►
From YouTube: 20200128 sig cluster lifecycle
Description
Kubernetes SIG Cluster Lifecycle meeting for January, 28 2020
A
B
A
And
we'll
do
the
standard
go-round
I
did
want
to
give
a
PSA
update
with
regards
to
a
bunch
of
conversations
have
been
having
with
the
CN
CF
regarding
our
toxic
kuchen.
Both
that
could
cover
year
going
forwards.
So
I
think
I
should
talk
about
that
unless
there
are
any
burning
other
things,
it
looks
like
Danielle.
C
Yeah,
it's
just
a
brief
thing
in
a
massive
just
advanced
call,
we
sort
of
decided
that
we
want
to
rename
our
repository
from
Eden
operators
to
cluster
add-ons
I.
Think
all
we
need
now
is
me
just
sort
of
give
his
plus
one
on
this
already.
We
just
won
this
on
the
issue.
I
think
we
can
move
forward
and
afterwards
I'm
going
to
announce
it
on
the
mailing,
this
and
everywhere
else,
which
is
for
consistency
of
reading
reasons
that
we
are
going
to
use
this
now,
I.
A
D
F
You
can
read
basically
I
propose
a
simple
format
where
we
can
have
a
list
of
general
questions
and
then
so
the
sub-projects
can
miss
their
project.
Specific
questions,
I
added
a
tentative
deadline
like
two
weeks
before
Keep
Calm,
and
this
gives
us
enough
time
to
prepare
dark,
shows
survey
across
the
API
already
provided
the
list
of
questions
thanks
to
this
for
the
rest
of
the
so
projects.
If
you
are
interested,
you
can.
F
A
Because
this
is
going
out
primarily
with
kuqali,
you
I
think
I'm.
Ok,
with
the
current
incantations,
we've
had
a
long-standing
set
of
problems
with
feedback
from
China
I,
don't
know
if
we
have
a
good
solution
that
works
more
generally.
Does
it
do?
What's
the
standard
survey
thing
that
everybody
uses
come.
A
A
A
G
F
Supports
that
my
descendent
of
the
charts
and
data
selection,
it's
every
month,
even
better,
so
this
is
like
a
detail
that
they
don't
really
know
that
much
about
I
have
questions
about
this
shared
account.
It's
only
two
people
are
allowed
to
share
the
account
very
much
I
do
not
know
it,
but
maybe
a
single
person
can
export
this
spreadsheet
and
share
it
with
everybody.
Maybe
that's
how
it
won't
work
yet.
B
A
Thanks
for
joining
us
I
think
we're
overdue
for
to
get
a
feedback
loop
again.
I
do
know
that
just
for
those
who
are
new
to
this
last
time
we
did
this.
It
was
actually
super
helpful,
Phyllis,
a
lot
of
interesting
data.
We
did
this
with
the
Ruby
mga
announcement.
It
actually
helped
shape
a
bunch
of
evidence
and
slightly
reprioritize
things.
We
knew
what
people
wanted
from
Rafi
back,
but
I
think
getting
the
raw
under
stilled
unfettered
pulse
from
a
large
swath
of
folks
is
always
beneficial
and
keeps
us
honest.
F
Potential
next
version
of
clad
go
is
going
to
include
a
context
argument
in
all
the
API
calls,
which
is
a
breaking
change,
so
the
plan,
for
maybe
a
missionary,
is
to
have
a
couple
of
clients,
one
that
has
the
context,
support
one
without
it.
But
you
have
to
decide
that
if
you
are
an
external
project,
nobody
is
going
to
update
you
with
this.
So
if
you
want
to
transition
to
the
context
client,
you
have
to
perform
some
sort
of
main
automatic
conversion
and
I
believe
Mike.
Denis
already
has
a
tool
for
that.
F
A
H
A
This
is
existence
for
what
it's
worth
going
back
to
years
of
your
I
was
actually
one
of
the
people
that
plunges
through
API
machinery
all
the
way
through
and
CD
a
long
time
ago.
I,
don't
know
why
they're
just
doing
it
now
for
the
work
or
Clank.
Oh
it's
it's
funny.
So
we
have
the
context
all
the
way
3
p.m.
you
should
read
for
what?
Where
is
this
I?
Don't
know
two
and
a
half
years
and
we
haven't
actually
enable
it
for
people
using
it.
I'm
sorry,
I'm
fetching
them.
A
F
F
A
A
What's
the
happy
path
going
forward,
but
the
TLDR
is
that
CNCs
said
that,
because
CF
is
growing
in
the
number
of
projects
that
the
CN
CF
cones
besides
communities
proper,
that
they
are
running
low
on
space
for
the
maintainer
drunks
right,
so
they
wanted
to
not
have
support
for
every
sub
project.
You
know
sick
coaster.
