►
From YouTube: Kubernetes SIG Cloud Provider 2018-10-03
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
All
right
30
so
today
is
October
3rd.
This
is
kubernetes
supervisor
meeting.
So
in
the
past
it
was
pretty
bad.
What
Chris
has
been
pretty
good
about
it,
but
I've
been
pretty
bad
about
doing
intros
at
the
beginning
of
every
sick
meeting.
So
there's
anyone
new
to
the
food
today
just
wanted
to
give
like
a
minute
or
two
just
to
give
a
chance
to
introduce
yourself.
B
C
D
A
E
Yeah
I'll
just
give
it
a
quick
thing
and
then
leave
it
to
Canberra
too
so
I
just
wanted
to
say,
hi
and
introduce
myself
I'm,
Aisha
and
I'll
be
the
113
release
lead
for
Cuban
ities,
which
started
the
recyclates
have
started
on
Monday
and
it
goes
till
the
first
week
of
December.
So
currently
we
are
planning
to
release
on
the
3rd
of
December,
which
means
it's
a
super
short
cycle
with
just
ten
weeks,
I
sent
out
an
update
to
say
joke
abilities
developers
about
the
schedule
itself
and
a
few
process
changes
there.
E
So
we
just
wanted
to
take
this
time
to
do
it.
To
means
to
just
highlight
that
the
release
is
super
short
and
hence
encourage
the
very
sick,
leads
and
feature
owners
to
plan.
Any
features
are
keeping
that
constraint
cycle
time
in
mind
and
Kendrick
was
going
to
will
start
formally
collecting
features
starting
next
Monday
October
8th.
So
in
case
you
feel
anything
needs
an
issue
open
in
the
features
repo.
E
Please
go
ahead
and
do
that
and
I
also
specifically
wanted
there
weren't
any
dogs
updates
that
were
being
planned
as
part
of
113
from
the
Dell
providers
day
itself.
I
remember
seeing
something
about
this
in
the
community
meeting
last
week,
so
I
just
wanted
to
find
out
more
about
it.
If
you
guys
have
any
info
at
this
point
Kendrick
do
you
want
I,
wanted
to
add
anything
yeah.
D
The
only
other
thing
really
there's,
there's
kind
of
just
one
big
outstanding
feature
here
and
that's
of
course,
out
of
tree
cloud
provider
support
that's
with
inside
of
the
key
features
repo.
It's
currently
not
tagged
with
a
milestone,
however,
with
inside
the
notes
and
said
that
it
could
make
it
in
113
could
make
it
to
114
depending
on
if
it's
gonna
be
stable
enough.
So
if
you
know
what
that's
gonna
be
now
that'd
be
cool.
I
can
tag
it.
If
not.
If
we
want
to
defer
it
to
114,
then
we
can
do
that.
D
A
B
G
We
are
we
committing
on
effect
that
we
want
to
GA
all
the
providers
out
of
tree
or
just
that.
The
fact
that
the
interface
that
we
have
is
considered
GA,
because
the
latter
I
think
it's
more
achievable,
obviously
having
everybody
on
a
tree
by
113
is
not
realistic
when
it
comes
to
the
actual
interface,
which
is
I,
guess
what
we
can
call
a
feature
in
kubernetes
I,
think
that
is
realistic.
Yeah.
H
Yeah,
but
maybe
there's
a
few
things:
we've
got
the
interface,
we've
got
the
out
of
tree
providers
and
we
have
the
current
batch
of
entry
providers
I.
Think
CCM
working
without
a
tree
providers,
probably
GA.
If
andrew
is
happy
with
that.
I'd
love
to
get
the
interface
to
at
least
be
in
staging
but
I
think
anything
beyond.
That's
probably
not
doable,
and
my
guess
is
that
for
the
for
the
majority
of
the
entry
providers,
just
getting
a
a
sort
of
alpha
level
out
of
tree
experience
is
probably
going
to
be
a
stretch
for
113.
D
Yeah
absolutely
I
kinda
have
the
same
exact
thoughts
as
what
Fabio
had
mentioned
as
well.
There's
I'm
not
going
to
anticipate
any
out
of
tree
providers
or
gonna,
be
ready
to
go
especially
GA,
because
I
know
there's
just
a
few
that
are
targeting
for
just
having
some
sort
of
a
release
for
113
I.
Think
why
I
was
more
or
less
looking
at
the
interface
itself
that
that's
gonna
be
ready
to
go
for
the
whole
ecosystems
are
developing
around
it.
Yes,.
