►
From YouTube: SIG Cluster Lifecycle 2020-07-28
A
A
A
Okay,
so
this
is
like
the
important
topic
for
today,
but
I
don't
see
any
representatives
of
the
cube
aws
project.
A
So,
basically,
that
the
discussion
here
was
that
the
cube
aws
project
wants
to
move
outside
of
the
kubernetes
incubator,
because
a
kubernetes
incubator,
so
they
have
a
different
reason,
but
something
that
is
happening
right
now
is
that
kubernetes
incubator
is
being
closed
and
cubadews
is
one
of
the
two
projects
that
remain
there.
A
B
A
Out
like
when
did
this
happen
like
how
did
we
end
up
being
the
super
project
owner
and
I
I
found
a
very
old
pr
that
did
some
honor
file
automation
and
I
don't
believe
that
any
sig
chair
at
the
time
even
agreed
to
this
ownership
of
the
occupied
awareness
project,
but
in
any
case
they
they
want
to
move
outside
of
incubator
under
kubernetes
6..
A
Originally,
I
said
we
had
a
discussion
with
andrew
psyking
with
aaron
cricken
barrier
and
the
q
by
the
west
folks,
and
I
said
that
maybe
this
project
already
overlaps
with
cops.
We
don't
want
too
many
projects
that
overlap
with
each
other.
A
sequester
lifecycle
level
and,
like
andrew
psykin,
said
that
hey,
maybe
we
can
move
it
to
a
squad
provider,
but
I
think
the
the
better
option
is
to
let
the
folks
move
to
a
separate
organization,
and
at
this
point
it's
a
matter
of
agreeing
discussing
with
them.
A
And
supposedly
this
meeting
was
the
forum
where
we
should
have
this
discussion,
but
given
they
are
not
here,
we
should
just
like
state
our
position
and
I
can
update
them
update
them.
On
slack,
we
have
a
private
discussion
with
the
maintainers.
B
B
C
Yeah,
I
I
would,
I
agree
with
the
idea
of
paring
down.
I
would
say
that
we
historically
have
tried
not
to
pick
winners
as
it
were,
so
we're
trying
to
identify
commonalities
and
if,
if
a
project
wants
to
wants
to
get
in
and
wants
to
like
take
on
the
the
code
of
conduct
and
the
ci
wants
to
like,
become
more
integrated
in
the
ci
signal
and
wants
to
report
in
and
be
subject
to
sort
of
the
governance
or
the
like
touch
governance,
then
I
feel
like
we
should.
C
We
should
accept
that
and
we
should
do
that
and
we
should
like
if
it
was
a
new
project,
we
just
strongly
encourage
them
to
contribute
to
something
else,
but
as
it's
an
existing
project,
I
am
not
certainly
from
the
point
of
view.
I
don't
have
a
problem
with
it
or-
and
I
I'm
you
know
we
have.
We
have
historically
said
that
we
are
not
going
to
have
one
tool.
C
Like
that's
a
non-goal,
we
will
have.
We
aimed,
I
agree
with
you.
Absolutely
we
want
quality
building
blocks,
and
so
we
don't
want
to,
like
necessarily
like
have
a
like
20
alternatives
for
like
cluster
add-ons
or
kubernetes
or
etsy
daily,
but
at
the
at
the
composition
level.
I
don't
think
it
hurts
us
to
have
multiple
options.
A
A
They
do
not
know
our
process.
They.
I
don't
think
they
even
know
knew
that
we
have
a
mailing
list
or
this
regular
meeting.
So
if
they
are,
they
agree
to
move
under
the
sig.
They
have
to
start
following
the
whole
process,
agree
to
the
whole
diagram
of
some
projects
that
we
have
it's
a
slightly
more
complicated
contract.
A
A
And
you
don't
have
to
obey
to
all
the
rules.
We
have
it's
a
much
complicated
like
contract
than
that,
and
you
know
so.
The
maintenance
started,
leaning
towards
you
know
extracting
cube
aws
under
there
is
a
cube,
aws
orc
on
github
already.
A
I
really
wish
that
the
maintainers
were
here
to
confirm,
but
like
in
terms
of
our
stance,
do
we
want
us
to
comment
on
the
pr
there's,
there's
a
pr
that
the
pr
is
for
the
kubernetes
sick
or
so
should
we
close
that
and
say,
like
I
don't
understand,
maybe
we
should
ask,
contributes
where
like?
How
is
this.
A
Because
the
existing
pr
is
for
the
old
idea.
C
I
just
I'm
very
much
in
favor
of
that
of
like
saying
look
if
they
want
to
go
into
into
c
closely
like
we're
not
opposed
it
feels
like
it
should
be.
