►
From YouTube: SIG Cloud Provider 2023-06-07
Description
Meeting Agenda: https://docs.google.com/document/d/1OZE-ub-v6B8y-GuaWejL-vU_f9jsjBbrim4LtTfxssw/edit#bookmark=id.qtjdzm7fl6m4
A
Welcome
to
Sig
Club
provider
for
Wednesday
June,
7th
2023
a
date
that
still
sounds
not
at
all
real.
We
have
a
few
people
here
and
a
few
agenda
items
here.
So,
let's
get
into
it
I'm
going
to
share
if
I
can
figure
out
how
to
computer.
B
A
Right
so
we're
we're
taking
a
quick
look
at
the
June
7th
agenda
for
bug
scrub,
slash,
triage
I
did
click
over
to
the
needs
triage
and
saw
we
have
a
bunch
that
have
the
label
because
they're
old
and
then
we
had
one
that
I
triage.
Just
now,
which
was
it
looks
like
we
have
a
test
flake
that
affects
Legacy
cloud
provider
Azure,
so
I
took
ownership
of
that.
Thank
you
again.
Antonio
and
I
will
follow
up
on
that
one,
but
we
can
certainly
go
back
and
retrieage
stuff
from
2021.
A
If
time
allows
but
I
think
that's
all
we
have,
we
don't
have
any
other
new.
A
You
know
issues
that
other
than
again
these
ones
that
we
should
probably
go
back
and
figure
out
if
we
want
to
keep
them
life
cycle
frozen
forever,
and
but
that's
that
covers
bug
scrub,
sub-project
updates,
I
wrote
a
little
bit
in
and
then
I
guess
we
can
just
take
them
verbally
from
others
if
desired,
for
provider
Azure.
We
did
do
releases
yesterday
for
our
active,
supported
versions,
and
these
releases
were
just
to
put
in
the
zero
downtime
migration
from
Nick
base
to
IP
based
back-end
pool.
A
So
you
know
a
little
bit
of
a
feature
ad
but
got
those
releases
out
yesterday.
Do
we
have
any
other
sub
projects
that
would
like
to
update
us.
A
I'm
going
to
take
the
silence
as
a
we
don't
have
a
specific
sub-project
update
from
any
of
the
other
folks
on
the
call
which
is
fine,
short
and
sweet
okay
into
the
agenda.
Joel
gave
us
a
point
to
look
at.
They
have
a
PR
up
to
change
some
health
check,
behavior
on
Azure,
to
resolve
an
outstanding
issue
and
to
be
more
consistent
with
other
providers.
A
The
most
current
status
on
this
was
one
of
our
colleagues.
One
of
my
colleagues
looked
at
it
and
said
we
are
thinking
about
this.
Let's
wait
and
then
Joel
I
think
would
probably
benefit
from
a
meeting
with
this
colleague.
Who's
based
in
Shanghai
I
took
I
took
a
look
at
the
times
and
obviously
this
meeting
is
not
an
ideal
time
for
our
colleague
in
Shanghai.
A
It
would
not
be
incredibly
difficult
to
find
several
hours
of
overlap
for
him
and
Joel
I
would
probably
not
be
in
this
3
A.M
meeting,
but
that
would
be
fine,
so
I'm
going
to
follow
up
on
that
and
see
what
works
best
for
them,
just
to
have
a
little
bit
more
real-time
discussion
of
that
exact
issue,
but
yeah
yes,
Walter.
C
Sorry
can
you
do
me
a
favor
and
just
quickly
open
up
which
files
are
being
modified
there.
Oh.
A
Sure
absolutely-
and
this
is
also
in
the
this
is
in
I-
mean
the
link
is
in
there.
So
it's
not
a
surprise.
So
the
the
request
is
to
change
it,
to
be
more
explicit
about
some
of
these
ports
and
to
change
the
health
check
path.
I
mean
it
all,
looks
very
standard
and
reasonable,
but
I
do
understand
the
perspective
that
perhaps
there's
something
that
could
break
and
so
I
understand
why
my
colleague
wants
to
make
sure
that
we're
not
gonna,
you
know
cause
any
breaks.
C
Yeah,
no,
no,
this
is
assuage,
is
my
concerns.
Mine
had
been
more
along
the
lines
of
we
have
policies
about
Legacy
Cloud
providers
in
KK
and
oh.