G
There's
some
sort
of
blocker,
like
you
know,
kind
of
a
psychological
blocker
and
you
know
holding
back
a
little
bit
on
the
work
for
a
tree
co-op
provider
because
they
interface
is
not
considered
stable
or
GA.
I
feel
like
there's
like
this,
which
is
a
misunderstanding,
because
interface
has
been
there
for
a
long
time
so
about
two
things
like
validating
that
and
making
it
official
that
is
GA.
So
people
can
be
more
confident
and
spending
time,
and
you
know,
working
on.
Pre
providers
will
be
a
nice
step
forward.
I
H
Sorry
I
think
we're
gonna
be
very
careful
what
we
define
in
that
grouping.
For
instance,
no
lifecycle
controller
still
pulls
in
cloud
from
the
cloud
provider
interface
and
while
I've
been
working
with
the
signo
to
get
that
gone,
I
I.
Don't
think
that
that's
necessarily
something
they've
committed
to,
and
that's
just
one
example
of
places
in
core
kubernetes,
where
we
still
have
a
pretty
strong
reliance
on
the
cloud
provider.
Interface.
C
That's
a
good
point
and
I
was
gonna.
Just
suggest
that,
typically,
when
an
API
moves
from
beta
to
stable,
it
requires
some
amount
of
soap
time
and
feedback
loop
in
order
to
ensure
that
the
API
is
in
fact
stable.
It's
very
hard
backwards.
Compatibility
guarantee
that
we're
offering
when
we
move
that
API
from
beta
to
stable-
and
it
is
really
unfortunate
when
we
find
that
we
either
have
a
breaking
change
after
version
the
API.
That's
much
more
challenging
to
accomplish
so
I.
Think.
C
C
A
E
F
B
A
Yeah
yeah,
so
there
seems
to
be
two
different
efforts
for
that.
So,
like
one
is
from
Zig
Doc's
around,
how
do
we
expose
users
and
then
the
other
one
is
like
what
we're
kind
of
working
on,
which
is
what
doc
mutation
does
every
provider
need
to
work
on,
that
we
can
kind
of
loop
together
into
that
other
doc,
and
so
just
to
just
to
clarify
that
what
we
are
pushing
for
is
the
latter
and
what
we
were
kind
of
hoping
that
sick
Doc's
would
do.
F
C
J
F
But
then
Cody
and
I
are
also
looking
at
at
the
you
know
flow
from
like
the
docs
homepage,
to
that
section,
where
the
cloud
providers
will
be
explaining,
like
the
difference,
just
solution
types
in
general,
so
that
users
have
a
better
understanding
of
like
you
know
what
might
be.
You
know
better
for
their
particular
situation.
Yeah.
J
The
reason
I
ask
is
I'm
working
on
some
pieces
of
documentation
right
now
and
I
don't
want
to
have
I,
don't
want
to
wait
until
the
structure
rules
are
in
place
because
I
edited
and
reformat
it
as
needed,
and
it
shouldn't
stop
documentation,
effort
that
is
underway
now
and
and
maybe
even
need
it
now.
Yeah
yeah.
F
I,
don't
think
that
it
should
hold
back
anyone
or
you
know
a
blocking
issue.
I
think
these
things
can
go
on
parallel
and
then
we
should
at
some
point
and
link
up
and
just
make
sure
that
it
all
you
know,
works
well
together,
but
yeah
I,
definitely
don't
want
to.
You
know,
hinder
your
documentation,
efforts
and.
C
To
be
clear,
there
andrew
is
actually
representing
two
roles
in
this
one
conversation
and
one
role.
He
is
the
GCP
representative,
who
is
bound
by
the
decisions
that
this
SIG
has
made
about
the
organization
and
structure
of
the
cloud
provider
document
documentation
in
the
other
role
he's
representing
cig
docks,
which
is
taking
on
the
how
to
reference
those
documents
that
documentation
across
cloud
providers
yeah.
C
J
A
Issues
and
to
choose
four
new
contributors
so
based
on
a
conversation
had
with
Caleb
a
while
back
and
initially
recently
we're
noticing
that
there's
a
lot
of
like
kind
of
easy
to
medium
issues
that
we
have
under
state
cloud
provider
that
we
can
be
kind
of
offloading
to
new
contributors.
Instead
of
us
trying
to
tackle
it
ourselves,
and
so
this
is
kind
of
something
I
figured.
We
can
try
out
and
see
if,
if
it'll
help,
people
kind
of
with
membership
or
whatever.
A
So
if
you
have
kind
of
like
generic
cloud
provider
or
just
specific
to
your
cloud
provider,
issues
that
you
think
are
like
easy
to
medium
that
you
want
to
put
on
the
agenda.
