►
From YouTube: Kubernetes SIG Multicluster 20171121
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).
B
A
So
there
isn't
much
I
guess
to
discuss
today,
so
we
might
end
up
meeting
short
unless
otherwise
there
is
somebody
comes
up
with
some
other
points
to
talk.
One
of
the
points
that
we
had
in
the
previous
meeting
was
about
subgroup
that
tradition
working
group
through
by
hand
the
same
and
starting
Nautica
for
the
next
meetings
for
it.
So
so
far
I
have
seen
to
the
same
9:30
PT
when
this
day
is
the
different
time
by
people.
Who
are
the
responded
to
that?
A
A
C
C
C
C
F
I
think
I
think
it
would
be
better
to
like
I
think
we
know
the
day
of
the
week
and
the
time
I
think
this
week
is
kind
of
an
exception,
because
there's
a
holiday
in
the
u.s.
that
probably
a
lot
of
people
are
gonna,
be
traveling
for
so
I
I
would
say
that,
like
we
should
definitely
meet
next
week.
I'm,
not
we
only
got
eight
responses
to
this
poll.
It
looks
like
so
it's
not
clear
exactly
how
many
people
are
going
to
be
in
at.
F
A
I'm
I
mean
I
personally
can
say
that
it
is
fine
again
if
we
start
next
week
also,
and
if,
if
there
are
people
from
the
group
who
actually
selected
the
time,
some
of
them
are
not
able
to
attend.
Probably
it
might
be
okay,
but
some
necessary
guys
should
at
least
be
like
redhead
somebody
should
represent.
Who
are
we
I
know
that
they'll
be
couple
of
people?
I
will
be
you
in
the
back
of
the
Clinton
will
also
find
anybody
interested
from
Google
to
join.
A
B
A
A
So,
but
what
we
are
seeing
right
now
is
that
we
don't
have
much
of
especially
reviewers.
We
don't
have.
There
are
little.
There
is
a
list
of
people
who's
there
as
parts
of
owners
and
reviewers,
but
I
am
not
sure
if
all
of
them
are
actually
actively
participating.
So
would
it
make
sense
to
have
a
relook
at
that
yeah.
B
The
amount
of
time
that
we
plan
on
spending
reviewing
so
I
I,
probably
provoke
propose
to
thing,
would
be,
did
to
trim
down
the
amount
of
Googlers
on
that
list,
because
it's
not
representative
and
I
probably
motion
to
at
this
point,
I
think.
The
last
time
I
saw
we
had
a
few
people
who
were
approved,
started
reviewer,
but
not
yet
approver
at
this
point,
I
think
they
would
be
eligible
for
being
approvers
now
I.
B
A
A
Okay
about
that
meeting
that
was
supposed
to
happen
tomorrow
we
I
mean
I
was
trying
to
think
about
some
agenda.
One
director
pointer
was
that
we
had
a
couple
of
meetings
earlier
and
there
was
a
note
that
workgroup
notes
were
maintained.
We
can't
proceed
from
that
or
we
can
set
up
a
new
agenda
any
citations
for
that.
This
is
a
brief
of
notes.
I
think.
C
F
C
Having
that
discussion,
I
think
that
there
was
this
idea
that
we
were
just
gonna
vendor
in
cluster
registry
and
maybe
that,
as
close
to
registry,
has
sort
of
become
a
thing.
Maybe
that's
not
a
viable
option
because
it
kind
of
me
in
this
case
somebody
wanted
a
cluster
registry,
and
you
know
they
would
have
to
just
pick
it.
They'd
have
to
choose
between
Federation
or
cluster
registry
versus
yeah.
I
have
a
in
cluster
registry
that
can
be
aggregated
and
then
you
can.
You
know
you
don't
have
to
choose
yeah.
F
C
B
B
F
F
So
me
comment.
Yes,
it's
up
already
I
think
that
that
Ivan's
working
too,
like
the
last
mile
of
like
it,
appears
based
on
what
I've
seen
that,
like
the
aggregation
part,
seems
to
be
working
correctly,
but
now
he
needs
to
had
delegated
autonomy
into
the
cluster
registry.
Api
server
for
the
whole
shebang
to
work
right.
F
B
A
So
what
I
wanted
to
know?