Lifecycle
is
a
federated
sub
projects
like
where,
basically,
we
have
a
bunch
of
different
sub
projects,
then
I
want
to
have
a
separate
talk
track
for
every
single
coop
con
for
the
entire
list
of
sub
projects.
A
So
as
a
compromise,
at
least
for
the
time
being
until
we
actually
defined
policy
up
the
stack,
we
kind
of
have
a
informal
agreement
to
say
that
for
the
time
being,
Saints
such
as
signature
lifecycle
could
request
to
the
three
talks
in
the
new
teenager
Gupta
and
we'll
have
to
do
a
rotation
periodically
for
some
of
the
sub
projects.
So
this
go-around
we've
submitted
talks
for
Canadian
and
cluster
API
for
could
come
on
EU
and
I.
Don't
know
if
the
making
folks
were
able
to
submit
their
talks
to
Nancy
and
get
them
accepted.
A
It
has
is
so
interested
that
we
might
want
to
be
giving
a
update
during
Kuk
on
North
America,
but
that
basically
means
that
other
talks,
if
many
kid
was
accepted,
as
well
as
per
diem,
that
those
talks
would
have
to
recycle
the
games
ground
for
a
different
kun
town
and
again,
this
is
more
like
a
PSA
for
the
timing.
It
will
try
to
work
out
form
the
policy
with
steering
to
folks
how
many
hot
takes
thoughts
objections.
It's
good,
it's
good
to
take
them.
Let
us
know
now,
because
we'll
take
them
to
steering
I.
A
I
Percussionist,
can
you
hit
me
yes,
so
I
had
something
to
do
with,
because
a
ship
is
a
project
under
open
infrastructure,
and
that
participates
in
using
the
kubernetes
cluster
through
copy,
as
well
as
middle
key.
So
if
they're
not
link
presentation
that
they
need
to
these,
will
that
be
part
of
the
con,
or
will
it
be
part
of
something
different?
You.
A
Can
always
there's
two
separate
tracks?
There's
the
maintainer
track,
as
well
as
the
actual
koukin
proper
self,
they're,
always
open
to
silly
talks
to
the
lame
you
have
to
go
through
the
standard
purity
process
to
get
accepted
to
KUKA,
because
you
are
not
actually
housed
by
the
sink
in
a
official
sub-project
of
the
city.
That
means
that
they
would
not
be
on
the
mutator
track.
A
E
Was
just
gonna
say
like,
as
this
talk
progresses
the
as
the
discussions
progress?
The
idea
of
having
some
form
of
informal
space
would
probably
be
a
nice
compromise
in
that
I
think
it
is
value
Paulo
in
addition,
and
like
I
know
that
they
quoted
a
fairly
large
number
for
the
price
of
a
room,
and
it
wasn't
clear
to
me
how
much
the
room
was
the
physical
space
and
how
much
was
the
people
and
the
video
equipment
and
sound
equipment.
So
like
whether
there's
room
for
negotiation
there.
A
Sometimes
it
doesn't
have
to
be
a
structured
talk
just
having
that
high
bandwidth
conversation
in
that
venue
is
enough
to
suffice
for
a
long
list
of
projects
and
I
think
that
is
maybe
even
more
beneficial
and
if
you
could
do
that
during
lunch
time,
when
there
are
other
things
scheduled,
anyways
and
they're
there.
Mr.
Lee
talks
I
think
that
might
actually
be
a
component.
So
I
will
raise
that
as
a
as
a
possible
option.
A
J
Yeah
I
was
I
was
thinking
something
along
the
lines.
I
think
what
Justin
was
saying
that
I
mean
I
I
had
a
question
actually
whether
you
know
whether
we
had
talked
or
whether.
What
do
we
think
about
top,
you
know
doing
talks
as
opposed
to
sort
of
a
two-faced
I.
In
my
experience,
the
base
basis
is
quite
useful.
You
know
I
guess
if
you
know,
if
we
can
do
it
informally,
then
that's
that's
fine.
A
A
good
question
and
I
think
I
think
we
need
to
hammer
that
when
I
with
contributes
to
as
well
traditionally
in
the
past,
we
did
face
to
faces
during
contributors
of
it
right
and
last
contributor
last
to
contribute
service.
I
think
we're
a
little
bit
weird
but
they're
in
it.
Work
like
the
traditional
style
face
to
faces
that
we
used
to
do
last.
One
was
very
weird,
the
one
before
that
was
just
awkward
because
there's
a
lot
of
contention,
so
certain
topics
dominated
the
face
to
face
discussion.
A
But
you
know
fool
me
once
I
fell
prey
to
that
again,
the
this
time
I
think
I
will
work
with
the
Petrovich
group
too,
as
well
to
figure
out
if
we
can
go
back
to
that
that
original
format,
which
I
think
was
super
beneficial
to
a
lot
of
the
contributors.