Let
me
agree
and
then
we
can
kind
of
assign
those
to
any
new
contributors
that
are
looking
to
make
contributions
so
that
one
we
have
FYI.
H
H
The
code
is
complicated
on
the
API
machinery,
side
and
I.
Think
the
the
a
lot
of
the
goals
and
how
we
do
things
and
what
we
need
to
do
on
cloud
provider
is
going
to
be
similarly
complicated
and
rather
than
having
people
get
lost
or
not
understand
where
certain
things
belong.
It
would
make
a
lot
of
sense
if
we
get
someone
who's
genuinely
interested
to
sort
of
assign
them
a
mentor.
K
I
totally
agree
with
that,
like
I
have
certainly
been
on
the
reviewing
side
of
PRS.
That
did
not
let
you
know
that
like
had,
we
worked
on
them
together
would
have
saved
a
lot
of
time,
so,
like
I,
am
very
much
of
the
like
work
together,
upfront
and
and
get
everyone
aligned
and
knowing
the
code
base,
so
that
then
the
you
know
this
sort
of
iterate
and
to
discover
incrementally.
So
it
is
actually
time-saving.
So.
C
B
Definitely
reach
out
to
Paris,
with
her
experience
with
running
kind
of
our
cohort
mentoring,
especially
for
setting
that
baseline,
like
this
is
the
code
base.
This
is
how
we're
going
to
try
and
review
your
pull
requests.
You
can
get
a
lot
of
I
guess,
efficiency
by
being
able
to
mentor
with
one
with
one
mentor
to
you,
know
an
mentees
to
for
that
kind
of
baseline
stuff.
A
A
So
next
one
is
also
mine,
which
is
proposing
the
cloud
provider
expression
working
group.
So
this
is
the
work
that
Walter
kind
of
started,
with
putting
all
the
provider
code
into
staging
and
then
dealing
with
the
dependencies
that
come
with
that.
So
I
had
proposed
that
me.
She
and
Walter
kind
of
take
the
lead
on
that,
and
the
point
of
proposing
a
working
group
is
so
that
we
just
have
more
focus
and
we
have
like
a
dedicated
timeslot
and
agenda
and
everything.
A
So
we
can
track
this
work
and
also
because
it'll
be
kind
of
temporary,
because
once
we
have
the
providers
out
of
Cabrini
/
kubernetes,
we
wouldn't
need
the
group
anymore,
which
kind
of
goes
into
next
point:
yeah
yeah.
So
if
you
read
a
PR,
there
are
some
comments
about
whether
the
working
group
should
be
a
sub
projects
and
so
Stephen
kind
of
has
a
proposal
separate
from
this,
but
kind
of
kind
of
changes.
The
way
we
might
want
to
to
outline
this
so
Stephen
join
me.
Give
an
update,
I'm,
not
cure
yeah.
B
B
Release
team
release
engineering
things
like
that,
and
we
realized
there's
not
really
an
efficient
way
to
to
kind
of
present
each
of
the
mechanisms
to
contact
people
or
understand
all
the
repos
that
that
you
can
reach
out
to
us
on
or
that
we
don't.
So
this
PR
is
it's
more
cosmetic
than
anything
that
kind
of
restructures
some
of
the
generator
code
that
allows
you
to
that
allows
you
to
define
sub
projects
as
if
they're,
either
Federation's
of
different,
repos
or
or
actual
I.
B
Don't
know,
committees
or
kind
of
actual
sub
projects
that
that
have
chairs
and
are
dedicated
on
some
specific
focus.
So
I
think
that
sig
cloud
provider
is
I'm,
not
sure,
because
I
haven't
been
following
all
the
meetings
or
anything
but
I
I
think
this
a
cloud
provider
is,
is
kind
of
like
ripe
for
this
use
case
as
well.
The
way
I
envision
say
cloud
providers
having
like
say
cloud
provider
proper
and
then
a
bunch
of
sub
six
that
own
all
of
the
bits
for
the
relevant
cloud
providers
underneath
them
I.
B
Think
right
now
the
I'm
not
sure
whether
the
cloud
provider,
AWS
and
and
the
ones
that
are
already
within
cyclic
writer
are
referencing
the
migration
from
from
entry
two
out
of
tree
stuff.
Or
is
it
saying
that
this
is
code
that
this
sake
or
this
former
state-owned
so
I
think
that
delineation
would
be
useful.
C
So
yeah,
maybe
just
a
little
context
on
this.
This
proposal
kind
of
broke.
My
brain
and
I
was
stunned
a
little
bit
into
silence
because
this
cig
started
as
sig
or
working
cloud
provider
working
group
with
exactly
that
string.