A
C
A
Yeah,
which
we're
not
oh
yeah,
absolutely
so
this
one
and
that's
a
really
good
point
actually,
because
we
and
I'm
certainly
I,
will
be
coming
to
you
probably
to
get
stuff
approved
if
that
one
over
here,
as
you
can
see
this
one
was,
this-
is
a
flake
in
the
Legacy
cloud
provider,
possibly
so
this
one
I
may
be
coming
to
you
for
approval
once
we
figure
out
how
to
deflake
it,
whereas
the
one
that
we're
discussing
that
Joel
and
pangfei
will
probably
work
through
the
details
of
is
in
club
provider
Azure,
so
I
think
that
one
will
be
okay.
A
Does
that
associate
your
concern?
We
have
a
thumbs
up
from
Walter,
excellent,
okay,
moving
right
along,
so
that
one
again
thank
you
Joel
if
you're
watching
the
recording,
but
we
will
have
to
follow
up
on
that
offline
next
up.
I
just
wanted
to
put
on
the
record
for
everyone
that
we
are
working
on.
The
Sig
leadership
change,
update.
A
I
was
actually
one
of
the
first
people
to
use
this
exciting
new
form
to
fill
out
that
and
then
I
put
and
then
I
pull
requested
in
some.
You
know
fixes
to
the
form,
and
so
that's
always
the
way
it
goes,
and
I
am
working
through
making
sure
that
El,
Nico
and
I
are
set
up
correctly
for
everything.
There's,
of
course,
we
found
a
few
places
where
things
needed
to
be
updated
worked
on
Etc,
but
this
is
just
for
complete
transparency.
A
This
is
the
safe
leadership
update
going
on
a
few
calendar
related
mailing
list.
Related
tasks
remain.
If
anyone
has
any
questions
or
concerns
about
that.
Obviously
we're
happy
to
talk
about
that,
but
I
don't
think
if
anything
more
really
needs
to
be
discussed
on.
It
is
just
working
through
the
whole
checklist.
A
Okay,
there
are
a
couple
of
things
we
do
want
to
discuss
and
I
put
a
bunch
of
detail
on
here
about.
It
occurred
to
me
that
the
this
auditory
migration
process
it
looks
like
we
got
bumped
out
of
1.28
on
account
of
not
having
production,
Readiness
review
tax
updates,
yada
yada.
This
all
happened
like
a
couple
days
ago
and
I
saw
it
all
when
I
was
prepping
for
this
meeting
and
I
thought.
A
Okay,
there's
some
work
to
do
here,
we're
probably
not
going
to
make
the
prr
freeze
like
this
week,
which
is
fine,
I
I.
Don't
I'm
not
concerned
about
us
getting
that
in
by
tomorrow,
but
we
are
going
to
have.
We
have
to
do
some,
some
writing
in
order
to
make
sure
that
we
can
actually
change
versions
and
anything
if
you
I'm
sure
if
you
had
changed
things
in
kubernetes
in
the
past,
and
then
you
go
to
try
to
do
things
now
and
you're
like
yeah,
it's
just
being
marked
as
beta.
A
Oh
no,
first,
you
have
to
fill
out
the
production
Readiness
review.
You
have
to
do
a
bunch
more
stuff
which
is,
which
is
all
fine
I
am
aware
of
this
I
just
put
this
in
here
for
transparency
that,
like
yes,
we
have
a
bunch
of
work.
We
need
to
do
here,
which
means
we're
probably
not
going
to
land
this
stuff
by
tomorrow.
A
If
anyone
wants
to
click
on
these,
read
about
these
put
some
comments
or
input
in
totally
welcome,
but
just
putting
this
out
there
that
it
looks
like
the
the
test
at
hand
are
going
to
be
the
production
ready
in
this
review
and
limiting
the
set
of
CI
jobs
to
the
ones
that
we
actually
think
are
crucial
and
I
need
to
talk
to
almaco
about
that
possibly
Walter,
possibly
Andrew
Etc,
whilst
you're
looking
pensive.
Do
you
have
anything
to
add
to
this.
C
C
And
so
I
think
El
Miko
was
going
to
look
into
what
it
would
take
to
actually
run
the
tests
with
that
build
Target,
which
I
think
would
be
go
a
long
way
towards
if
we
can
set
a
CI
job
up
with
that.
I
think
that
goes
a
long
way
to
meeting
dims's
concerns
and
in
fact
mine,
yeah.
