
solution
with
us.
You.
D
B
Yeah
yeah,
okay,
okay,
yeah,
that's
a
good
suggestion
and
I
think
majority
of
the
people.
The
users
actually
go
through
documentation,
so
they
go
to
the
documentation.
And
you
just
say:
okay,
the
release
is
available
here
and
then
let's
try
to
use
it,
there's
a
big
bunch
of
people
who
actually
go
and
land
that
wherever
the
binaries
are
that
fit.
So
that
also
needs
a
bit,
but
that
is
sort
of
I
am
sort
of
waiting
for
the
release
to
have
in
some
ways
that,
after
which
I
can
update
the
documents,
I
mean.
C
Previously
I
think
it
was
possible
for
people
to
just
go
to
the
you
know:
kubernetes
release
and
just
you
know,
pull
the
Federation
artifacts
out
of
that
or
use
those,
because
that's
no
longer
possible.
We
have
an
opportunity
to
provide
some
documentation.
As
for
the
state
of
the
world,
so
some
people
might
read
the
docs,
some
people
might
go:
hey
Federation,
I'm
gonna
go
consume
the
binaries,
oh
it's
not
there.
Where
do
I
get
it.
B
B
This
is
that
in
debase
occur
from
the
last
meeting,
which
was
there
so
I
thought
some
Kubik
on
update
selected
two
related
to
the
people
joining
this,
but
I
don't
think
there
is
enough
attendance
today
for
that
kind
of
a
pig
and
I.
Remember
that
one
of
the
steel
guys
mentioned
that
he
would
want
to
speak
in
the
Federation
sink.
I
did
individually.
Send
him
a
invite
I'm,
not
sure.
If
he
is
aware
that
meeting
is
today
or
not.
E
B
B
Okay,
I
have
one
item
to
discuss
like
if
nothing
else,
it
is
related
to
Federation
yep
like
currently
we're
using
Cooper
Fett
to
deploy
the
condition
kind
of
plain
components
right.
So
as
an
alternative,
can
we
think
of
a
line
chart
like?
Are
we
reinventing
the
same
stuff
like?
It
is
growing
too
huge
to
adapt
to
all
sorts
of
customizations?
C
A
A
As
of
yet
I
mean
I
haven't
looked
personally
myself
and
I
haven't
looked
recently,
but
I
don't
think,
given
that
that
it's
going
to
be
possible
to
make
helm
charts
work,
as
is
I,
think
if
you
wanted
to
use
helm
or
charts,
you
would
have
to
be
an
alternative
strategy
or
a
less
integrated
setup
solution.
Where
there's
more
manual
steps
for
the
user
to
do
I.
B
Over
there
yeah,
and
also
like
why
I'm
asking
this
is
like
currently
D
deploy
etcd
as
a
board
within
the
same
deployment
as
the
fh
over
right
now
or
be
an
ideal,
be
like
how
the
user
wants
to
use
it.
So
probably
they
have
already
deployed
an
eighty
city
server
and
then,
above
that
they
probably
want
to
use
a
perdition.
So
right
now
the
condition
control
plane
sits
on
just
the
one
post
caris
cluster.
Maybe
it
is
not
that
ideal.
Also,
like
many
people
are
asking
like
how
do
we
spread
this
control
plane
across
clusters?
B
B
B
I,
don't
think
so
if
I
understand
correctly.
Currently
it
is
all
intrinsic.
The
ET
city
is
deployed
like
a
single
pod
within
the.
If
it's
Edward
deployment,
no
I
think
she
is
talking
about
Dever
already.
There
was
already
a
discussion
about
about
not
trying
to
add
more
flags
and
if
any
point
of
time
we
come
with
land
into
a
situation
that
we
need
more
flex
to
better
move
to
using
some
mechanism
of
config
like
kuba
ADM
is
using
I
mean
that's
for
an
example
sure.
C
Right
right,
I
just
wasn't
sure
that
there
was
like
for
some
reason,
I
was
under
the
impression
that
it
was
possible
to
configure
an
existing
at
CD
cluster
I
I.
Would
ask,
though,
seems
like
the
common
challenge
here.
Is
that
helm
doesn't
support
the
features
that
would
be
required
to
do
like
SSL?
That
seems
like
that
would
be
a
common
enough
scenario,
but
you
know
we
should
file
an
issue
with
helm
and
see
if
they
can
fix
it
rather
than
having
to
continue
maintaining
our
own
tool.
