►
From YouTube: SIG Cloud Provider 2023-09-13
Description
Meeting agenda: https://docs.google.com/document/d/1OZE-ub-v6B8y-GuaWejL-vU_f9jsjBbrim4LtTfxssw/edit#bookmark=id.rhsekr1jd8gl
A
Welcome
to
kubernetes
Sig
Club
provider
for
September
13th
2023.
We
are
going
to
kick
off
the
the
meeting
here
with
a
reminder
that
we're
governed
by
the
kubernetes
code
of
conduct.
So
in
short,
please
be
excellent
to
one
another.
Be
kind:
have
fun
talk
about
Cloud
providers,
it's
great!
It's
probably
fine!
A
Okay,
jumping
into
the
agenda!
The
first
thing
I
wanted
to
look
at
while
people
fill
out
their
sub-project
updates.
Is
we
actually
have
a
couple
of
issues
to
triage?
So,
let's
that
have
said
cloud
provider
listed
on
them
and
needs
triage.
So
let's
take
a
look
at
this
one
first
and
see:
whoa.
Okay,
sorry
I
just
moved
to
the
wrong
I.
Just
don't
know
what
happened.
A
I'm,
sorry,
that
was
super
weird
I
tried
to
drag
the
zoom
controls
that
were
obscuring
my
view
of
GitHub
out
of
the
way,
and
it
took
me
to
another
desktop
I.
Don't
know
if
that
made
a
difference
to
what
was
displaying
or
not
sorry
about
that.
Okay,
so
service
node
Port
cannot
be
released
promptly.
Let's
see
if
this
has
had
some
work
on
it.
It
looks
like
Antonio
says
this
is
a
Sig
Network
problem.
A
A
So
what
do
we
think
Cloud
controller
manager
is
adding
a
file
license.
A
Or
at
least
that's
how
we
test
it
all
right.
Does
anyone
see
anything
in
here
that
they
want
to
follow
up
on
or
investigate.
C
B
A
Okay,
I
mean
it,
it
does
sound
semi-familiar,
but
again,
it's
kind
of
unsure.
Basically,
it
looks
like
network
is
definitely
Antonio's
going
to
be
taking
a
look
at
this.
Is
this
something
where
we
want
to
because
he'd
act,
the
problem
he
has
not
removed
the
knees
triage,
so
I'm
kind
of
reluctant
to
have
us
remove
these
triage.
Since
then,
I
won't
show
up
in
TIG
networks
triaging
next
time
they
run
through,
but
does
do
you
makeover?
B
A
A
Okay,
yes,
it's
of
course,
provider
specific
tasks.
Okay,
what
do
we
think
it
looks
like
some
test
runs
have
been
green.
Do
any
of
our
Google
friends
want
to
take
a
look
at
this,
our
chairs
accepted.
If
we
have
anybody
yeah.
A
All
right,
so
those
are
the
only
two
that
are
the
current
ones.
Hopefully,
anyone
else
who
wanted
I
don't
have
anything
in
sub-project
update
today,
specifically
myself,
it
looks
like
IBM
Cloud
we
do
have
one.
Do
you
want
to
tell
us
what's
going
on
with
your
a
project
yep.
E
The
open
source
repo
we
have
is
being
impacted
right
now,
just
to
we
have
two
different
environments.
We
have
in
IBM
Cloud.
We
have
a
classic
environment
and
a
VPC
environment,
and
right
now
both
of
them
are
open
sourced
and
eventually
we
want
to
deprecate
the
classic
environments.
We
want
only
open
source,
the
VPC,
so
we're
working
on
refactoring.
So
we
can
set
that
up.
So
we
only
open
sourcing,
one
environment
at
the
other,
so
just
to
refactoring
the
code.
A
Yeah
and
you
want
to
have
something
consistent,
yeah
all
right,
well,
more
power
to
you
on
that
and.
F
A
You
the
best
there
that
is
always
super
fun
and
then
does
anybody
else
have
any
other
sub
project
updates
that
you
want
to
go
into
before
we
jump
into
our
agenda.