A
better
fit
might
well
be
a
better
fit
somewhere
else,
but
like
contrabax
like
they
have
these
pieces
of
functionality,
how
can
they
get
them
in
the
cube
ads,
kubernetes
github
org
like
how
can
they
get
ci
signal?
How
can
they
get
the
lgtm
block
there
and
make
them
like
actually
do
that
because
it
it?
C
B
B
So
I
think
we
should
push
back
to
make
sure
that
if
they
want
certain
capabilities
that
the
governance
model
of
associated
projects
should
allow
for
folks
to
enable
licensing
and
bot
enablement-
and
it's
fine
to
be
an
associated
project,
that
basically
means
you
adhere
to
the
cla.
B
So
that
way,
if
a
person
commits
to
a
cli,
they
don't
have
to
like
do
five
different
hoops
to
jump
through
for
cncf
associated
projects,
which
is
like
that's
really
like
one
of
the
biggest
benefits
to
be
honest
from
a
commercial
contribution
perspective,
so
we
can
follow
up
there.
I
do
like
the
point
that
I'm
not
opposed
so
long
as
that
they'd
adhere
to
all
the
other
guidelines
and
that
they've
historically
not
been
engaged
at
all.
B
So
you
know
it
would
be
a
change
in
operating
mode.
I
was
taking
a
look
at
their
commit
history
and
it
looks
like
a
couple
of
people
that
go
up
on
once
a
week
or
so
once
every
two
weeks
looking
through
and
doing
some
fixes
here
and
there.
B
So
I
know
I
don't
know
about
how
active
the
community
is
versus
like
it
seems
like
a
couple
of
people
who
use
it
regularly
for
some
things,
but
I'm
not
truly
addressed
to
what
its
broader
implications
are
usages
in
the
wild,
because
I
do
think
that
might
matter
too,
as
well.
A
Okay,
so
as
an
action
night,
I'm
going
to
open
the
discussion
in
the
country
back
swag
channel,
maybe
instead
we
should
use
the
mailing
list.
Any
preference.
A
All
right,
the
second
topic
I
had
is
psa
that
the
deadline
for
the
kubecon
votes
is
this
friday
at
noon.
Pacific
time.
Basically,
for
this
cube
call,
we
have
to
record
the
sessions
in
video
format.
We
have
to
submit
them
to
a
certain
google
drive.
Folder
more
details
can
be
found
in
this
email.
This
is
a
psa
for
the
two
projects
that
we
have
sessions,
for.
I
believe
this
mini
cube,
cube
admin,
cluster.
B
A
Hopefully,
the
maintainers
that
signed
up
for
the
calendars
can
send
these
votes
in
time.
I
guess
we
also
have
to
find
time
to
record
the
signature
session.
C
A
We
we
can
chat
all
this
was
like
ideas,
but
we
have
to
book
some
time
because
otherwise,
yes,
it's
not
going
to
happen.
Otherwise,
I'm
going
to
have
to
just
sit
down
and
record
it
myself.
A
So,
okay,
let's,
let's
record
yeah,
okay,
any
questions
from
the
subproject
representatives
about
this.
D
A
Okay,
so
project
readouts
are
at
kubernetes.
We
don't
have
any
major
updates.
We
are
in
cold
freeze
for
119..
A
The
only
major
thing
that
we
had
to
deal
with
last
week
was
we
had
a
very
peculiar
failure
in
our
container.
This
setup
we're
basically
running
container
dean
docker,
and
we
had
a
very
strange
issue
trying
to
resolve
it
took
some
time
like
we
managed
to
fix
it
like
earlier
this
week.
You
have
more,
you
can
take
a
look
at
this
see
some
of
some
of
the
details.
A
C
Yes,
so
I
think
the
cross
project
thing
of
potential
interest
is
that
we
are,
or
we
have
a
contributor,
adding
support
for
arm,
and
so
I
think
one
of
the
things
that
came
up
is
like
how
we're
gonna
name
the
arm.
C
64
images,
whether
we
put
it
in
the
name,
whether
we
put
it
in
the
tag,
whether
we
do
manifests
well,
I
think
we
broadly
decided
to
follow
what
we
think
is
the
pattern
from
kubernetes,
which
is
to
put
them
in
the
name,
the
architecture
and
the
name
and
use
the
same
tag
and
just
have
a
and
not
necessarily
have
a
a
multi-arch
manifest,
but
nothing
stops
us
doing.
So
I
just
want
to
surface.
B
A
B
B
A
C
Yes,
I
don't
see
lee
here,
so
I'm
just
going
to
give
an
update,
which
is
that
you
know
we
have
the
google
summer
of
code
person
working
on
that
and
she's
doing
a
wonderful
job
and
is
like
maintaining
momentum
and
just
basically
knocking
knocking
ideas
out
as
pr's
as
fast
as
we
can
throw
them
at
her.
