►
From YouTube: Kubernetes SIG Multicluster 20180130
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
A
B
To
expound
on
Neal's
point
and
I:
don't
want
to
take
words
out
of
your
mouth,
but
I
think
the
challenge
isn't
so
much
speaking
about
things
in
the
context
to
this
meeting,
it's
more
like
how
do
we
communicate
to
a
broader
audience
that
doesn't
participate
in
these
meetings?
What
we're
working
on
and
what
the
status
versus
yeah.
C
C
Sometimes
they
don't
get
enough
replies
like
the
I'm
one
view
I
had
was
maybe
be
a
building
a
concern
perception
that
is
there's
so
many
people
in
the
sink
but
like
what
are
they
doing
like
so
many
questions
about
Federation
or
like
what's
happening
in
Federation,
but
like
not
as
many
people
I
guess
so
many
people
here
and
Sigma
Tings
and
on
email
discussions,
but
not
everyone
like
working
on
a
federation
or
other
issues.
So
we
should
clarify
and
we
we
send
out
an
email
to
cover
latest
users
when
we
rename
the
sig.
C
D
C
A
E
Yeah
hello,
so
my
name
is
Matt
Iacobucci,
sorry
I
just
had
to
run
in
from
we've
got
a
community
space
and
so
I
just
grabbed
a
room.
So
yeah
I'm
working
for
f5
and
I
got
interested
in
sig
Multi
cluster
tending
Q
Khan
last
Q
Khan
in
Austin
and
so
I'm.
Just
right
now,
I'm
just
trying
to
lurk,
learn
as
much
as
I
can
about
Federation.
You
may
seen
how
we
could
possibly
contribute
contribute
in
the
future.
F
G
Actually
build
a
kubernetes
appliance
and
I'm
heavily
involved
in
the
storage
suggestions.
So
right
now
we
are
trying
to
add
support
for
kind
of
a
hybrid
Road
in
the
multi
cluster
flyer,
from
listen,
opera
and
so
I'm.
Getting
involve
lots
of
the
actively
looking
if
I
have
any
questions
helping
and
going
forward,
I'm
willing
to
take
up
some
of
the
work
to
make
things
work
in
a
multi
custom,
ROM
I'd.
Thank
you.
H
H
I
D
Program
manager,
my
name,
is
Michael
Rubin
I'm,
a
manager
at
Google
for
a
whole
bunch
of
stuff
in
kubernetes,
reindeer
from
like
storage,
Multi
cluster
operating
systems
and
sort
of
random
stuff.
I,
like
the
word
stuff
and
I've,
been
involved
in
multi,
cluster
and
Federation
for
about
two
years
now
so
I've
seen
this
group,
you
know
evolve
and
change
and
grow
into
where
it
is
today
very
happy
to
be
here.
A
B
F
A
D
D
My
main
focus
is
to
make
sure
that
we
address
multi
cluster
problems
and
not
necessarily
focus
on
federating
phonetics
API,
as
is
as
a
result
of
that
we
have
various
efforts
like
Federation
working
group
itself,
which
is
focused
on
developing
api's,
that
we
think
map
to
use
cases
that
people
are
trying
to
solve
in
the
multi
cluster
world
and
at
Google
we
have
our
own
projects
that
are
all
reflected
as
well
in
the
open
source
and
those
are
cluster
registry
and
cube
mci,
which
is
a
coaster
ingress
to
the
main
developer.
There
hi.
J
Hi
I'm
Craig
Carmen
I'm
working
on
cube
mci
with
the
hill
I've,
been
on
the
scoops
in
September
and
ramped
starting
work
on
controller
stuff.
D
A
document
that
I'm
gonna
send
over
to
the
sink
is
right.
Now
our
landing
page
doesn't
really
reflect
any
of
what
we
just
discussed
as
we
went
around
the
table
in
that
there
are
separate
ongoing
efforts
unless
you
attend
every
single
meeting
and
even
there
I'm
not
sure
you
don't
really
know
where
to
go,
to
find
all
the
efforts
that
we've
covered
in
the
last
three
minutes.