If
somebody
is
there
in
this
meeting
right
now,
passively,
listening
or
whatever
it
is
they're
somebody
who's
willing
to
contribute.
They
can
certainly
sign
up,
probably
by
attending
the
next
work,
took
meeting
that
happens
of
addition,
or
they
can
directly.
I
am
one
slack
okay,
and
what
I
wanted
to
clearly
know
was
currently.
Is
there
any
problem
the
because
it's
a
little
tricky
to
set
up
the
Federation
and
say
start
developing
something
in
it?
If
there
is
any
problem,
they
can
voice
this
over
there.
H
B
H
There
you
go
yeah
thanks,
so
yeah,
so
that
stuff
is
complete,
so
you
can
now
do
all
the
building
and
and
exploring
in
the
repo
without
having
to
so.
It
can
be
also
vendor
din
as
well,
since
it
now
contains
all
the
generated
code
and
the
vending
code
as
well
he's
also
worked
on
getting
the
community
docs,
the
P
are
merged,
so
the
plan
and
the
API
for
cluster
registry
has
been
merged
and
he's
spent
time
going
through
and
creating
issues
to
reference.
H
So
that's
a
good
progress
there
and
then,
as
was
previously
discussed,
I'm
working
through
getting
the
aggregated
and
delegated
changes
to
the
cluster
registry,
both
in
the
CRN
it
tool
and
the
cluster
registry
API
server.
So
the
the
work-in-progress
PR
is
up.
If
you
want
to
like
poke
around,
it's
not
ready
to
be
reviewed,
but
it's
certainly
up
okay,
so
those
are
kind
of
latest
happenings
for
cluster
registry.
A
Yeah
then
next
is
CI
updates.
So
this
this
is
a
bit
about
the
Federation
CI.
So
there
used
to
be
a
soap
job
for
Federation,
which
we
have
the
move
right
now.
Does
anybody
see
the
need
of
the
same
or
use
it
of
the
same,
so
that
I
mean
usage
of
resources
versus
the
utility
of
the
same
any
commands
somebody
might
have
I
have.
C
C
That
are
more
important
than
worrying
about
scale.
I.
Think
the
one
of
the
biggest
things
that
we
need
to
address
in
the
near
term
is
to
have
automated
validation
of
more
recent.
The
most
recent
version
of
cube
that
we
can
vendor
base
in
the
state
of
the
staging
repos.
The
present
we
can't
automate
that
there's
no
like
just
a
manual
process
to
determine
what
what
hash
of
cube
Rico
we
can
use,
but
fixing
that
is
is
probably
I.
C
A
G
Yeah,
actually
yeah
so
I
think
the
CA
job
currently
does
validation
against
latest
commits
from
kubernetes
right
now,
so
we
are
going
to
make
resubmit
job
or
to
validate
against
vendor
kubernetes
instead
of
our
latest
kubernetes
commit.
So
that's
that
change
I
am
going
to
work
in
the
coming
week.
Probably
that
that's
the
one
I
think
Maru
is
talking
about
right,
not.
C
C
C
We
just
need
a
place
to
publish
those
results.
It's
not
as
I
mean,
there's
definitely
more
work
to
be
done.
On
the
staging
repo
side.
That's
all
I
mean
and
that
that'll
take
a
long
time,
but
the
short
term
fix
is
is
simply
being
able
to
automatically
discover
which
versions
of
kubernetes
in
the
staging
repo
we
can
safely
vendor.
G
C
Can
do
that,
but
because
we're
essentially
chasing
head,
that
means
that
we're
always
on
a
leg.
So
maybe
that's
a
good
stopgap
but
I
would
say
ideally
we're
gonna
want
to.
Maybe
we
don't
want
to
Reve
ender
like
against
head
all
the
time,
but
we
want
to
be
able
to
test
against
head
as
often
as
possible
to
detect
breakage
yeah,
and
so
it's
not
like.
We
want
a
vendor.
We
just.
We
need
to
be
able
to
know
which,
which
versions
of
cube
and
the
staging
repos
we
can
safely
test
against.
Essentially.
G
A
Okay,
okay,
yeah:
we
can
talk
more
about
this
offline,
also
yeah.
We
also
noticed
that
some
ingress
tests
English
currently
that
it
English
currently
works
for
GC
and
some
of
those
tests
are
broken
and
they
have
been
so
since
quite
some
time.