A
Okay,
Andrew,
it
looks
like
you
have
our
first
topic.
You
want
to
take
it
away.
D
Yeah
so
we
discussed
it
in
the
extraction
so
project
meeting
and
it
was
discussed
on
slack
as
well,
but
dims
kind
of
flip.
This
or
actually
did
we
discuss
it
in
the
last
meeting
because
I
missed
the
last
meeting.
A
D
S
yeah
so
I
think
we
should
with
so
assuming
that
we're
going
to
keep
those
changes
for
129..
We
probably
want
to
put
a
lot
of
thought
and
effort
into
documentation,
blog
posts.
Things
like
that,
making
sure
that
it's
a
pretty
big
highlight
of
the
release
and
ensuring
folks
know
what
things
are
going
to
break.
A
D
So
my
understanding
is
beta.
Apis
are
not
on
by
default
in
the
runtime
config,
but
beta
feature
gates
are
still
on
by
default,
and
so,
with
the
promotion
of
this
feature,
gate
to
Beta.
We
are
turning
it
off
by
default
and
because
there's
no
runtime
API
associated
with
it
it'll
just
be
on
by
default,
which
is
means
it's
off
by
default.
D
I
agree:
okay,
I,
don't
know
we
probably
want
to
just
sync
up
with
the
release
team
as
well
and
just
make
sure
they
understand
the
magnitude
of
the
change
and.
F
B
Yeah,
so
I've
got
the
pr
up
there
for
I
just
linked
to
PR.
That's
updating
the
metadata
but
I'd
like
to
follow
up
on
that
with
the
there's,
a
PR
that
dims
or
there's
an
issue
that
dims
had
started
where
he
was
collecting
notes
about
like
the
migration
stuff,
because
I
think
that's
the
other
part
of
the
cap
we
need
to
update.
Is
we
need
to
put
some
some
general
guidance
about
what
it
you
know?
What
components
do
you
need
to
update
and
or
to
migrate?
B
And
how
do
you
migrate
them
and
then
we
it
sounded
like
we
can
defer
some
of
the
platform
specific
details
to
those
specific
platforms,
but
at
the
very
least
we
need
to
talk
about
like
what
do
you
need
to
do
to
kubl
it?
What
do
you
need
to
do
to
API
server?
What
do
you
need
to
do
the
cube
controller
manager
and
then
and
then
talk
about
you
know?
You
also
have
to
deploy
a
CCM,
but
that's
kind
of
in
their
hands
at
that
point.
B
So
I'd
like
to
get
back
in
and
update
that
part
of
the
cap
as
well,
but
so
far,
I've
just
put
up
a
PR
that
kind
of
took
dim's
work
and
just
you
know
carried
it
forward
with
updating
the
versions
and
putting
us
into
beta
phase
then
sounds.
A
B
It
was
the
docs
updates
that
were
the
big
part,
so
like
I,
think
that
dims
and
the
second
link
that
I
put
there,
that
has
kind
of
like
the
issue
and
the
the
cap
I've
been
trying
to
put
all
the
links
in
that
issue
because
that's
like
where
the
discussion
is
kind
of
happening.
So
that's
where
that's
where
I
said
I'd
like
to
take
dim's
work,
Gathering
some
of
these
notes
and
then
transferring
them
into
the
cap.
So
that's
probably
the
next
thing.
I
need
to
do
basically.
A
Okay,
I'm
putting
a
notice.
You
can
see
that
we
need
to
look
at
the
deadlines
for
1.29
and
make
sure
I
know
the
release
like
cycle
just
started,
but
I
think
we
gotta
look
at
the
deadlines
very
carefully
and
make
sure
we're
going
to
make
the
production
Readiness
review
and
documentation
deadlines
because
I
don't
want
to
set
ourselves
up
for.
Oh
no,
we
just
realized.
We
have
to
do
it
all
in
the
next
three
days
or
else
like.
Let's
make
sure
we
know,
let's
make
sure
we
know
yeah.
B
Yeah,
that's
why
we
wanted
to
turn
the
tests
on
early
in
the
in
the
cycle
so
that
we
know
we
kind
of
get.