So
I
think
we
need
to
address
that
on
the
community
page
and
I.
That's
my
action
item
for
for
this
week,
I'll.
D
L
M
This
is
a
crystal
Luciana
another
note
in
there
for
possible
things,
so
I
work
on
mostly
Canadians
networking
and
resource
management
working
group
stuff,
but
I've
also
can
we
get
involved
in
this
from
the
sto
perspective,
because
we're
interested
and
how
the
cluster
API
is
working
out.
Hello
in
a
gray
with
mesh
inside
of
us
do.
N
D
D
Why
would
that
be
something
that
people
would
find
useful?
You
know
I
know,
I
see
some
new
faces
and
I
know
that
there
are
maybe
more
wider
faces.
We
could
advertise
this
in
the
community
study.
The
community
meeting
not
have
a
whole
cig
update,
it
just
say
eh'.
If
you're
interested
in,
what's
going
on
we're
going
to
have
you
know
an
hour
session,
20
minutes
of
prep
and
then
40
minutes
of
today.
D
So
then,
let's
say
that
we'll
well
we'll
take
an
action
item
in
this
room,
at
least
to
get
that
started
on
the
mailing
list
and
kick
it
off,
and
you
know
it's
really
really
low
cost
to
do
on
the
members
of
the
state
and
I
think
it
might
be
a
way
to
you
know,
bring
everyone
up
to
speed,
I
think
the
biggest
thing
is
just
getting
those
like
three
three
minutes
or
three
seconds
inside
the
community
meeting.
That
should
be
that
difficult,
probably.
A
D
I
think
the
community
meeting
and
having
a
say,
update,
I
think
we
should
do
a
sig
update
for
multi
cluster.
Definitely
but
I
think
this
sort
of
meaning
is
something
that
could
go
on
for
a
while
and
be
very
interactive.
And
what
I've
noticed
in
the
community
meeting
is
that
it's
not
always
the
best
forum
to
have
a
lot
of
people,
pepper,
pepper,
other
people
with
questions.
So
you
know
I
what
I
like
about
this
meeting.
That
happened
with
the
storage
taken
at
you
and
the
scheduling
stick
was
that
it
was
really
interactive.
D
I
think
that
when
it's
done,
we
should
probably
definitely
do
a
multi
cluster
update
I
think
we
are
getting
a
little
past
due
for
that,
but
I
feel
like
that's
more
of
them.
Information
coming
out
as
opposed
to
a
dialog
I,
don't
know
if
a
committee
meeting
lends
itself
well
to
the
dialogue
and
I've
said
that
three
times
in
three
different
ways:
it's
all
this
okay.
A
A
D
The
community
meeting
first,
let
a
wider
audience
know
that
we're
going
to
do
this
and
say
if
you're
interested
subscribe
to
our
mailing
list.
Otherwise
we'd
subscribe
to
the
list,
then
we've
advertised
and
people
wouldn't
know
you
know
how
to
contribute
to
the
time
and
place.
I
expect
that
might
even
be
a
doodle
or
scheduling.
D
D
A
K
D
One
go
to
that
meeting
and
say:
hey
we're
gonna,
throw
a
party
as
opposed
to
throwing
a
party
to
the
people
that
are
in
the
know
already
and
after
the
safe
you
want
to
learn
about
the
party
come
to
our
list.
Then
people
would
come
to
the
list.
We
could
schedule
the
party
together
I'm
starting
I'd
like
to
say
party,
and
then
we
would
have.
Hopefully
you
know
we
brought
in
the
group
of
people
that
we
would
be
able
to
engage
with
this.
That
way,
as
opposed
to
we
started
it
on
the
list.
D
D
L
D
D
Me
it
sounds
like
office
hours
yeah,
it
might
be
a
little
office
hours,
but
I
think
it
should
be
office
hours
with
an
initial
like
20
minute.
Here's
where
we
are
status
update
for
not
a
super
detail.
This
is
where
the
code
lives,
and
this
is
the
you
know
the
pieces
of
the
API
but
more
10,000
status
update,
because
otherwise
you
and
I
mean
I,
don't
know
about
you,
but
sometimes
I
would
go
to
office
hours
in
school
and
not
even
know
what
I
didn't
know.