So,
although
they
are
not
necessarily
a
priority
for
us
right
now,
but
if
somebody
is
willing
to
volunteer
or
if
somebody
wants
to
have
a
look
and
just
he's
taken.
C
C
The
past
google
has
been
maintaining
that
and
they've
had
the
expertise.
Not
it's
not
really
clear
to
me
where
we
go
in
the
future.
If
they're
not
going
to
be
maintaining
that-
and
it
may
be,
that
we
have
to
look
at
integrating
whatever
tooling,
they
intend
to
support
their
customers
separately
from
federation
like
Nikhil's
work
on
an
ingress
configuration
solution.
Maybe
we'll
have
to
just
reuse
that
because
then
we'll
have
something
and
it's
supported
by
like
Googlers.
A
B
B
C
Think
continuing
to
Nora
Nora
is
is
the
path
for
now
and,
like
I,
said
I
I
think
that
if
we
can
find
a
solution
where
we
can
reuse
integration,
that
Google
is
actually
willing
to
support.
That
would
be
ideal.
So
I'm
not
sure
we
don't
have
enough
resources
to
you
know
be
cloud
specific
at
this
point
right.
C
I
think
the
goal
also
I
mean
we
also
have
integration
for
federated
services
for
Google
DNS
and
core
DNS
and
route
53
I
think
a
near-term
goal
needs
to
be
enabling
extension
for
things
that
are
cloud
specific,
so
it
can
be
done
out
of
repo.
If
someone
cares
to
implement
it
support
it,
it
can
be
done
without
burdening
the
core
effort
by
things
that
we
I
don't
know
it's
just
not
it's
not
general.
That
makes
sense.
B
A
A
Okay,
so
we
are
issues
this
week.
We
don't
have
any
specific
her
which
need
attention
yeah.
This
was
the
Solana
Kubik
on
which
we
have
a
reserve
slot.
Is
there
something
in
the
last
thing
that
we
had
whoever
was
attending
on
what
was
what
was
willing
to
attend
the
but
probably
compiled
what
they
are
going
to
present?
Maybe
put
a
link
over
here,
I'm,
not
sure
fiction.
Are
you
taking
the
lead
in
this
or
somebody
else,
I.
B
Of
course,
if,
if
Quinton's
there-
and
he
rather
speak
for
what's
going
on
in
Federation,
I
went
I
wouldn't
mind
if
he
took
that
part
I'm
to
me,
the
bigger
unknown
is
the
salon
and
what
we
want
to
make
out
of
that
two-hour
block
I
a
lot
of
it
as
depending
on
how
much
we
get
done
next
week
in
in
the
Federation
working
group.
I
think
most
of
the
time
should
be
spent
for
Federation
working
group.
I,
don't
see
any
any
raised
blockers
any
blocking
issues
or
PRS
in
cluster
registry.
B
There's
really
only
two
people
working
on
it
right
now
and
I'm,
not
sure
they're
there
President
Jonathan
will
not
be
there.
So
basically
comes
down
to
do.
We
use
this
to
our
block
efficiently,
given
that
not
everybody's
there,
or
we
just
kind
of
cancel
it
or
get
together
with
what
are
people's
opinions.
You.
D
Through
some
of
the
other
stuff
that
I
see
out
there
for
what
other
teams
are
doing
with
their
salon
time,
it
seems
like
some
are:
providing
a
more
detailed
update
to
the
community
and
also
allowing
folks
an
opportunity
to
work
with
the
technology
hands
on
I.
Don't
know
you
know
who's
gonna
be
there.
That
could
support
that.
I
can
certainly
help
and
I'm
willing
to
to
help
out.
If,
if
we
get
at
least
agree
on
what
it
is,
we
want
to
cover
I.
B
C
I'm
not
sure,
there's
precedents,
I
mean
I
was
acute
con.
The
last
couple
of
years
and
I
mean
last
year.
There
was
a
lot
of
presentations
during
the
main
conference
days
and
then
he
had
kind
of
an
unconference
day.
Where
was
mainly
just
developers
like
not
really
users
so
I,
don't
really
know,
I
think
that
they're
kind
of
evolving
at
it
and
I'm,
not
sure
anybody
has
experience
with
how
its
involved
in
this
misfit
iteration.
