►
From YouTube: SIG Cloud Provider 2019-10-30
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,
hello,
everyone
today
is
October
30th
2019.
This
is
the
bi-weekly
cloud
provider,
say
cloud
provider
meeting
so,
as
usual,
we'll
start
with
sub-project
updates
and
then
we'll
go
into
the
agenda.
I,
don't
see
anyone
from
Alibaba
here
so
skip
that
provider
AWS
any
updates
worth
calling
out
you
guys
for
me
I'm
having
issues
with
my
yeah,
okay,
cool
I.
Don't
think
we
have
too
much
we're
working
on
ete
tests.
A
So
hopefully,
next
couple
weeks,
we'll
we'll
have
a
first
implementation
at
and
then
J
do
you
have
anything
else,
yeah
I'm,
not
yeah,
I,
think
that's!
It
I
think
we
we
need
to
triage.
Some
of
those
issues
that
you've
been
pointing
us
to
I
saw
that
one
today.
So
that's
also
on
our
plate
and
we're
we're
cool
and
that's
IDI
test
for
the
the
true
provider
or
for
both
I
mean
well
it'll,
technically
be
for
both
because
out
of
tree
and
poor
Slyke
a
seat
but
yeah
it's
out
of
trick.
A
Yeah,
okay,
cool,
so
I
do
want
to
call
out
one
thing
here,
so
we
do
have
a
cloud
test
suite.
It
only
has
one
test
but
I,
think
if
possible,
it'd
be
good.
If
we
can
move
those,
if
you
can
add
those
tests
to
that
suite
so
that
other
providers
can
can
run
it
yeah
as
far
as
they
know
like
we
have
the
cloud
package
in
the
test.
A
The
new
testing
framework,
and
the
only
thing
you
test
right
now
is
deleting
ec2
instance
and
then
ensuring
that
the
control
plane
deletes
the
associated
notes,
that
instance,
but
yeah
like
if
there
are
more
common
ones
like
add
them
here,
so
other
folks
can
benefit
from
it
got
it.
Okay,
yeah
right
now
is
just
kind
of
aw
specific
stuff,
but
yeah.
If
there's
anything
comment
other
thing
I
did
there,
we
were
I,
was
sort
of
looking
at
your
I.
A
Think
there's
a
été
testing
cup,
but
it
I
think
it's
mostly
just
saying
like
you,
should
have
at
least
180
tests
and
publish
the
results
yeah.
That
was
more
for,
like
the
publishing
side
of
it,
like
it'd,
be
great
if
we
want,
as
many
providers
like
running
conformance
and
then
uploading
it
to
test
grid.
A
A
B
B
A
A
D
A
All
right
so
yeah
also
saying
earlier,
we
ran,
we
were
running
into
issues
on
vSphere
where
I.
Well,
it's
not
really
an
issue,
but
we
have
an
assumption
and
vSphere
that
you
have
one
Nick
for
VM
and
then
you
would
assign
you
know
that
one
Nick
to
internal
or
external
IP
in
the
careering
snowed,
and
so
this
is
pretty
much.
This
is
a
common
pattern
for
most
cloud
providers.
Like
you,
you
have
one
IP
for
internal
IP,
one
and
one
IP
for
external
IP,
but
then
going
forward
we're.
A
A
A
When
Nick
I
mean
the
we
have
multiple
en
eyes
on
our
note,
but
they're
allocated
to
pods
right
so
like
when
the
node
is
there
a
scenario
scenario
where
the
node
tries
to
look
up
its
private
ipv4
address
and
it
accidentally
looks
up
the
you
know:
a
dress
instead
of
the
actual
node
address,
so
I
think
it
knows
how
to
determine
that,
because
the
node
is
always
8
0,
but
we
do
I.
You
know
like
we
do
have
a
bunch
of
other
interfaces
on
there.
A
So
if
there
was
code
that
I
don't
know
had
assumptions
about
that,
I
just
want
to
make
sure
that
that's
an
acceptable
use
case
right,
okay,
cool
good
to
know
all
right,
so
extraction
migration.
One
thing
that
we're
trying
to
do
often
is
try
to
get
kind
of
a
status
roll-up
from
all
the
providers
on
where
they're
at
for
the
extraction,
migration
efforts.
A
So
just
kind
of
giving
a
quick
summary
of
what
this
is
tracking.
This.
The
top
here
is
entry
providers,
which
ideally
nothing
really
changes
here,
because
we're
kind
of
removed
that
but
we're
trying
to
get
more
updates
on
the
other
tree
work
and
ideally
try
to
get
this
as
close
to
what's
in
tree
or
ideally
better.
A
A
A
A
E
A
My
understanding
is
that
this
is
a
controller,
that's
specific
to
Google
yeah,
that's
correct!
The
cube.