And
then,
when
the
scope
expanded,
it
migrated
from
being
a
working
group
to
being
a
sig
partially
specifically
because
there
would
be
ownership
of
some
code.
C
So
from
from
my
perspective,
this
seems
like
it
is,
may
be
useful
to
break
out
into
a
sub
project
so
that
folks,
who
are
interested
in
cloud
provider
related
topics
that
are
not
specific
to
extracting
the
cloud
providers
might
get
benefit
out
of
the
meeting.
But
it
seems
more
like
meeting
time
organization
and
agenda
jockeying
and
a
separate
working
group.
Yeah.
G
B
B
H
B
So
I
think
to
to
the
subject
project
versus
working
group
point.
We
still
have
an
issue,
maybe
not
us
ourselves,
but
there
seems
to
be
some
confusion
about
what
the
scope
of
a
sub
project
is
versus.
What
the
scope
does
a
little
working
group
is,
and
there
are
some
active
PRS
from
like
steering
and
stuff,
but
the
one
of
I
think
one
of
Tim's
and
cleaner
Clare's
questions
were.
Is
you
know?
B
Don't
necessarily
find
that
to
be
true
and
and
Tim
doesn't
feel
that
way
either
so
I
think
there's
a
confusion
around
like
from
the
governance
level
like
what
is
a
sub
project
and
I
think
that
a
sub
project
can
be
whatever
sake
wants
it
to
be
as
long
as
it's
properly
organized
so
I
think,
there's
I
think
there's
flexibility,
for
if
this
is
a
federation
of
some
some
some
sets
of
code
versus
like
these
are
committees
that
live
within
the
state.
To
do
some
action.
B
G
It's
actually
a
sub
project
that
owns
code
actually
has
some
people
that
are
officially
reference
as
owners
of
those
dead
code
I
mean
owners
file,
that's
what
they're
being
used
for.
So,
if
the
this
new
sub
project
or
what
are
we
going
to
call
it,
it's
going
to
own
code,
we're
going
to
have
an
owner
file
that
owner
file
will
have
people
in
it.
So
they're
still,
you
know
the
accountability,
part
of
things
they
can
bear
up.
Even
though
there's
no
like
clear
upon
you,
know
clear,
appointed
leaders
or
chairs
in
that
specific
sub
project.
B
H
Well,
sub
project,
or
whatever
we're
calling
it
should
actually
get
turned
down
once
all
the
cloud
providers
are
gone,
not
having
been
said,
I
think
there
are
going
to
be
two
models
of
code
ownership.
One
is
for
cloud
provider,
specific
code
and
one
you
know
AWS
GCE
and
one
is
for
code,
which
is
general
across
cloud
providers
for
things
like
rap
controller
and
the
cloud
node
controller.
Things
like
that.
C
So
I
think.
If
we
organize
the
agenda,
this
meeting
typically
has
an
hour
and
we
don't
use
the
whole
hour.
So
even
as
a
it
sounds
like
there's
already
at
a
meeting
time
for
the
extraction
initiative,
which
is
great.
Keep
that
if
it's
useful
I
would
say
please
reach
out
and
make
sure
that
there
are
not
folks
in
Europe
or
Asia
who
are
missing
it
because
of
the
time
but
other
than
that.
A
Yeah
and
I'll
be
honest,
like
I
literally,
like
the
only
reason
I
chose
what
in
group,
because
it
lets
me
like
set
a
time
and
a
like
a
China
origin
so
like
that
works
with
stuff
projects,
s
I
like
what
you
said,
though,
about
just
allocating,
maybe
like
the
last
15
minutes
or
the
whatever
remaining
time
you
have
in
this
meeting
to
talk
about
the
extraction
instead
of
a
separate
time
slot
Walter.
Do
you
think
that
might
be
a
good
idea
going
forward.
H
B
A
A
H
I
think
I
mean
whether
it's
a
working
group
or
a
sub
sub
sig
or
whatever
we're
calling
it
I
think.
While
this
initiative
is
going
on
having
the
place
where
sig
release
sigdoc
sig
no
sig
lifecycle
can
come,
you
know
anyone
who's.
An
AOA
WS
customer
with
concerns
can
come
and
chat
about
what
we're
doing.
Where
we're
doing
it.
Why
we're
doing
it?
How
we're
doing
it,
how
they
can
help
is
really
valuable
and
so
I
think
having
something
beyond
a
meeting
has
some
value,
whether
it's
a
working
group
or
it's.
H
A
A
B
A
J
H
Of
this
is
detailed
in
the
cap,
but
I
think
I,
don't
think
we've
actually
organized
it
in
that
way.