It
start
to
get
an
impression
of
what's
happening,
but
yeah
like
I'll,
make
a
note
to
double
check
on
the
dates
you
know
just
to
see
what
kind
of
time
I
have
to
get
this
stuff
done
on
the
cap
updates,
yeah.
D
Great
progress,
though
129
so
far,
the
fact
that
we
have
like
the
CI
pipelines
green
with
this
with
the
feature
kit
already
flipped,
is
really
good,
but
yeah
Bridget,
I
kind
of
agree
that
we,
we
probably
pulled
the
trigger
a
little
early
doing
the
feature.
Gate
update
before
the
camp
update,
but
I
could
take
that
off
with
dims,
because
he
was
kind
of
pushing
yeah.
A
I'm
happy
I'm
happy
that
we
went
that
direction
because
I
remember
dim's,
being
the
one
kind
of
shutting
us
down
in
the
past,
so
like
Jim's
being
enthusiastic
about
moving.
This
forward
sounds
great:
let's,
let's
keep
with
that
momentum,
but
we've
got
to
make
sure
we
check
out
all
the
boxes
so
that
this
doesn't
get
booted
out
of
the
release.
B
A
D
Sure
so,
there's
still
like
one
admission
controller,
an
API
server
that
is
applying
some
metadata
on
Precision
volumes
if
it's
used
by
one
of
the
entry
Cloud
providers
and
our
current
our
plan
in
previous
releases
was
to
add
a
web
hook.
Integration
in
the
CCM
that
can
replace
that
logic,
but
I
think
a
lot
of
that
work,
kind
of
stalled
and
the
feature
to
add
the
web
for
hosting
in
CCM
I'm
kind
of
stuck
in
Alpha,
not
really
ready
for
beta.
D
So
what
I'm
proposing
instead
here
is
just
create
an
external
project
that
is
specifically
just
a
simple
admission
web
hook
that
replaces
the
persistent
volume
label
admission,
controller
and
point
users
in
that
direction.
If
they
still
need
that
volume
that
emission
controller.
D
What
the
answer
was:
yes,
yeah,
so
it's
a
the
project's
already
created
the
repos
and
develop
it's
a
it's
a
kubernetes
six
project
and
basically
it's
just
going
to
be
a
Web
book
server
that
Imports
the
Legacy,
the
Legacy
cloud
provider
packages
and
just
it
it's
basically
just
a
wrapper
around
the
current
logic.
But
but
just
as
an
external
admission
controller.
A
B
A
C
I
just
wanted
to
remember
so,
like
I
started
like
working
on
the
next
steps
for
using
it
as
a
proper,
a
Web
book,
because
right
now,
whatever
Nick
has
done,
the
changes
are
only
to
start
as
a
admission
Web
book
server
but
the
actual
con.
It
doesn't
update
the
episode
with
the
new
configurations
to
make
the
webbook
in
place
so
I
started.
C
Work
on
that,
like
I'm,
hoping
to
I
mean
have
some
initial
PR,
like
maybe
this
week
or
early
next
week,
sometimes
I
hope
to
have
some
update
before
the
next.
This.
D
Awesome
so
just
to
be
clear,
there
there's
two
separate
things
we're
talking
about
like
the
first
one
is
the
original
plan
which
Kirsten
just
mentioned,
which
is
the
ability
for
the
cloud
controller
manager
to
host
a
webhook
server,
which
you
can
use
as
kind
of
a
generic
like
plug-in
mechanism
to
do
admission
control
for,
like
you
know,
random
resources
like
Services
service,
load,
balancers
and
persistent
volumes.
D
Things
like
that,
but
because
we
are
moving
forward
with
this
promotion
to
Beta
in
129
and
the
and
that
web
hook,
integration
is
not
going
to
be
ready
for
129.
We
are
working
on
a
separate
effort
to
have
an
ex
to
have
a
project
that
users
can
use
right
away.