The
controller
manager
has
a
special
iPad
mode
that
is
just
for
Google.
We
had
I
had
put
this
read
because
I
didn't
know
what
the
future
planned
for
for
a
do.
This
AWS
is
but
I'm
pretty
sure.
Like
most
cloud
providers
are
gonna.
Have
this
as
unsupported
is.
A
A
A
If
you
were
to
ask
Walter
he'd,
probably
say
you
you
may
want
to
support
it
in
the
future,
because
largely
what
it
allows
you
to
do
is
do
pot
allocation
on
a
per
pod
basis
and
not
like
allocating
a
subnet
per
node
which
allows
for
more
efficient
allocation
of
pot
subnets
because
you're
doing
it
and
based
on
the
number
of
pods
and
not
the
number
of
subnets,
you
can
allocate
per
node
right.
So
our
defaults
C&I
does
not
do
a
subnet
per
node,
but
I
don't
know
what
the
difference
I
I
guess.
E
A
E
I
guess
for
this
specific
thing,
which
is
you
know,
the
migration
of
cloud
providers
outside
of
KK
is
this?
Is
the
cloud
node
I
Pam,
something
that
we
need
to
work
on
something
in
in
the
AWS
cloud
provider
and
tree
to
make
it
extractable,
or
is
that
already
done?
I
guess
I'm?
That's
what
I'm
missing
yeah.
A
D
A
That
for
now
and
then
we
can
follow
up
with
Walter
and
see
if
we
should
sounds
good
change,
then
cool
all
right
so
CSI
is
being
worked
on,
is
Amazon
working
on
a
proxy,
the
API
server
Network
proxy,
not
right
now,
okay,
but
you
plan
to
support
it.
That's
also
TBD,
okay,
so
yeah
we
are
we've.
We
have
interest,
but
we're
not
100%,
sure,
okay,
cool,
okay
and
test
crude
and
potential
provider,
both
in
progress
right.
A
C
C
A
B
D
D
Yeah
I
guess
I
need
to
look
yeah
because
we
had
continual
conformance
testing
running
though
I
haven't
looked
at
it
in
a
while
for
the
cloud
provider.
But
there
is,
you
know
all
of
it.
The
tests
that
run
on
every
PR
are
there
on
test
grid
as
well,
but
I
should
double-check
on
conformance,
probably
before
we
put
that
green
okay.
A
Cool
this,
so
it
seems
the
new
work
trying
to
get
the
decision
package
moved
to
the
cloud
provider
repo
so
that
when
you
try
to
vendor
in
all
the
CCM
stuff,
you
don't
have
to
actually
vendor
in
the
main
community
street
piece
on
the
call
so
we'll
skip
this
update.
But
he's
been
good
progress
on
this
one
credential
provider
who
this
year.
A
A
E
A
Mean
like
I,
am
happy
to
take
the
work
and
I
know.
We
discussed
that
but
I
like
you,
had
a
talk
with
him
at
some
point:
Andrew
and
and
I
did
he
say
that
he
was
going
to
or
yeah
so
I
have
not
been
able
to
get
in
touch
with
them.
I
think
it's
I,
think
it's
okay
to
say
at
this
point
that
you
should
take
the
work
because
it's
been
yeah.
It's
been
about
like
three
weeks
since
we
got
a
response,
so
I
think
yeah.
A
A
So
let's
go
to
the
gender
there's
just
one
item
here:
I
just
wanted
a
quick
update
on
if
we're
all
good
to
go
for
the
cube
con
sessions
in
San,
Diego,
so
I
know,
Nick
and
Fabio
are
set
up
for
the
intro
and
then
you
know
and
Walter
are
set
up
for
the
deep
dive
yeah.
So
me
and
Fabio
have
been
we've
had
it.
We've
met
a
couple
times.
We
are
working
on
the
slides
right
now,
so
will
will
be
meeting
again.
F
A
A
G
Let
you
know
I've
got
a
meeting
scheduled
for
next
Monday
before
the
steering
committee,
so
I
thought
there
were
a
number
of
indications
that
indicated
that
this
would
be.
You
know,
kind
of
a
straightforward
process
but
I
believe
I'm,
the
first
one
going
through
this.
So
I'm
learning
some
of
these
things
as
I
go
along
and
it
does
have
a
web
portal,
but
I
don't
have
a
slack
channel
or
a
zoom
meeting
account
or
anything.
So
we
haven't
actually
started
and
right
now,
I've
got
a
proposed
charter
in
a
PR.
G
G
These
technically
are
pr's,
but
they're
I,
don't
know
what
you'd
call
them
tickets
or
something
to
open
up
a
request
for
a
slack
channel
and
a
zoom
and
those
aren't
completed
yet
either
and
the
people
doing
that
told
me
that
they're
gonna
wait
until
the
steering
committee
acts
on
the
Charter.
