►
Description
Leader migration, Cloud provider flag in kube-up, and more.
B
C
Okay,
hey
everyone:
this
is
the
bi-weekly
cloud
provider,
extraction,
migration,
sub
project
meeting
and
I
will
share
my
screens.
C
C
C
A
All
right,
so
that's
my
five
yeah
sorry
for
that.
Okay,
okay,
so
I
finished
every.
I
finished
the
whole
process
of
me
migrating
from
casey
and
molly
to
them
with
scalps,
which
is
great,
so
I
I
would
so
first.
I
would
like
to
I'd
like
to
have
a
shout
out
for
everybody
who,
like
helped
with
this
process.
A
Although
I
didn't
see
anybody
anybody
involving
any
anybody
involved
like
in
this
meeting
so
for
now,
I'm
trying
to
ask
like
do
anybody
have
any
experience
to
do
leader
migration
outside
of
the
helps,
because
in
order
for
that
to
graduate,
I
should
have
the
automated
test
part
I
get
ready
very
soon,
because
I
just
need
to
automate
the
test.
The
whole
process
will
scale
up,
but
I,
but
I
still
need,
like
signals
from
other
qualifiers,
to
see
if
they
have
experience
doing
the
data
migration.
E
Yeah
we're
we're
doing
it
or
we've
done
it
a
number
of
times
in
test
environments.
I
don't
think
we
currently
have
a
running
automated
job,
but
we
we
have
tested
it.
F
So,
to
be
clear,
the
ga
requirements
only
call
for
one
automated,
but
we
need
a
sign
off
from
each
of
the
cloud
providers
planning
to
do
it
that
it
worked
for
them.
So
I
think,
as
far
as
amazon
goes
as
long
as
you're
willing
to
say
it
worked
for
you.
I
think
we're
happy
on
amma's
the
side.
I
think
then
probably
our
biggest
concern
would
be
microsoft
and
I
know
el
mico.
D
C
Yep
and
I
can
take
an
action
item
to
follow
up
with
you
jared
on
like
if
there's
any
more
testing,
that's
remaining
for
on
the
gk
side
of
things.
F
Okay,
I'm
probably
saying
a
little
much
googly
but
make
sure
if
you
could
follow
up
jared-
and
this
is
not
an
lss
requirement.
But
if
you
and
andrew
could
follow
up
with
the
warsaw
team.
A
F
So
I
I'm
sorry
introductions
are
probably
in
order.
Do
you
want
to
do
a
quick
round
of
introductions
mike
andrew.
C
Yeah
sure
sure
I'll
go
first,
I'm
andrew
cloud
provider
co-chair.
I
recently
actually
joined
google
last
month,
so
pretty
new,
but
been
working
in
this
area.
For
for
a
bit
walter.
Do
you
wanna
go
next.
F
Sure
hi,
I'm
walter,
I'm
also
at
google,
I'm
an
emeritus
chair
and
the
current
tl
for
one
of
the
current
tls
for
sig
cloud
provider.
I
also
have
the
lead
for
the
extraction
working
group
and
for
things
like
the
api
server
network
proxy.
D
Want
to
go
next,
yeah
sure,
hey,
I'm
mike
michael
slash,
elmico
will
call
me
online.
I
work
for
red
hat.
I
primarily
work
on
things
like
cluster
api
and
the
auto
scaler
and
stuff
like
that,
but
our
team
has
crossover
into
these
cloud
providers.
So
I
like
to
come
here
and
see
what's
going
on
and
help
out
when
I
can.
E
Nick,
hey,
I'm
nick
turner,
other
co-chair
of
sick
cloud
provider
and
I
work
at
aws
on
eks.
E
A
Hi
this
is
jiahui
or
jarek,
so
I've
been
working
for
the
extraction
working
group
for
the
leader
migration
feature
and
I'm
currently
trying
to
bring
that
feature
to
gaa
so
yeah
and
help
will
be
greatly
appreciated.
B
Yeah
and
jacob
I'm
from
google
I'm
right
recently.
I've
joined
a
project
either
to
the
cloud
providers
and,
as
as,
as
I
understand,
it's
kind
of
migrated
to
my
area
of
responsibility.
B
G
Lebron
hello,
this
is
lebron.
I
work
for
vmware.
I've
been
working
on
cloud
provider
vsphere
for
a
while,
but
this
extraction
group
is
new
to
me.
C
Awesome,
I
think
that's
everyone
right
cool,
so
going
back
to
the
topic
here
about
migration.
C
I
think
did
you
mention
jared
about
like
you're
working
on
e3
tests
right.
A
Right,
it's
right,
it's
basically
all
the
main
tests
for
to
upgrade
the
cluster
and
then
roll
it
back
with
chaos.
C
Using
which
they're
using
the
apps,
yep,
okay,
cool
and
then
yeah,
I
saw
your
pr
for
the
v1
types
I
guess
like
that'll,
be
dependent
on
like
they
do
test
them
or
not.
But
that
sounds
good.
A
So
we
used
to
have
a
resource
log
field
because
we
use
this
for
magnesius
endpoints
et
cetera,
et
cetera,
but
but
nowadays
we
can
only
do
we
can
only
support
lenses
because
I
think
voicetech
remove
support
for
anything
else
like
in
1.94.
A
So
in
so
there's
no
point
keeping
this
field
anymore.
This
was
discussed
during
the
cap
during
the
capital.
Sorry.