D
Yeah
I'm
not
sure
you
there
I've
been
to
all
the
group
cons,
but
this
is
going
to
be
the
biggest
one
ever
in
terms
of
attendance,
so
it's
gonna
be
tough
to
gauge.
You
know
how
many
folks
would
be
interested
in
and
with
the
right
topic
is
going
to
be
for
the
audience.
That's
that's
gonna,
be
tough,
I
think
what
we'll
find
with
so
many
people
attending.
It
is
possibly
a
lot
of
folks
that
could
use
the
more
you
know,
level.
C
I
think
that
there's
there
I
would
hope
that
there's
been
enough.
I
mean
people
have
been
sort
of
excited
about
Federation's
potential
in
the
past,
because
it's
been
sold
pretty
well
like
in
the
regular
conference
talks
I
mean
if
we
could
actually
get
users
who
have
you
know,
feedback
and
put
whatever
on
what
they
would
like
to
see
not
just
for
a
while
they
just
in
a
in
a
multi
cluster
context.
I
mean
the
common
complaint
is
I.
Have
lots
of
clusters
I
have
difficulty
managing
them
all.
C
There's
not
really
any
community
supported
way
to
do
that
the
Federation,
maybe
isn't
what
I
want,
or
maybe
it
is,
but
I
guess
I'm
mine.
This
is
asking
questions
for
me
like
how
do
we
actually
know
who's
going
to
attend
what
the
number
is
going
to
be?
You
know
what
type
of
users
or
developers
they're
going
to
be.
Is
there
any
way
to
find
that
out
people
just
kind
of
randomly
show
up
I.
D
Think
we
could
see
who
is
listed
as
added
it
to
their
schedule,
but
that
might
not
be
representative
of
who
actually
shows
up
you
know.
Sometimes
you
will
put
something
in
their
schedule
and
then
decide
not
to
be
there
and
then
other.
You
know
you
can
also
see
a
lot
of
people
that
just
show
up
that
haven't
put
it
in
their
schedule.
C
D
I'm
happy
to
put
more
work
into
it
and
you
know
try
to
put
together
a
straw
man
what
we
should
cover
and
then
review
it
with
the
group
I'll
take
that
on
and
try
to
get
it
done
this
week.
D
But
but
if
you
have
ideas,
please
let
me
know
too,
because
I,
you
know,
my
own
ideas
might
be
different
than
what
you
guys
think
we
should
cover.
So
I
was
thinking
something
like
you
know.
If
we
can
show
people
how
to
use
Federation
or
multi
cluster,
just
with
mini
cube.
You
know
something
really
simple,
that
we
know
wouldn't
require
a
lot
of
external
dependencies
than
them.
That
would
be
good,
but
I,
don't
know
how
possible
that
is.
A
I
can
also
say:
I
support
in
I
mean,
for
example,
there
could
be
a
simple
hands-on
which
says
that
this
is
how
you
can
deploy
the
control
plane,
or
this
is
how
you
can
actually
create
a
Federation
in
tiny
things.
All
most
of
the
stuff
is
there
and
which
is
their
interpretation,
but
the
users
didn't
try
it
out.
Yeah.
B
I
think
the
point
Dan
makes,
which
is
valid,
is
demonstrate.
I
mean
there's
a
lot
of
demos
on
Federation,
but
at
least
what
I'm
hearing
with
mini
cube
is
demonstrate
how
it
is
possible,
if
you
lower
the
barrier
to
development
entry,
to
get
more
people
to
hop
on
like
I.
Did
it
does
the
compelling
story,
I.
C
C
When
I've
been
doing
local
testing,
I,
don't
use
mini
cube
because
Rosie
foster
up
is
a
little
frankly
a
little
bit
nicer
in
a
Linux
box,
but
same
thing,
where
I
bring
up
a
cluster
I,
deploy
the
control
plane
to
it.
I'm
gonna
join
the
cluster
yeah.
It's
not,
you
know
spreading
load
or
dealing
with
multiple
clusters,
but
it's
it's
the
simplest
possible.
You
know
deployment
of
Federation,
it
lets
people
sort
of
use
it.