So
we
mentioned
things
that
need
to
be
owned
and
need
to
be
handled
and
need
to
be
moved.
But
if
you
think
there's
value
in
actually
just
saying
we're
planning
on
taking
over
these
pieces
and
that
mid,
there
may
be
some
value
in
that
I
think
we
can
certainly
enhance
the
cap
if
we
think
detailing
the
data
that,
in
that
straightforward
away,
is
valuable.
B
Yeah
cuz
I
got
a
I,
got
a
question
today,
Sekou
sure
about
like
the
CSI
plugins,
Roger
and
I
was
like
I.
Don't
know,
I,
don't
like
it's
lives
in
someone's
repo
right
now,
and
we
want
to
migrate
into
kubernetes.
Sakes
is
like
cool
I.
Would
love
for
you
to
do
that,
but
I
don't
know
what
the
list
of
all
the
things
that
I
need
to
I
need
to
kind
of
catbird
or
at
first
so
I
think
it
would
be
valuable
when
the.
H
Csi
one
is
is
a
special
interesting
case
because,
to
date,
the
volume
none
of
the
volume
logic
has
ever
actually
lived
in
the
cloud
provider
interface.
So
storage
has
always
done
that
on
the
side
and
I
believe
the
CSI.
The
basic
CSI
framework
is
in
one
repo
outside
of
kubernetes
and
then
at
least
the
GCE
PD
CSI
plug-in
is
yet
another
repo
that
right
now
sod
stands
from
storage.
G
H
G
We
actually
had
a
question
the
other
day
internally,
because
CSI
is
not
supposed
to
be
covering
any
specific,
there's
more
like
a
standard,
so
people
are
gonna,
you
know
I
would
say
they
were.
You
know
looking
at
the
fact
that
most
people
want
to
use
a
you
know
a
couple
kubernetes
club
provider
XYZ
to
store
CSI
blogging
that
are
supposed
not
to
be
used
only
with
their
identities,
so
maybe
having
a
separate
vehicle
for
that
made
sense.
But
I
guess
that's
up
for
up
to
interpretations
any
more.
A
Yeah
I
think
we
might
be
good
to
have.
Maybe
someone
in
this
take
lead
on
this
storage.
Stuff.
I
know
Walter
you.
You
have
a
lot
of
information
on
storage,
but
you
seem
pretty
busy
with
all
the
actual
extraction
itself
as
well,
so
might
be
good
to
have
someone
to
say
kind
of
follow
up
on
that
week
week.
H
A
F
H
So
I
don't
know
who's
going
to
China,
suddenly
I'm,
not
but
I
I
think
it
would
be
good
if
we
try
to
make
some
time
for
all
of
us
to
meet
up
in
each
who
anyone
who's
actually
going
to
be
cute
guns,
so
I'll
definitely
go
to
Seattle.
So
if
we
can
try
and
meet
up
at
some
point,
maybe
during
the
contributors
summit
yeah.
J
C
Andrew
said
we
have
both
an
intro
which
I
wouldn't
expect
to
see
many
of
you
at
and
a
deep
dive
session,
which
I
hope
to
see
many
of
you
at
in
both
Shanghai
and
Seattle,
and
the
deep
dive
description
that
we
came
up
with
was
more
about
what
the
intention
of
the
build
and
release
process
is
so
I'm
hopeful
to
reach
out
to
some
folks
from
cig
release
as
well
to
expand
the
scope
beyond
the
process
and
into
the
build
and
release.
What
are
the
binaries?
C
What
happens
in
a
world
where
the
cloud
provider
artifacts
are
not
together
with
the
kubernetes
build
itself,
and
what
does
discoverability
look
like?
What
is
combination
of
those
things?
Look
like
what
is
validation
and
versioning,
and
so
to
be
an
interesting
conversation,
somewhat
dependent
on
who
shows
up
and
what
their
current
pain
points
are.
But
please
do
stop
in
for
that
one.
So,.
J
B
B
G
Just
want
to
throw
it
out
there,
given
them
I
guess
many
of
us
will
be
a
definitely
cute
girl.
Maybe
will
be
a
good.
You
know
idea
they
start
thinking
about.
If
you
want
to
do
it
like
a
face-to-face
meeting
like
a
proper
one,
you
know
it's
gonna.
If
I'm
not
mistaken,
you
can
shadow
will
end
up
being
right
after
code
freeze
or
right
before
good
freighter,
113,
which
means
there
will
be
a
good
point
to
start
talking
about
114,
which
may
be
a
you
know,
potentially
the
reason
what
we're
gonna
have
end
up.