D
O
Shirts
sure
so,
can
you
guys
hear
me
yeah,
okay,
yeah,
so
Josh
purpose
as
a
last
week
in
kubernetes
development
mailing
list
that
it
most
people
here
are
familiar
with
like
kubernetes
weekly,
but
that's
like
from
a
broader
ecosystems.
You
know
reference
who's,
writing
what
blog
post,
what
presentations
happened?
O
What
meetups
are
going
on,
but
this
is
basically
a
summary
for
folks,
like
developers
who
are
contributing
the
kubernetes
who
want
to
know
what's
going
on
in
other
states,
so
that
might
be
a
good
place
for,
like
a
you
know,
kind
of
a
last
weekend,
Multi
cluster,
what
PRS
got
merged?
What
the
release
schedule
looks
like
is
there
anything
being
blocked
at.
N
O
High
level,
so
you
can
imagine
like
one
sentence:
one
leak,
maybe
three
or
four
links
a
week.
If,
if
it's
pertinent-
and
you
know
that
way-
the
entire
community
gets
to
see
what's
going
on
here
and
if
they
want
more
information
than
they
whoo
come
join
the
cig,
or
you
know,
talk
about
the
office
hours
to
that
Michael
was
referring
to
earlier.
L
A
O
I
was
thinking
of
that,
but
you
know
there's.
Sometimes
there
are
a
lot
of
notes
from
different
SIG's
to
cruise
through
and
and
what
I
was.
You
know,
I
think.
The
purpose
of
this
is
to
be
a
high-level
summary
for
people
who
are
deep
into
the
development
and
also
people
who
are
all
in
the
project
in
general,
who
may
not
even
know
which,
which
you
know
SIG's,
are
out
there
and
what
notes
to
peruse
three.
But
it's
much
it's
much
more
high-level
than
than
the
deep
detail
of
meeting
minutes.
Alright,.
D
Are
you
volunteering
to
distill
those
details
into
high-level
notes
for
a
weekly
newsletter,
because
that
would
be
great
to
be
yet
another
sort
of
periodic
task
where,
as
you
can
tell
I'm
a
talker,
so
yeah?
Oh,
it
seems
to
make
more
sense
like
maybe
once
a
quarter
to
have.
You
know
an
out-of-band
session
to
let
people
have
a
dialogue,
but
if
you
want
to
do
like
a
weekly
newsletter
and
I'm,
a
huge
fan
of
lwn
but
I
also
know
that's
a
lot
of
work.
O
Unless
we
have
someone
motivated
to
do
stuff
so
Mike
how
about
if
we
added
us
in
a
section
to
the
agenda
each
week,
that
would
go
in
the
newsletter
and
if
everyone,
you
know,
with
booking,
come
to
some
consensus
about
what
those
three
high-level
topics
are
I
can
just
send
them
out.
As
part
of
this
it's
last
week
in
kubernetes
development,
I,
don't
mind
that
sounds
great
okay,.
D
O
C
A
Federation
wise,
there
isn't
some
very
specific
update
except
couple
of
PRS
that
have
been
merged
and
should
exist,
and
some
really
scripts
that
you
have
open
on
question
registry
yeah
I
know
that
there
has
been
one
more
release,
cut
out,
literacy
fix,
which
was
we
yeah.
Apart
from
that,
if
there
is
some
update
Jonathan,
you
think
on
the
top
of
it
yeah.
A
K
K
So
yeah
the
big
big,
really
thing
this
past
week
was
the
V
zero
point.
Zero
point
three
release
some
other
things
that
have
been
going
on.
I
circulated
a
few
weeks
ago
with
a
smaller
group,
a
testing
plan.
It's
an
intention
to
try
to
decide
determine
what
sort
of
testing
strategy
made
sense
for
the
cluster
registry.
I
will
share
that
out
with
the
cig.
More
broadly.
Anybody
who
wants
to
take
a
look
at
that
and
maybe
contribute
comments
and
feedback
is,
is
certainly
welcome.