So
all
right,
what
I
can
do
I
think
afterward
is
maybe
even
a
write
up
on
a
summary
of
the
prop.
The
whole
process
can
Walter
yeah.
A
F
So
quick
questions
I
mean
so.
Thank
you.
So
much
if
you
do
write
that
up,
I
think
that'll
be
hugely
useful
for
a
lot
of
us.
I
know
that
we
already
have
I
thought
some
user
groups
sort
of
for
several
the
cloud
providers
under
this
seg
and
I
do
know.
We
definitely
have
several
cloud
providers,
slack
channels
that
are
owned
by
this
SEC
yeah.
G
A
G
F
Not
saying
there
isn't
utility,
but
my
memory
was
and
I'm
happy
to
have
someone
like
Nick
correct
me,
but
the
various
cloud
provider
channels
seem
to
be
have
slightly
different
emphasis
for
different
cloud
providers.
So
my
memory
was
like
the
AWS
slack
Channel
and
group
under
this
seg
tend
to
be
more
end-user,
and
some
of
the
other
ones
were
definitely
more
developer.
Focused
no.
A
It
definitely
is
and
I
thought
that
we
so
I
guess
I
didn't
really
understand
what
the
point
of
migrating
to
user
groups
was
and
I
thought
that
was
kind
of
I
thought.
User
groups
were
replacing
sub-projects
in
like
for
cloud
providers
under
sake
cloud
provider
like
you
optionally,
if
you
wanted
the
slack
Channel,
if
you
wanted
that
stuff,
then
you
would
become
a
user
group,
but
it
sounds
like
there's
actually
two
different
options.
You
can
either
be
a
sub-project
or
you
can
be
a
user
group
yeah.
G
G
F
A
G
It
might
be
unique
to
our
scenario,
because
with
film
where
we
have
multiple
infrastructures
for
running
kubernetes,
for
example,
people
could
run
it
on
our
desktop
hypervisors,
which
are
workstation
infusion,
and
they
really
would
have
nothing
to
do
with
the
vSphere
cloud
provider
at
all.
So
having
a
user
group
to
capture
that
I
think
makes
sense,
but
if
this
might
not
apply
to
other
cloud
providers.
Well,.
G
And
by
the
way,
I
just
I
posted
a
link
to
that
PR
on
the
Charter
in
the
chat
here
and
then
I.
Think.
If
you
look
in
that
PR,
it
might
have
its
there's
there.
Certainly
the
proposed
Charter
is
one
of
the
files
being
checked
in,
but
I
think
it
may
have
left
comments
pointing
to
some
of
the
prior
steps,
but
when
this
is
all
over,
I'm
I'd
be
happy
to
write
this
up.
Just
in
case
somebody
follows
my
footsteps:
yeah.
E
A
Context
is
that
most
so,
like
the
cloud
providers
are
a
special
case
because
they
were
cigs
and
they
already
have
established
end-users
attending,
and
so
the
transition
to
use
group
may
not
make
sense,
but
for
most
user
groups,
they're
kind
of
starting
off
to
just
create
a
form
for
users
to
discuss
certain
topics
without
actually
having
to
be
sponsored
or
owned
by
sig
as
a
project,
and
so
like
big
data
is
the
classic
example
of
there's
no
sig
and
there's
no
code
to
actually
own.
But
there
are
a
lot
of
users
who
are
running.
A
You
know
big
data
platforms
on
kubernetes
I
just
want
to
talk
about
it,
and
so,
like
I,
agree
like
if
it
doesn't
exempt
sense
for
AWS,
because
the
sub-project
forums
are
fine.
Then
then
yeah
no
need
to
create
the
user
group.
But
if
you
do
need
one
down
the
road
you're,
always
you
always
have
the
option.
Yeah.
G
And
I
believe
that
the
big
data
group
actually
predated
the
effort
to
create
these
user
groups
that
are
cloud
focused,
they've
been
out
there
for
a
while,
and
one
could
argue
retro
actively
that
the
IOT
edge
working
group
by
man
probably
should
be
a
user
group,
because
it
really
is
a
group
of
disparate
users,
but
it
dates
back
over
a
year
before
they
had
any
concept
of
having
user
groups.
So
some
of
these
things
were
done
in
a
legacy
world.
That's
a
little
different
than
what's
being
shaped
today.
G
G
I've
got
two
listed
in
my
PR,
but
I
I
identified
people
who
are
willing
to
go
and
account
bigger
than
that,
so
that
that's
another
step
is
actually
lining
up.
Users
and
I
think
maybe
someday
I'd
wish
that
you
know
for
this.
First,
let's
call
it
an
auger
a
year
I
the
vendor
am
going
to
be
a
chair,
but
I'd
hope,
maybe
in
a
future
year
the
users
might
actually
take
this
over.