That
replaces
the
one,
the
the
very
specific
capability
that
the
web
hosting
on
CCM
was
supposed
to
fix,
which
is
persistent
volume
labeling.
D
So,
like
we're
gonna,
the
current
plan
is
to
just
do
them
in
parallel,
like
we're
going
to
have
the
external
project
that
replaces
just
the
persistent
volume
labeler
as
an
external
web
hook,
and
then
we're
also
going
to
continue
working
on
the
generic
web
hook
thing
that
we
can
eventually
use
for
other
things
as
well.
B
Just
had
a
I
have
a
I
just
have
a
question
I'm
wondering
like
do
we
need?
Do
you
think
we
need
to
have
a
discussion
of
this
of
the
new
controller,
the
new
admission
controller
you're,
creating
in
the
cap
like,
should
we
also
add
that
information
to
the
cap,
okay,
cool?
Thank
you.
A
I'm
I'm
nodding
like
yes
and
I,
also
feel
like
I'm
a
little
I
want
to
make
sure
I
understand
and
have
the
timelines
correct.
Because,
with
my
naive
understanding
of
this,
it
kind
of
sounds
like
what
we
would
be
saying
to
the
world
is
hey.
Our
cup
is
looking
great.
We
are
promoting
it
to
we're
promoting
this.
You
know
change
to
Beta
And.
Also,
we
made
a
completely
major
load-bearing
change.
A
A
D
We
can
provide
a
simple
external,
like
just
admission
web
hook
that
basically
Imports
the
same
packages
as
what's
in
tree
and
does
the
exact
same
logic
as
the
API
server,
but
it
just
does
it
as
an
external
webbook
and
not
as
an
entry
admission
controller,
and
that
would
that
would
be
the
last
thing
that
we
need
to
replace
that's
entry
and
from
there
we
should
effectively
have
like
complete
feature
parity
within
tree,
we're
just
not
using
the
built-in
web
capabilities
in
the
CCM
to
do
that,
because
that
feature
is
not
ready.
D
A
I
mean
the
functionality
all
sounds
great.
The
thing
I'm,
the
I'm
slightly
concerned
about
is
the
part
where
we're
saying
something
can
go
to
Beta,
even
though
one
of
the
features
that
is
theoretically
essential
for
it
is
actually
going
to
be
an
external
Alpha
thing.
That's
the
thing:
I'm
going:
what's
production,
Readiness
review
going
to
say
about
that
because
I'm,
a
little
worried
they're
going
to
turn
around
look
at
that
and
go
until
the
essential
underpinning
components
can
be
marked
beta.
Do
you
really
think
this
can
be
beta?
D
Yeah
so
like
I,
guess,
I
guess
this
is
a
matter
of
like
what.
What
do
you
mark
As
Alpha
Beta,
because
the
like
this,
the
web
hook
capability
in
the
CCM,
is
Alpha
for
sure.
D
But
if
we,
if
we
create
an
external
project
which
is
just
a
webhook
server,
that
is
importing
the
exact
same
code,
AS,
what's
already
an
API
server
and
it's
not
new
code,
it's
just
like
porting
the
same
code,
but
as
an
external
binary
like
yeah
like
if
we're
saying
like
that's
Alpha,
then
okay,
maybe
we
need
to
wait
longer.
If
we're
saying
it's
it's,
we
could
mark
it
as
stable,
because
it's
just
the
same
code
in
a
different
binary,
then,
maybe
not
so
I
I
think
we
can.
B
I
think,
like
another
thing,
we
probably
it's
probably
important
to
mention
while
we're
talking
about
this
new
project
is
kind
of
the
scope
of
impact,
because
my
understanding
and
I'll
I'll
try
to
mention
this
in
the
cut.
But
my
understanding
is
that
the
scope
of
impact
is
relatively
limited
for
this
because,
like
most
most
new
users
to
kubernetes
are
going
to
be
using
CSI
drivers.
If
they're
using
storage
and
the
CSI
drivers
will
most
likely
handle
the
persistent
volume
labeling.
B
So
like
this
admission
plug-in
that
that
labels,
the
persistent
volumes,
is
only
really
needed
in
some
very
specific
cases
on
on
specific
Cloud
providers.