To
do
so.
K
One
of
the
next
things
I
wanted
to
talk
about
is
that
there
has
been
some
some
trouble
for
people
ven
during
in
the
cluster
registry,
which
is
something
that
I've
taken
note
of
there:
a
potential
strategy,
an
idea
I
had
proposed
and
as
an
issue
in
order
to
fix
this.
But
at
some
point
we'll
probably
want
feedback
on
is
I'm,
considering
splitting
out
into
two
repositories
having
the
cluster
registry.
Api
live
more
or
less
by
itself.
In
one
repository,
it
should
have
a
very
small
set
of
dependencies.
K
Api
server
and
CRN,
it
itself
would
I
think
would
probably
have
a
harder
time
of
that
since
CRN
it
and
the
cluster
registry
will
have
more
tight,
definite,
have
more
dependencies
and
have
more
tight
definitions
of
how
they
want
their
dependencies
of
what
versions
of
their
dependencies
they
want
to
use.
So
that
I
think
is
something
that
will
probably
happen.
Assuming
that
there
is
agreement
or
that's
if
somebody
else
proposes
another
strategy,
which
would
be
a
better
strategy,
we
will
move
forward
and
try
and
work
on
that
over
the
next
few
weeks.
K
Other
than
that
there
are,
we
had
a
discussion
a
few
weeks
ago
about
bringing
the
API
to
beta
and
came
up
with
two
major
points
that
we
wanted
to
to
make
sure
that
we've
worked
out
before
the
beta
timeframe,
one
of
which
was
more
contentious
than
the
other.
So
the
more
contentious
point
was
whether
the
cluster
resource,
the
CLE
object
itself,
should
be
namespace
scope
rather
than
cluster
scopes.
K
So
it's
something
that
we
want
to
consider
carefully
before
doing
it.
I
think
those
are
probably
the
main
updates
at
this
point
yeah,
as
always,
we're
looking
for
contributors
want
to
help.
There's
issues
in
the
repository
that
we
definitely
want
feedback
on,
and
there
are
things
that
our
mark
does
help
want
it
in
the
repository
yeah.
K
I,
don't
think
it's
too
early
they've
almost
had
a
beta
version
of
the
cluster
registry
by
mid-february.
At
this
point,
I
think
it
might
be
better
to
think
about
it
by
the
end
of
February,
depending
on
how
soon
the
discussions
can
be
completed.
But
I
would
hope
to
have
a
decision
on
that
issue
within
the
next
or
by
the
next
sig
meeting
to
have
a
decision
on
that
issue
and
then,
potentially
over
the
next
few
weeks,
to
do
the
implementation
work
necessary
to
enact
that.
K
I
will
have,
there
will
be
a
there
will
be
testing
plan
which
I
will
send
around,
which
is
a
Google
Doc
I
can
write
up
the
thoughts
about
the
beta
ization
of
the
API
that
we
had
try
and
coalesce
them
into
a
doc
as
well,
so
that
people
can
comment
it
on
it.
I
will
try
to
do
that.
I'm
gonna
send
that
out
both
of
those
out
later
this
week.
M
M
K
A
Okay,
thanks
thanks
Jonathan,
so
traditional
updates
are
something
that
majority
of
the
effort
over.
There
is
now
focused
on
figuring
out,
v2
API,
and
once
we
are
done
with
that,
we
would
probably
have
some
update
strategy
from
the
current
project,
our
API
and
so
far
what
we
are
doing
in
terms
of
the
key
consider
the
transition
Griffo,
what
the
verb
that
is
going
on
over.
A
There
is
one
outstanding
task
which,
because
of
because
of
the
rest
of
the
world
that
has
been
happening
around
the
v2
API,
has
taken
a
back
burner
which
was
updating
the
documents
at
the
current
documentation,
which
is
around
Perdition,
so
that
the
users
get
the
correct
idea
of
what
transition
is
happening
or
how
the
transition
is
happening.
I
should
be
able
to
complete
that
within
the
bill
in
the
coming
week
by
the
next
Signet
in
at
least,
there
is
some
update
about
the
v2
API
work
group
discussion.