B
C
This,
what
I'm
saying,
if
it
does
it
today,
I'm
just
saying
if
that's
a
deficiency
in
helm
I,
can
see
a
need
for
this
capability
for
a
lot
of
different
things,
at
least
within
the
kubernetes,
like
eco
system,
that
you
know
the
upstream
development
community
works
on,
let
alone
downstream,
so
I
mean
if
we
solve
that
in
general,
then
everybody
wins
versus.
Oh
we're,
just
gonna
continue.
You
know
extending
something
like
you
bad,
because
home
has
this
deficiency,
that's
imposing
a
major
cost
on
the
development.
A
Yeah
I'm
on
board,
with
that
I
have
a
vague
recollection
of
helm,
either
saying
we
don't
have
time
to
do
this,
or
this
isn't
something
that
we
particularly
want
to
do.
Unfortunately,
I
can't
remember
which
or
if
that's
true,
so
it's
definitely
worth
filing
an
issue
and
letting
them
respond
again.
If
that's
not
half
they
want
to
take
home
down.
I
could
understand
if
they
don't
want
to
add
too
much
logic
into
helm
and
they
don't
want
to
say
open
the
door
to
like
well.
C
Many
other
option
is
that
we
work
with
in
helm
to
do
I.
Think
about,
like
oh
I,
need
some
sort
of
dependency
mechanism
for
getting
the
service
and
an
ssl
certificate
like
matching.
We
there
could
be
a
pod.
That
actually
does
that
that
you
know
helm,
deploys
and
figures
out
how
to
do
it
like.
It
seems
like
something
we
should
be
able
to
solve
with
a
little
bit
of
work.
We
just
have
to
justify
that
effort.
It.
A
Might
also
be
the
case
that
I
mean
a
lot
of
what
Cupid
does
could
be
migrated
into
home
charts
with
the
bits
that
have
to
be
done
with
a
sort
of
with
logic
living
in
cube
bed.
And
then,
though,
some
charts
exist
independently
for
people
to
use
in
a
setup
that
doesn't
use.
You
bet
necessarily
right,
which
would
be
I,
think
a
win,
because
then
people
would
be
much
it'd,
be
much
easier,
I
think
to
deploy
without
a
net
CD.
A
C
I
mean
the
fact
that
that
helm
is
like
nominally
just
a
bunch
of
configuration,
means
that
it's
much
easier
for
customization
like
downstream
versus
cube
pad,
where
you
might
actually
have
to
change
go
code.
I
would
be
much.
I
would
be
very
much
in
favor
of
making
it
easier
to
allow
downstream
customization
I.
B
A
You
and
there's
really
been
nothing
much
there's
been
a
little
bit,
I
guess
done
since
then,
but
not
really
much.
The
goal
now
sort
of
to
work
through
the
rest
of
the
items
to
the
rest
of
things
to
just
work
through
the
items
in
the
milestones.
I
have
a
goal,
probably
over
the
next
few
weeks,
to
try
and
have
a
more
comprehensive
testing
plan
for
the
cluster
registry
and
then
to
start
working
on
actually
writing
test
harnesses
and
test
frameworks,
I'm
in
figuring
out
what
kinds
of
tests
we
want
to
have
I.
A
Think
there's
some
there's
some
other
just
issues
that
need
to
be
worked
on.
One
thing
that
will
be
hopefully
coming
up
soon
is
that,
as
the
cluster
registry
I'd
like
to
start
thinking
more
carefully
about
how
to
integrate
the
cluster
registry,
replace
the
Federation
cluster
API
with
the
cluster
registry,
there's
an
aggregated
server
or
as
some
sort
of
direct
inclusion,
I
think
now
it's
in
a
state
where
he's
ready
to
start
talking
about
that
and
figuring
out.
If
there
is
anything
left
blocking
it.
A
The
one
thing
I
know
about
is
that
the
cluster
object
doesn't
yet
have
a
status
I.
Don't
think
that
would
be
that
difficult
to
add
the
just
the
status,
that's
similar
to
what
Federation
uses
and
I
think
that's
a
perfectly
reasonable
thing
that
anybody
who's
doing
is
clusters
and
doing
them
and
act.
Any
active
reconciliation
would
have
some
sort
of
status.
I
mean.