So
I
think
it's
also
important
to
note
that
when
we
talk
about
this
like
really,
this
is
to
help
people
who
are
bringing
older
clusters
forward,
who
are
not
necessarily
engaged
with
some
of
the
newer
storage
patterns.
B
So,
like
that's,
my
understanding
of
it
and
I'd
like
to
maybe
get
the
details
there,
so
that
when
we
talk
about
this
and
I
get
what
you're
saying
Bridget
like,
even
though
we're
taking
the
code
from
another
project
and
it's
still
a
whole
new
project.
We're
setting
up
a
new
way
of
configuring
and
everything,
but
we
need
to
let
people
know
that
really
this
a
minority
of
users
are
going
to
be
using
this,
because
it
is
kind
of
like
an
edge
case
for
modern
kubernetes
deployments
and
that's
just
kind
of
my
opinion
on
it.
A
A
Oh
I'm,
saying
I
just
I've
been
through
production,
Readiness
reviews
Gauntlet
a
few
times
so
I'm
just
making
sure
that
we
have
our
ducks
in
a
row
in
terms
of
what
we
assert
is
going
to
be
fine
and
what
we
where
we
believe
risks
exist
and
what
we
can
document
about
them,
because
I
think
that's
going
to
be
really
germane
to
whether
or
not
this
gets
the
green
light
from
the
prrr
committee.
Yeah.
B
B
You
know
on
slack,
if
I
run
into
some
trouble
kind
of
trying
to
figure
out
what
I
need
to
be
looking
at
here
to
get
these
details.
Yeah.
D
Sounds
good
and
we
might
have,
we
might
have
to
Lupine
Michelle
for
the
specifics
on
the
storage
bits.
A
This
is
very
exciting
and
I
think,
given
that
we
have
some
real
me
documentation
and
and
prr
work
to
do
here.
Let's
make
sure
that
before
our
next
meeting
we've
at
least
done
some
sort
of
sync
on
that.
B
A
Okay,
great
I
mean:
is
there
anything
else
about
this,
that
we
should
take
a
look
at?
Do
you
want
to
step
through
any
of
this,
or
what
do
you
think
Andrew.
D
No
I
think
I,
think
yeah.
Just
making
sure
like
this
is
I
agree
like
make
sure
it's
all
the
numerated
enumerated
in
the
cap,
so
I
think
yeah,
let's
just
yeah
when
the
Cap's
ready,
let's
just
review
the
cap
together
and
make
sure
it's
on
there.
A
Okay,
let's,
let's
just
write
down
review.
D
A
To
go,
look
I
mean
it's
there.
We
seem
to
be
much
sooner
and
docs
deadlines
and
prrr
deadlines.
I
think
that
the
release
it
I
think
we're
only
a
week
or
two
in
but
I
think
it's
like
four
weeks
or
something
I
mean
it's
two
like
we
probably
don't
have
two
more
of
these
meetings.
After
this
before
we
have
to
have
at
least
some
of
that
data
ready.
B
A
Okay,
so
I
noticed
that.
D
A
Individual
came
on
our
Sid
cloud
provider
channel
on
select
and
brought
this
up
and
I
thought
hey.
This
is
interesting
because
this
is
one
where
and
again
Antonio
was
looking
at
it
and
thinks
there's
some
failure,
and
maybe
there's
like
provider
id
missing,
API
call
stuff
going
on
and
the
tagged
Andrew.
So
this
is
just
a
few
hours
ago.
So
I'm
sure
you
haven't
had
a
chance
to
look
at
this
yet
Andrew.
But
are
you
thinking
this
looks?
I
mean
it's
right
now
it
says
it's
trash
and
I'm.
D
Yeah
we
should
probably
yeah
I
think
we
should
review
it.
Okay,
all
right,
I
think
yeah.
It's
a
valid
ask.
A
So
this
is
a
triage
accept,
do
you
want
to
look
at
it
or
do
you
I
think
Antonio
assigned
you,
but
like
do
you
want
to
try
to
bring
anyone
else
in
or
what
are
you
thinking.