D
C
No
I
shouldn't
sorry
I'm,
not
trying
to
advertise
for
mini
Hulme
I
think
it's
probably
a
better
solution
for
people
who
aren't
OpenShift
savvy.
It's
just.
If
you
know
it
it'll
run
on
local
docker,
Linux
box
versus
mini
cube.
They
only
run
it
a
vm,
so
it
just
adds
overhead
for
developer,
but
for
a
user
everybody
keeps
a
better
option.
Has.
D
H
H
That's
documented
on
running,
like
pac-man,
the
web-based
pac-man
application,
that's
federated
across
cloud
providers,
so
you
could
either
do
like
gee
ket
or
you
could
do
AWS
and
uke,
and
it's
even
documented
with
Azure.
Even
though
there's
some
DNS
issues
with
Azure
have
since
been
fixed,
but
there.
D
Might
be
a
good
thing
to
show,
as
you
know,
like
one
of
the
the
walkthroughs,
that
they
don't
do
themselves
like
here's,
something
really
cool
that
you
could
do
with
Federation
or
instead
of
like
having
them
walk
through
it
I'm
not
opposed
to
doing
that.
I
think
that
you
know
highlighting
something
like
that
would
be
really
cool
for
people
to
see.
You
know
live
and
in
person
something
like
that
working
I
think
as
far
as
running
it
locally
on
their
their
laptops
or
trying
it
themselves,
it
might
be
a
little
too
complex,
mm-hmm.
H
C
E
C
D
D
C
B
Okay,
I
guess
suppose
we
get
the
agenda.
I'll
try
to
get
I.
Take
somebody
and
see.
If
you
have
contains
the
description
of
what
these
sessions
are,
then
the
fact
that
you
knew
what
the
other
sessions
were
doing
was
probably
speaks
to
the
fact
that
they
are
promoting
their
sessions
better.
So
we'll
try
to
get
an
updated
description.
A
A
Is
there
any
any
specific
pointers
so
that
we
need
to
separately
talk
about
this
like
a
this
maintains
proper
release
process
and
there
are
roles
which
are
which
actually
are
taken
up
by
people
and
set
of
issues
and
all
that
stuff
which
is
marked
for
the
treaties.
But
here
we
don't
necessarily
have
that,
because
we
are
sort
of
carrying
over
a
lot
the
state
what
it
was
say
a
month
back
and
then
using
the
same.
A
Finally,
in
the
same
shape,
and
probably
some
bug
fixes
in
so
any
other
suggestion,
apart
from
that,
like
we'll,
be
able
to
cut
out
the
tar
files
and
being
probably
able
to
find
out
a
bucket
or
a
location
which
is
public,
and
we
can
publish
that
this
is
there,
the
binaries
are,
apart
from
that,
is
there
anything
else?
Do
we
need
to
take
care
about
this
I.
C
A
C
No
real
plan
for
coordinating
release
of
sub
projects
with
cube,
so
we're
kind
of
pioneering
that
here
I'm
assuming
that's
true,
for
you
know,
cluster
I
just
do
or
anything
else,
I
guess,
I'm,
not
sure
cluster
industry
is
intending
to
relay
the
same
requirement
of
compatibility.
That
Federation
does
but
anyway,
all
have
to
say,
like
we
can
put
the
file
somewhere.
I
was
just
you
know,
there's
there's
a
certain
amount
of
coordination
that
goes
on.
You
know
around
the
release,
artifacts
around
blog
posts.
You
know
all
that
sort
of
stuff
I
mean
I.
C
C
Is
here
too,
and
just
have
pointers
to
the
release
artifacts?
Recently:
okay,
but
I,
don't
know
I've
been
kind
of
hesitant
to
want
to
drag
the
release
team
into
thinking
about
that
because
they
got
their
hands
full,
just
trying
to
stabilize
things
right
now.
Coach
reason
all
so,
maybe
when
things
are
settled
down
they're
just
waiting
for
things
to
clear,
we
can.
A
Okay,
but
that
makes
sense-
and
in
case
we
don't
have
any
specific
requirements
directly
I
mean
to
collaborate,
the
dub
person
who's,
writing
the
blog,
putting
it
up
at
the
same
place.
We
can
always
have
for
us
simply
log
for
Federation
as
saying
that
that
alesis
reasons
here
or
whatever
additional
features
are
there.