A
Yeah,
oh
I
think
we
all
I
think
we're
all
agreeing
with
the
concerns
like
no
one
wants
to
make
every
single
cloud
provider
and
every
single
end
user
said
that
would
be
bad
so
anyway,
I
know
that
El
Miko
had
a
conflict
with
today,
but
I'm
gonna,
say
action
item
I'm
gonna
follow
up
with
him
to
gather
that
information
specifically.
C
A
C
I
can
push
for
Google
to
be
another,
and
if
we
can
get
more
I'm
I'm
all
for
more
I,
just
don't
want
to
see
less
than
two.
A
Okay,
cool
we'll
see
if
we
can
talk
to
Nick
and
see
if
he
can
also.
A
Okay,
so
that
that
is
investigation
between
now
and
the
next
call
I
guess.
A
Alice
we
should
talk
about
on
this
one.
Before
we
move
on
to
cloud
provider.
Equinix
I,
don't
know
how
to
pronounce
it
so
I'm
gonna
have
to
find
out
from
Chris
metal,
Equinox
equinix
excellent,
so
they
put
a
request
in
I,
saw
it
and
I
thought
I
hope.
Somebody
looking
at
you
Walter,
who
has
been
around
City
Club
router
a
little
longer,
can
give
us
some
guidance.
Is
this
like
I
I,
recognize
that
this
is
one
of
those
things
where
people
comment
and
give
support?
A
Is
this
one
of
those
things
where
we
does
a
cloud
provider?
Do
we
have
like
a
a
checklist
of
hey?
If
you
would
like
to
make
what
your
cloud
provider
an
official,
Sig,
Club
provider,
you
know
item,
why
don't
you
do
X
I,
just
I'm,
not
positive,
if
I'm
allowed
to
just
be
like
come
on,
come
all
or
if
there's
some
sort
of
you
know,
you
must
be
this
High
to
ride
the
club
provider.
Ride
I
have
no
idea.
C
Great
questions
we
have
traditionally
pretty
close
to
said.
Come
one
come
all
I'm
glad
Chris
is
here
because
it
it
there
is
one
or
two
things.
I
would
like
to
chat
about
if
a
person's
okay
with
that
yeah,
so
one
just
so
we're
totally
transparent.
If
you
want
to
move
to
KK
as
opposed
to
having
your
own
repo,
the
one
I
mean-
and
this
is
true
for
Amazon
as
I.
B
C
100
I'm
using
I'm
using
it
that
term
a
little
more
broadly
than
it
technically
applies
so
good
call
out,
but
anything
under
the
kubernetes
cncf
is
open
source
so
as
as
I
kind
of
just
referenced,
the
AWS
cloud
provider
under
cncf
is
largely
maintained
and
owned
by
a
googler,
and
this
is
allowed
because
it's
open
source
and
anyone
can
do
it
right.
So
I
just
want
to
be
very
clear
if
you
move
to
cncf.
This
doesn't
mean
this
is
my
company's
little
private,
followic
and
Oakley.
We
get
to
be
owners
in
there
and.
C
C
It
is
a
good
thing
to
be
to
to
be
very
transparent
about
upfront,
so
I
mean
one
of
the
things
related
to
this
is
that
Bridget
I
El,
Nico
and
Nick.
Since
we
are
the
the
leads
for
say,
cloud
provider
are
automatically
going
to
be
backup
owners
in
case
you
all
disappear
and
there's
an
open
source
contributor
who
starts
contributing,
and
so
that's
going
to
be
the
big
one.
The
other
one
I
will
mention,
and
you
can
come
and
ask
questions
in
either
in
any
of
the
Sig
cloud
provider
about
the
details.
C
C
There
are
and
you're
probably
and
six
storage
would
be
good
for
the
CSI
stuff.
But
you
may
well
want
your
own
CSI
provider
and
those
are
things
that
are
normally
stuck
in
this
kind
of
repo.
C
D
We
don't
we
don't
have
a
yeah,
we
don't
have
a
CSI
or
cni.
Maybe
there's
some
discussion
over
whether
we
have
a
cloud
provider
or
a
CCM,
though
so
that's
traditionally
a
sticky
point
for
me
and
where
the
line
for
one
ends
and
Begins
for
the
other.
C
So
there
are
four
there's
generally
four
operators
that
have
cloud
provider,
specific
logic
and
those
are
going
to
be
disappearing
from
the
KCM.
C
So
it's
things
like
the
node
ipam
controller
and
you
are
welcome
to
run
your
own
CCM
and
pull
in
the
default
operators.