A
So
could
you
just
go:
go
up
a
little
bit
yeah
we
used
to
have
a
we
used
to
have
a
resource
log
field.
Could
you
go
to
the
right
yeah,
putting
it
louise
right
instead
of
this
one
yeah,
we
used
to
have
a
reason
yeah,
so
it
should
be
the
instance
or
endpoints,
but
endpoints
were
was
removed
in
124,
which
means
this
field
can
only
be
nonsense
and
because,
because
it
is
defaulting
to
lenses
and
only
can
only
take
an
instance,
then
we
may
as
well
just
remove
this
field.
A
C
Yeah
sorry,
what
I
meant
earlier
was
that,
like
we
wouldn't
want
to
promote
the
config
type
or
api
to
v1,
like
we
want
to
do
that,
along
with
the
feature
gate
promotion
right.
That's
what
I
meant
by
lockdown
that
you
do.
A
A
So
that's
right!
You
can.
You
can
delay
that
if
you
think
and
if
you
think
you
can
so
I'm
just
putting
this
like
a
little
bit
earlier,
so
that
we
can
so
that
we
can,
you
know,
have
some
have
some
reviews
a
little
bit
so
make
some
so
like
this.
C
A
F
I
I
actually
think
that
this
is
probably
a
good
idea
to
split
if
nothing
else,
one
of
the
things
that
is
inevitably
going
to
come
with
an
api.
An
api
change
is
whether
or
not
we
need
to
go
through
api
review
and
especially
for
api.
If
we
do-
and
I
don't
think
we
should
have
to
go
to
api
review,
but
I
think
it
is
likely
to
be
a
discussion
and
if
we
do
having
it
broken
out,
the
way
that
jared
had
is
probably
a
good
idea.
C
Yeah
makes
sense,
I
I
guess,
like
the
the
only
concern
I
was
raising
is
like
because
it
happens,
it
happens
before,
like
you,
promote
the
api
and
then
we
forget
about
the
gate,
or
vice
versa.
So
just
just
things
like
that,
but
like
no
objection
to
splitting
it
in
general,
unless,
unless
we're
saying
explicitly
that
we
don't
need
the
we're
okay,
decoupling,
the
feature
gate,
but
the
api
version.
A
C
A
Your
head
turn,
so
maybe
we
can
just
have
like.
If
so,
maybe
you
can
maybe
take
a
like
make
it
a
preview
for
the
api
type.
So
if
everybody's
okay
with
the
api
type,
then
it
will
be
much
faster
like
when
I
finish
off
the
testing
part
and
we
can
just
merge
those
pr's
all
together.
F
Gotcha
quick
mention
jared.
I
notice
we
still
have
v
one
alpha
one.
We
probably
should
go
ahead
and
delete
that
okay
and,
if
you're,
if
you've,
actually
noticeably
changed,
v1
beta1,
which
you
seem
to
have
by
removing
the
leases
field,
it
probably
makes
sense
to
deprecate
v1
beta1
and
try
and
delete
it
as
soon
as
we
can.
C
C
Okay,
walter,
I
think
you
have
the
next
one
about
the
cube
up,
supporting
the
clock
provider
option.
E
F
I
I
can
find
it
and
link
it.
The
the
short
version
is
there's
a
pr
that
is
merged,
which
now
gives
us
a
the
option
to
turn
off
the
cloud
provider.
Flags
when
you
run
cube
up.
F
That
is
a
precursor
to
being
able
to
test
the
two
flags,
or
at
least
one
of
the
two
feature
gate
flags
for
cloud
provider.
F
So
we
need
to
go
ahead
and
create
a
test
job
which
should
be
pretty
light,
and
then
we
can
get
a
very
good
idea
of
which
tests
are
relying
on
a
cloud
provider
being
turned
on.
E
I
have
a
question
related
to
the
disabled
cloud
fighter
and
disabled
credential
provider
flags
for
the
disable
credential
provider
flag
so
that
obviously
disables
the
credential
provider.
E
C
Yeah,
it's
one
of
those
like
weird
confusing
integrations
but,
like
the
credential
provider,
is
a
separate
subsystem
of
cubelet
independent
from
the
cloud
provider
flag
to
my
knowledge,
so
yeah
like
that's.
Why
and
that's
why
there's
two
feature
gates
because
there's
no
like
good
way
to
like
logically
bundle
it
into
the
same
feature
gate
so
setting
cloud
provider
external
but
not
disable.
The
cubic
credential.
E
Not
a
problem
necessarily
yeah,
so
the
reason
why
I'm
asking
is
just
because
we
are
deciding
whether
or
not
to
ship
our
next
version
with
a
credential
provider
and
how
decoupled
that
is
from
disabling
cloud
provider
is
a
consideration.
So
if
we
have
cubelet
running
with
cloud
provider
equals
aws,
you
know
and
we
have
the
credential
provider,
then
the
external
credential
provider,
then
there's
two
code
paths
for
getting
credentials.
F
C
C
C
Okay,
so
I
guess
the
action
item
for
this
is
to
run
tests,
maybe
like
create
a
proud
job.
C
Yeah,
probably
all
of
them
find
that
out
later
cool
anything
else.
C
E
Started
working
on
it,
I
I'll
bring
a
pr
to
either
this
meeting
or
the
other.
As
soon
as
I
have
it,.
C
Cool,
let's
end
it
here,
then
thanks
all
stay
tuned
thanks
later.