That
probably
is
relevant
to
the
signal.
A
So
so
we
have
been
recently
meeting
as
regularly
as
twice
a
week
and
twice
a
week
happening
from
last
week
onwards.
So
we
had
two
meetings
last
week
and
we
had
one
meeting
on
this
Monday
this
week
and
another
scheduled
tomorrow,
so
that
API
seems
to
be
shaking
earth
taking
quite
a
good
shape,
at
least
the
areas
around
which
we
should
be
focusing.
A
B
L
D
A
B
Me
just
to
follow
on
with
that,
like
it's
I
was
a
bit
vague.
The
goal
that
I'm
trying
to
get
to
is
to
basically
have
some
few
representative
types
like
secret
replicas,
sat
and
be
able
to
define
those
federated
types
to
find
public
placement,
maybe
overrides
and
have
them
propagated
to
member
clusters
and
trying
to
sort
of
take
the
design,
as
we've
discussed
it
so
far,
kind
of
chart
a
course
through
that
and
then
give
people
an
opportunity
to
look
at
it
play
with
it.
B
A
O
A
O
J
M
It's
it's
compiling
together,
a
few
things
like
JSON,
it's
terraform
and
general
cloud
provider,
specific
things,
and
it
was
a
way
of
having
a
common
staged
rollout
for
your
kubernetes
cluster
for
any
given
cloud.
Back-End
that
someone
shares
to
implement
it
for
the
first
step
would
basically
be
like
select
how
you
want
your
cluster
to
kind
of
be
designed,
then
select.
M
If
you
want
to
use
something
like
chorus,
bootstrap,
ignition
or
kubernetes
or
Kubb
ADM,
then
the
third
stage
was
actually
used,
coop
ADM,
to
provision
the
cluster
on
top
of
your
machines,
which
was
using
terraform
to
get
the
machines
and
then
the
final
step
was
launch.
Any
additional
add-ons
feels
very
similar
to
like
what
some
of
the
examples
were
for
the
cluster
API
at
least
moving
forward.
The
GCP
example,
if
I
saw
so
that's,
maybe
kind
of
what
drove
me
to
bring
it
up
here.
D
We
could
be
like
dam
where
VMs
or
a
bare-metal
storage
or
the
network
infrastructure,
if
I
understand
it
correctly
and
I
may
not
be
I.
Think
what
you're
describing
are
that
the
central
operations
that
you
need
to
portably
provision
and
manage
a
kubernetes
cluster.
You
know
in
a
bunch
of
different
places
to
me.
I
think
that
the
scope
of
this
group
was
already
pretty
wide,
has
to
deal
with
the
interactions
between
clusters
sort
of
more
of
a
north
direction,
whereby
no,
how
do
we?
D
M
J
D
So
I
guess
my
youichi
talked
to
the
cluster
lifecycle.
Folks
I
talked
to
them
a
bunch
and
you
know
we
can
take
this
offline.
Maybe
I
don't
want
to
take
too
much
signal
to
cluster
time
on
it,
but
I
think
my
understanding
is
kubernetes.
Everywhere
is
sort
of
an
effort
and
a
goal
and
cluster
api
is
a
product
to
help
us
achieve
that
goal.
D
D
D
M
Hoping
that
helps
yeah
I
think
that
TLDR
is
that
cluster
API
would
potentially
be
used
downstream
like
open.
It
is
anywhere,
but
doesn't
inherently
replace
it.
D
D
A
D
O
Consensus
about
what
the
top
two
or
three
things
we
wanted
to
discuss
for
this
meeting
in
the
news,
whether
we're
going
to
be
the
biggest
thing
like
the
takeaway
that
I
was
following,
was
that
we're
gonna
start
in
office
hours
meeting
that
we
need
to
kind
of
evangelize
and
get
out
there
if
they
were
going
to
be
to
other
things
that
we
add.
What
would
that
be.
A
L
J
G
Been
turning
the
last
few
as
few
meetings
action
I
was
involved
initially,
but
now
we
are
getting
very
very
serious.
I
was
yeah.
Thank
you.
Well,.