In
my
opinion,
I
found
that
to
be
more
beneficial
than
over-fitting
towards
new
contributors,
I
think.
J
A
D
So
I
actually
had
a
recent
discussion
with
Bob
about
this
and
there's
a
survey
out
there
right
now
asking
when
folks
would
like
to
have
the
contributor
summit
for
Austin,
whether
whether
it
should
be
a
zero
like
it
currently
is
or
de
minus
one,
and
if
it
does
get
moved
to
day
minus
one.
One
of
the
things
that
they're
looking
at
is
potentially
offering
space
to
cigs
for
things
like
face
to
face
and
other
types
of
collaboration
on
day
zero.
F
This
is
mostly
kept
update
for
Kapadia.
We
had
a
big
discussion
on
the
kept
for
moving
Cuba.
The
amount
of
kaykai
with
seek
release
mainly
and
we've
decided
that
it's
not
appropriate
to
move
it
in
1:18,
because
the
work
is
so
much
too
many
variables
too
many
unknowns.
So
this
is
PC
for
comedian
move.
A
Any
questions
from
the
work
I
do
have
a
question.
The
topic
of
cluster
API
being
NIDA
believable
never
ceases
to
end
and
the
the
topic
of
trust,
the
state
of
the
comedian
operator.
Could
we
link
to
that
and
maybe
even
talk
about
that
similar
instead
of
like
campy
documentation,
I,
think
that
would
be
beneficial
to
folks
who
incessantly
ask
this
question?
Well,.
F
C
A
G
So
we're
trying
to
just
like
set
like
kind
of
like
if
the
first
release
that
lucky
first
row
provider
can
rely
on
almost
feature
complete,
but
we'll
see
like
what
what
can
be
done.
The
other
thing
that
we're
doing
it
is
we're
rooming
the
backlog
really
closely
and
I'm
trying
to
look
upon
things
that
like
we
cannot
do
or
they
feel
extra
and
also
late
waiting.
Like
some
good
first
issues
this
week
we
do
have
a
good
math
draft.
G
G
G
That's
that's
a
good
idea
and
I
think
these
are
big,
could
copper
so
with
that,
but
hopefully,
like
you
know
what
conversations
like
in-person
and
with
companies
that
are
using
it
that,
like
we
can
add
them,
get
them
to
commit
something
about
maturity.
Like
I
mean
the
alpha
scope
is
like
it
is
what
it
is
I
would
say
like
we're,
not
trying
to
be
naive
or
anything
we're
mostly
trying
to
keep
things
simple
for
now,
so
that
we
can
accomplish
our
things.
Writing.
A
K
Yeah,
thank
you.
So
there's
great
work
happening
in
the
sick
right
now.
A
few
contributors
are
working
on
couplet,
flag
migrations,
which
is
the
Kubla
configuration,
but
some
of
the
untouched
portions
of
what's
remained
in
the
CLI.
We
have
a
contributor
who's,
working
on
coop
controller
manager,
config
and
making
sure
that
we
push
something.
K
And
then
we
have
OB
tech.
It's
links,
Alex
and
like
toughen,
who
you
all
probably
know.
Well,
they
will
be
taking
ownership
of
presenting
our
maintainer
track
session,
submit
that
and
then
just
as
a
final
note
with
all
of
the
work
that's
been
happening
in
the
sig
just
want
to
thank
everyone
for
their
sponsorship
and
extended
thanks
to
the
folks,
who've
stepped
up
and
contributed,
as
most
of
these
people
are
doing
this
on
their
own
time
than
or
not
nothing
paid
to
do
so.
K
F
May
I
have
a
quick
question
with
regards
to
the
couplet
flagged
application.
So
I
saw
the
ticket
that
somebody
created
about
the
list
of
flags
that
we
want
to
remove
in
deprecated
and
I
saw
somebody
sitting
up
ER
immediately
to
deprecate
the
cube
config
flag
on
the
coverlet,
which
is
the
cube
config
used
for
client-server
communication
with
the
API
server,
which
is
quite
disruptive
to
existing
bootstrap
TOS
bootstrap
processes.
F
F
K
Personally,
I
think
there's
a
lot
to
be
done.
I
know
that
you
and
I
have
shared
some
opinions
about
like
supporting
flags
long
term
in
accordance
like
in
compliance
with
their
version
de
pais.
Let
me
be
even
having
like
an
API
version
flag
that
applies
to
the
command
line,
but
yeah.
That's
I'm,
not
aware
of
the
status
of
that
issue
that
you
talked
about
so
I'm
looking
for
it
now.
F
Yeah,
maybe
this
is
a
petition
discussion
with
Michael
I
asked
her
Oh
Steve.