But
it's
going
to
be
things
like
the
node
ipam
controller
and
the
node
lifecycle
controller
and
especially
like
the
node
lifecycle,
controller,
I,
you're,
probably
going
to
want
to
think
about
how
that
applies,
because
as
an
example,
one
of
the
things
that
the
node
lifecycle
operator
does
is
if
your
node
is
for
the
cubelet
running
on
your
node,
is
failing
to
respond
to
or
or
generate
heartbeat
messages.
C
After
a
certain
period
of
time,
the
operator
is
going
to
degrade
it
and
then
is
going
to
want
to
check,
call
to
you
and
determine
if
that
VM
still
exists,
and
there
is
no
kubernetes
way
of
doing
that.
So
every
that's
the
sort
of
thing
that
generally
every
cloud
provider
ends
up
having
to
implement
for
themselves.
D
Sure
that's
not
kind
of
my
question,
though,
like
are
you
saying
that
the
CCM
bits
should
be
in
a
separate
repo
from
the
cloud
provider.
C
B
D
Yeah
should
all
be
good.
There
I
have
a
PRN
to
kind
of
get
our.
You
have
some
like
default
templates
for
everything
that
I'm
trying
to
get
us
more
in
line
with
so
I'm
working
out
that
bits,
but
everything
else
looks
like
we
should
be
in
compliance,
but
well.
A
A
I'm
gonna
put
an
for
me
to
comment
with
support.
We
also
maybe
we'll
need
a
couple.
Other
people
Walter,
perhaps
could
go
comment,
but
I
think
the
main
thing
I'm
just
going
to
look
at
is
hey.
Do
you
have
any
completely
wacky
out
there?
A
D
Looked
at
it
and
that's
fine,
yeah
I
I
have
a
PRN
that
replaces
our
license
file
with
the
default
Apache
2
license
file
used
by
everything
else.
I
think
our
license
file
says
it's
Apache
too.
It
just
doesn't
state
it
with
the
same
text.
Expressions
that
you
do
so
I
I
just
want
to
I.
I
am
conscious
of
that
one
thing
being
the
only
stickler
potentially
from
an
open
source
perspective,
but
it's
open
source
either
way
and
yeah.
We
have
the
same
goal:
I
just
want
to
make
sure
anyway,
yeah.
A
A
You're
here
and
we
do
not
want
to
keep
you
in
any
way,
I
just
don't
want
to
accidentally.
Do
it
wrong
and
make
you
have
to
start
over
in
an
annoying
way.
A
D
C
A
A
Just
hate
to
bother
you
well
I
hate
to
bother
him
when,
when
we're
coming
up
to
like
freezes,
you
know
I
think
actually
Walter,
you
might
be
the
other
person
who
needs
to
go
approve.
Those
I
will
message
them
to
you
separately.
A
A
D
D
D
C
I
will
also
mention
it
would
be
good
if
you
continued
to
show.
We
do
have
changes
in
coming
for
cloud
providers
and
it
can
be
good
to
stay
on
top
of
it.
We'd
also
just
love
your
feedback
on
various
things,
yeah.
D
That
that
I
mean
that's
one
of
the
goals
of
contributing
was
to
get
more
in
line
with
Upstream
changes,
features,
goals,
Etc
and
also
we're
looking
for
another
place
for
employees
to
contribute
that
counts
as
yet
another
project
that
they're
contributing
to
Upstream,
because
they
we
have
cluster
API.
If
we
can
have
this
under
the
k6
as
well,
that
helps
getting
people
into
the
org.
So.
A
That
is
fantastic
and
to
the
point
of
Walter,
alluding
to
the
fact
that
we
definitely
need
perspectives
on
changes
like
you're,
probably
going
to
be
the
sort
of
folks
who,
for
example,
could
really
get
into
test
Grid.
Or
you
know
any
of
the
other
places
where,
like
contributions
to
improving
tests
and
decreasing
flakes,
is
actually
probably,
in
my
opinion,
one
of
the
really
valuable
contributions.
Because
that
means
you
don't
waste
people's
time,
because
if
somebody's
sitting
there
watching
proud
and
then
they're
like
and
afflict.
B
A
Great
okay,
so
I
will
I'll,
go
check
and
find
those
Piers
or
those
Cherry
picks
that
I
mentioned
and
send
them
to
Walters
separately.
Is
there
anything
else
we
should
talk
about
today
on
this
call.
B
I
recently
joined,
like
I,
mean
AWS,
so
I'm,
like
kind
of
trying
to
take
over
like
not
take
over
but
like
Nick,
is
kind
of
like
moving
into
Earth
kind
of
thing
she
got.
