►
From YouTube: Kubernetes SIG Multicluster 20180703
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
I'm
wondering
how
many
people
were
gonna
get
at
this
meeting.
I
know
that
the
Huawei
guys
are
probably
watching
soccer.
A
A
A
A
A
B
I
see
yeah
because
I'm
not
only
working
for
this
service
smash
across
mod
cluster,
but
I'm
also
working
for
the
mod
classes,
ports
for
the
communities.
So
we
have
two
requirements
so
the
first
one
that
we
want
to
make
the
multi
class
or
too
many
the
resources
is
to
manually
corinthis
resources
and
another
one
that
we
want
elaborately
easier
to
manage
all
the
applications
running
on
table
alternatives.
B
So
my
requirements
for
the
federate,
node
and
fell
report
is
actually
the
requirement
for
the
resource
demand
for
the
clusters,
because
there
are
some
customers
want
to
have
an
overview
of
all
the
resources
that
they're
managed
across
different
clusters.
So
she
sees
a
cura
requirement
that
they
want.
Only
one
and
component
will
all
be
resulted
from
differences
right.
A
Not
sure
that
Federation
is
necessarily
a
good
fit
directly
like
the
idea
of
pods
like
pods
and
nodes
are
explicitly
not
supported
in
Federation,
like
that's
by
design,
not
by
accident,
and
it's
not
that
there
wouldn't
be
the
potential
for
adding
kind
of
like
an
aggregation
layer
on
top
that
would
collect
data
about
underlying
clusters
and
allow
it
to
be
viewed
through
like
a
single
pane
of
glass.
That
sort
of
thing,
but
Federation,
isn't
really
designed
to
have
be
a
central
point
for
propagating
configuration
down
to
clusters.
A
Well,
there
is
I'm
not
sure
if
there's
issues
for
it
I
think
there
is.
There
are
plans
to
consider
different
mechanisms
for
reporting
and
aggregating
status
of
underlying
clusters.
It's
not
strictly
like
a
federation
capability.
It
could
be.
It
could
very
well
be
something
that
is
implemented
and
so
that
it
can
be
deployed
separately
from
Federation
that
make
sure
I'm.
Okay,.
B
A
What
multi
cluster
is
kind
of
like
a
very
broad
sort
of
problem
space
that
basically
involves
any
problem
of
coordination
across
multiple
clusters.
Federation
is
a
more
specific
special
case
in
which
the
goal
is
to
provide
a
way
to
define
configuration,
and
you
want
to
be
common,
largely
common
across
clusters
and
mechanisms
for
scheduling
things
into
different
clusters,
but
it's
basically
a
subset
and
the
reason
that
that
change
happened
last
year
was
the
Federation
was
a
very
monolithic
solution
that
didn't
solve
everyone's
problems
and
it
was
kind
of
difficult
to
support
new
use
cases.
A
The
way
it
was
implemented,
and
so
the
decision
was
made
to
broaden
the
scope
of
the
sig
to
be
multi
cluster,
which
could
encompass
solutions
other
than
Federation
and
kind
of.
In
addition
to
that,
Federation
was
sort
of
rebooted
this
year
to
be
more
modular
to
be
implemented
in
a
way
where
bits
and
pieces
could
be
used
independent
of
one
another,
and
it
would
make
it
easier
for
contributors
to
add
features
to
support
the
use
cases
without
having
to
know
about
everything,
because
it's
it's
less
monolithic.
A
In
terms
of
supporting,
like
features
like
like
reporting
like
aggregating
status
across
clusters,
I
think
that's
definitely
a
valid
multi
cluster
use
case.
I.
Think
in
the
near
term,
there
are
there's
a
desire
to
experiment
with
different
mechanisms
as
collecting
and
reporting
status
within
Federation
I'm,
not
entirely
sure
that
pods
and
nodes
would
be
in
the
scope
of
Federation.
Just
because
it's
not
something
we're
doing
right
now,
but
I
think
it's
definitely
within
scope
of
a
multi
cluster.
It's
just
sort
of
considering
like
like
documenting
the
use
case
like
Pacifico
way.
A
But
but
the
goal
I
think
would
be
providing
enough
detail
as
to
like
what
you
want
to
do
and
why
so
that
people
can
provide
feedback
and
and
sort
of
I
mean
and
also,
if
it's
something
other
people
will
find
useful,
then
maybe
you
can.
You
can
get
other
people
to
help,
but
the
challenge
is
kind
of
like
coming
in
and
saying
I
want
this
feature
without
providing
the
backstory.
It
makes
it
difficult
for
people
to
reason
about
whether
it's
something
they
want
to
and
whether
they
want
to
contribute
to
implementation.
A
Yeah,
okay,
I
want
to
thank
you
by
the
way
I
mean
you
kind
of
come
into
the
scene
and
like
you've
been
implementing
like
enjoying.
We
haven't
had
a
lot
of
outside
contribution.
At
this
point
it's
been
largely
RedHat
and
why
away
from
a
code
perspective,
so
very
much
appreciate
your
contribution.
I
just
wanted
to
make
sure
that
we're
kind
of
converging,
rather
than
diverging
in
terms
of
intention
yeah.
A
C
B
Hope
that
the
mode
classroom
meeting,
maybe
I'm
not
sure
if
we
can
risk
areas
but
yeah
I,
think
I
call
the
feedback
to
answer
the
proposer
that
maybe
we
can
add
another
meeting
or
we
can
and
use
some
elsewhere,
such
as
dance
one
week
you
asked
a
man
not
week
in
Chatham,
but
I'm,
not
sure.
If,
if
we
can
do
this
because
internally,
we
may
not
have
you
know,
people
contributed
products,
yeah,
so
yeah.
A
Okay,
but
I
guess
we'll
need
to
have
some
more
sort
of
stakeholders
chime
in
if
it's
a
Federation
meeting,
it's
sort
of
a
smaller
group
of
people
and
I
think
that's
more
doable.
The
multi
closer
meeting
is
a
little
bit
harder
because
there's
contributors
from
Google
that
I
mean
a
lot
of
the
meetings
are
9:00
a.m.
or
later
Pacific,
just
because
that's
regular
business
hours
for
Google
in
the
Pacific
time
zone
so
we'll
have
to
see
how
they
feel
about
shifting
that
forward.
B
A
Of
the
repos
there's
also
Cuba,
NCI
and
there's
also
cluster
registry
I
think
those
are
the
three
primary
development
efforts
being
sponsored
by
the
sig.
At
this
point,
yeah
and
chances
are
teacher.
The
Federation
will
probably
spawn
more
repos.
The
goal
is
to
try
to
break
things
up,
so
it
can
be
consumed
in
too
badly.
So
the
idea
of
having
like
one
repo
is
this
kind
of
how
kids
started
but
I
think
there's
there's
a
move
away
from
that,
because
it
becomes
difficult
to
scale
yeah.