He
can
join
the
working
group
meeting
today,
but
I
think
he
was
away
from
keyboard.
I
was
cannot
join,
but
this
is
an
interesting
topic
whether
we
want
to
keep
these
flags
forever
or
how
are
we
going
to,
for
instance,
candle
upgrades
in
Cube
ATM
system
defense,
because
this
is
a
breaking
change.
Well,.
A
A
Well,
we
do
actually
have
a
formalized
process
for
duplication,
but
there's
no
way
first,
because
there
are
so
many
flags
on
the
order
of
hundreds
there's
no
way
for
individual
contributors.
I
see
how
what's
the
transitive
closure
of
all
the
people
are
using
simply
a
lot
of
times
they're
like
well,
we
don't
use
it
so
I'm
going
to
remove
it
or
we
could
potentially
shift
this
into
a
separate
location.
So
a
lot
of
times.
What
happens
typically
is
they
will
try
to
do
something
which
they
think
is
the
right
thing
to
do,
not
realizing.
A
H
E
Yes,
so
cops
update,
basically
have
pretty
much
business
as
usual,
like
rolling
forwards
and
cherry-picking
and
releasing,
and
then
I
have
previously
said.
There
was
nothing
of
cross-project
relevance
and
I
remembered
that
we
have
made
some
interesting
or
potentially
interesting.
Tweaks
are
we
who
are
either
attests
so
previously
we
had
these
released
branches
for
cops
and
we
did
not
like
you
test
them
at
all
when
we
had
a
PR
against
them
and
we
now
do
et
tests
against
those
order,
kubernetes
versions,
which
I
think
is
obvious.
E
D
D
E
J
A
few
nice
updates
there
QPR's
in
review
or
one
that's
going
to
improve
the
end-to-end
testing
story.
The
cover
a
TDM
join,
then
there's
a
command
to
renew
certificates
and,
finally,
an
improvement
to
make
join
more
tolerant
of
transient
failures.
So
if
there's
you
know
network
glitch,
while
you're
joining
you
don't
have
to
reinvent
it
a.m.
unless
you,
unless
you
want
to
say
that
that's
going
to
be,
this
is
going
to
be
fly
for
that
yeah.
That's
that's
holy
update
right
now.
I
do.
A
A
J
The
support
get
it
get
introduced
in
3
4,
but
the
the
default.
You
know
when
you're
joining
us
still
as
a
you
know,
there's
a
regular
member
so
there's
something
that
that
I
think
is
going
to
be
coming
in
3
I'm,
not
sure
if,
especially
in
3
5
but
Otto.
You
know
this
notion
of
Auto
promotion,
but
I
think
before
that
I
think
the
having
a
learner
mode
be
the
default.
I
think
that
that
is
going
to
come
in
the
future.
J
A
A
If
that's
the
latest
edition
curricular
writing
3
3
10,
which
is
out
of
date
at
least
a
bare
minimum
update,
2
3
3
18,
which
is
the
latest
release
version
of
3
3
I,
would
I
would
move
to
go
to
3
4
in
this
cycle
for
118
plus,
just
because
we
want
to
get
that
learner
mode
feature
in
as
quickly
as
we
can.
Is
that
super
important?
Is
there?
Anybody
who
wants
to
I'd
be
happy
to
training
somebody
on
this?
F
The
server
is
at
34.3
the
client
and
not
sure
why
it's
trailing
behind.
This
is
usually
maintained
by
Joe,
Bates
and
folks
who
decide
if
this
is
the
appropriate
moment
to
bump
the
client.
I
have
no
information.
What
by
the
client
is
not
that
you
forget,
but
I
can
tell
you
a
couple
of
things
about
learner
mode,
so
in
qadian
we
gain
nothing
from
because
learners
have
a
cup
of
maximum
one
learner
per
question,
and
we
still
have
to
have
our
logic
for
handling
the
like
weight
between
the
members.
So
that's
that's.
A
major
problem.
F
A
A
That
CD
having
multiple
members
join
at
a
time,
so
we
wanted
to
basically
have
multiple
members
become
learners,
and
then
then
you
can
just
declare
didn't
say:
I
want
three
notes.
The
problem
today
is
that
we
can't
do
that
all
concurrently,
so
instead
we
you
do
almost
like
a
rolling
change
inside
of
Kathy
and
we'll
probably
have
to
continue
to
do
that.
It.
L
A
M
Had
some
conversations,
I
think
there's
still
gonna,
be
some
ongoing
discussion.
I,
don't
think,
there's
any
like
major
external,
hopefully
I
think
you
may
still
need
to
discuss
whether
you
know
what
Google
wants
to
investing
in
I.
Don't
expect
a
problem
if
it's
not
like
officially
resolved
on
our
side.