D
B
I
mean
I'm
not
necessarily
signing
up
I'm,
also
not
like
shirking
away
from
it.
I
just
wanted
to
comment
that
I
think
this
is
a.
This
is
a
good
issue
for
new
reviewers
to
get
into
as
well,
because
it's
not
specific
to
one
cloud
provider
and
it
involves
much
more
generic
Behavior,
so
I
I,
like
I,
think
this
is
a
great
place
if
we
have
people
who
want
to
get
more
involved
in
the
Sig.
This
is
like
a
great
review
to
get
on.
A
Okay
and
I
saw
a
couple
of
hands,
go
up
so
cheers
then
yeah.
A
A
Okay,
I
am
and
then
I
think
Caleb.
You
were
also
raising
your
hand.
F
Yeah
I'm
new
here
so
I
think
I'm
to
be
a
good
opportunity
to
get
started.
A
Mm-Hmm
yeah:
do
you
wanna?
Why
don't
I
I'll
do
a
triage,
accept
and
assign
both
of
you
to
take
a
look
Kirsten?
What's
your
GitHub.
A
Awesome:
okay
I
just
wanted
to
draw
attention
to
that
one
because
it
came
up
on
the
on
the
chat
and
I
thought.
Oh
okay,
see
what
else?
If
anything,
do
we
want
to
talk
about
today?
Caleb
since
you're
new
here
you
wanna,
tell
us
about
yourself,
say
hi.
F
Hi
everyone
I'm
Caleb
and
I'm,
based
in
Nigeria
I've,
been
using
kubernetes
for
a
while,
but
I've
always
been
interested
in
contributing.
So
this
is
me
taking
some
practical
steps
to
that
effect.
A
Wonderful
welcome
and
thank
you
of
course
cloud
provider.
Is
you
know
very
so
collaborators,
her
Club
provider,
specific
geospecific
Cloud
providers
that
you're
usually
working
with.
A
Wonderful
there's
a
this
whole
cloud
provider
extraction
migration
project
means
that
there's
a
lot
of
moving
of
stuff
out
of
tree
that
there
is
a
Thursday
extraction.
Migration
meeting
that
if
you
joined
you,
probably
have
on
your
calendar
as
well,
so.
C
Thank
you.
I
just
saw
one
note
on
the
extraction
meeting
like
I
think
the
dog
is
out
after
it,
so
hint
it
is
the
one
which
I
got
from
the
committee's
repo.
It
mentions
the
meeting
as
1
30.
PST.
A
A
B
Yeah
I
can
I
can
follow
up
on
Kirsten's
on
Kirsten's
issue
here
and
make
sure
it
gets
updated.
Are
you
saying
is
it?
Is
the
agenda
wrong
or
is
the
docs
in
GitHub
wrong,
I
guess.
C
So
the
Dark
Link,
the
Dark
Link,
which
from
the
GitHub
when
I,
moved
to
it
like
it,
shows
the
meeting
as
130
PST,
okay,
I'll.
C
C
Yeah
I
think
maybe
they
can
because
I
have
still
my
Google
shows
like
it's
1
30ish,
so
maybe
both
at
least
I
think
fixing
the
Google
doc
to
mention
the
meeting
is
at
1
30,
like
9
30
paste
test
of
138
that
hence.
E
B
Was
gonna
I
was
going
to
mention
to
Caleb
as
well
there's
a.
We
also
have
a
sub
project
meeting
on
Wednesdays
that
alternates
with
this
meeting
on
alternating
Wednesdays
for
the
API
server
Network
proxy,
which
is
also
called
connectivity,
and
that's
another
project
that
we're
we're
kind
of
my
we're
moving
towards
General
availability
on
it
and
it's
another
area
to
get
involved.
If
you're,
if
you're
curious
or
looking
for
new
projects
to
kind
of
get
into
and
whatnot.
A
Okay,
wonderful
have
a
wonderful
day
and
week
and
hope
everybody
stays
safe
and
has
fun.