So
he
asked
me
to
like
just
see
to
the
cloud
Toyota
stuff.
A
little
bit
I
was
I
was
I,
mean
I'm,
aware
of
the
process,
because
I
was
like
involved
in
sick
apps
before
so.
B
I
just
have
like
one
doubt
regarding
the
AWS
thing
for
the
CI
part,
so
my
understanding
was
that
AWS
code
is
moved
out
of
Upstream
report
completely.
C
So
if
you
don't
mind,
I'll
answer
that
yeah
yeah,
so
here's
the
thing
there
are
two
P.
There
are
two
groups
that
will
benefit
from
this
and
and
so
I'm
gonna
make
both
benches.
C
C
You
also
are
probably
more
care.
If
kubernetes
someone
makes
a
change
that
makes
kubernetes
no
longer
work
with
Amazon,
or
at
least
the
Amazon
repub,
so
your
CCM.
C
So
in
all
cases,
cloud
provider
is
set
to
external
and
then
the
Deployable
all
it
then
installs,
the
CCM
it
actually
installs
the
CCM.
The
credential
provide
the
the
Google
credential
provider
and
the
group
The
Google
CSI
driver,
and
then
it
runs
the
full
CI
suite
and
so
what
it
does
is
we
have
a
job
on
the
cloud
provider
gcp
that
is
generate.
That
is
generating
a
known
good
version
of
each
of
those
components
and
we
are
pulling
those
images
in
our
install
script
and
then
running
the
suite.
B
C
So
that's
the
sort
of
CI
job
we're
talking
about.
In
fact,
what
I'd
like
to
do
is
the
it
should
be
Superfluous
to
to
have
the
cloud
provider
external
flag.
In
fact,
what
I'd
like,
which
is
what
I
was
mentioning
earlier,
that
alniko
is
working
on
is
to
be
able
to
run
those
jobs
with
binaries,
where
the
cloud
provider
code
has
just
been
compiled
out,
and
so
that,
specifically,
is
what
we're
talking
about
trying
to
run
a
CI
job
and
that's
when
we
think
you're
likely
to
want
to
run
in
perpetuity.
B
Okay,
yeah
I
just
want
to
like
get
that
answer
that
okay.
This
is
more
for
long-term
thing.
To
make
sure
that
KK
code
changes
are
not
breaking
the
cloud
provider
than
more
about
removing
the
in-cloud
provider.
C
Well,
it's
actually
about
remote,
removing
the
enclosure.
So
so,
once
we
get
at
least
two
greens
from
cloud
providers
that
we've
removed
the
code,
at
least
in
test
and
nothing
broke,
then
the
goal
would
be
to
actually
take
all
of
the
code
that
the
that
those
build
flags
have
are
disabling
and
actually
just
delete.
That
code.
B
Second
part,
because,
like
at
least
my
understanding
is
that
currently
there
is
no
code,
I
mean
we
can't
deploy
in
I,
mean
AWS
cloud
with
entry
with
Cloud
providers
internal,
it
has
to
be
always
external,
because
the
cube
the
Legacy
Cloud
product
doesn't
have
any
code
related
to
AWS.
B
C
May
be
true,
but
I
will
say
that
whether
that
is
true
or
not
more
than
50
percent
of
each
of
the
binaries.
That
I
just
mentioned
is
cloud
provider
code.
Okay,.
B
B
A
B
A
C
Just
quickly
for
Kirsten,
it
would
be
good
for
Bridget,
I
and
El
Miko
to
understand
if
Nick
is
moving
on
to
Sig
auth.
Is
he
giving
up
his
lead
chair
in
sick
cloud
provider,
or
is
he
planning
on
maintaining
that
that.
B
I'm
not
sure
I
can
check,
but
I
can
I
can
definitely
check
with
him,
but
she
yeah
sure
I'll
I'll
ask
him
like
I'll
mention
about
this.
A
Oh
yeah
thanks
definitely
important
and
it's
one
of
those
things
where
we
definitely
should
get
because
I
did
I
just
went
through
that
big
checklist
that
you
saw
of
like
changing
people
around,
and
so,
if,
if
more
changing
around
needs
to
happen,
we
should
get
that
done.
B
Okay,
I
will
message
him
and
asked
to
update
and
absolutely
thank.
A
All
right,
I
think
that
we're
good.
Thank
you
all.
So
much
I
will
see
you
all
next
time.
Bye,
bye
thanks.