There
are
also
some
other
contributors,
but
satoshi
now
has
about
a
month
of
google
sort
of
code
left,
and
we
are,
I
guess,
thinking
about
what
what
thing
should
fill
that
month.
C
So
if
anyone
has
any
ideas
about
things
to
look
into,
any
ideas
are
very
welcome.
We
have
a
couple
of
ideas,
but
I
don't
think
any
of
them
are
sort
of
like
the
next
level
of
of
complexity
or
or
impact.
C
So
any
ideas
are
very
welcome
there
if
you
want
to,
if
you've
been
wanting
to
do
cluster
out
on.
If
you
wanted
to
use
cluster
add-ons,
but
you
felt
some
feature
is
missing
or
some
fundamental
thing
is
missing.
Then
this
is
a
great
time
to
come
up
with
those
ideas
and
issues
on
the
on
the
repo
or
in
our
office
hours
for
bi-weekly
meeting
any
of
those
or
here
any
of
us
are
very
welcome.
Thank
you.
C
It's
three
months
and
we
are
on
the
months
and
we
are
ending.
We
are
in
the
ending
the
second
month
now,
so
it
ends
the
end
of
august.
A
C
Yes,
I
mean-
obviously
yes
I'd,
say
I'd,
say
she's
already
like
achieved
the
goals
of
the
of
the
program.
It's
that
you
know,
she's
a
wonderful
contributor
and
you
know,
has
added
a
lot
of
functionality,
a
lot
of
documentation,
a
lot
of
like
polish,
and
if
there
are
other
things
that
people
want,
then
we
can.
We
have
an
opportunity
to
to
get
a
good
chunk
of
work
done.
A
Yeah,
that's
great.
Google
server
of
code
is
once
you
get
successful
like
one
year
of
success
in
terms
of
students
completing
tasks
next
year.
From
my
past
experience,
I
guess
next
year
more
students
might
be
attracted
to
this.
So.
C
Potentially
and
we
yeah-
we
actually
had
like
a
lot
of
when
it
comes
time
next
next
year.
I
think
we
should
put
up
more
projects.
We
had
a
bunch
of
people
apply,
students
apply
to
the
thing,
but
we
only
had
one
task
slot.
C
I
think
if
we
were
to
apply
with
more
projects
slots,
I
think
they're
called
then
we
would
perhaps
have
more
more
ability
to
pick
more
people.
We
just
had
to
turn
people
away
this
time.
A
C
Yes,
sir,
I
was
saying
like,
although
all
the
sub
projects
should
consider
doing
it
like
based
on
our
experience,
I
wasn't
saying
we
should
have
ten
cluster
add-ons,
several
co-participants,
so
just
more
like
it
seems
like
a
great,
it's
been
a
great
experience
from
my
point
of
view,
at
least
in
terms
of
like
getting
my
contributions
for
the
project
is
great,
so
I
encourage
people
to
think
about
it.
Next
year.
A
Yeah,
hopefully
we
get
more
projects
by
then.
So
any
questions
comments
for
questions.
D
Yeah,
so
we've
shifted
our
milestones
around
a
little
bit,
so
we've
pushed
back
our
1.13
release
to
match
up
with
119,
because
we've
been
having
some
issues
supporting
the
release
candidate
for
it.
So
we're
gonna
take
a
step
back
and
do
that
and
instead
we'll
release
a
point
release
this
week
to
improve
backwards,
compatibility
between
versions
of
mini
q.
That's
been
an
issue,
an
ongoing
issue
for
us,
and
we
have
finally
gotten
a
new
zoom.
D
The
new
chat
will
be
put
in
use
next
monday
and
yeah.
We
have
two.
We
have
two
contacts,
one
on
the
18th
and
one
of
the
19th.
A
That
is
great
that
you
were
able
to
book
a
couple
of
socks
for
crypto.
Hopefully
there's
some
attendance
enough
attendance
in
the
interest.
D
We
don't
we
don't
actually
know
what
happened.
There
was.
I
remember
something
happened
for
all
of
the
sigs
where
we
had
to
change
something
because
we
were
getting
like
zoom
bombed
or
something
and
starting
then
that
then
we
suddenly
our
our
meeting
was
password
protected.
I
I
joined
too
late,
so
I
was
like
I
joined
the
mini
cube
group
like
late
enough
that
the
zoom
call
already
existed.
So
I
had
no
idea
how
it
was
set
up
initially.
D
B
Thomas
reached
out
to
me
a
while
ago,
and
to
be
honest,
like
this,
is
weird
like
I've
been
managing
the
zooms
for
the
whole
sig
for
a
while,
and
I
really
don't
think
that
should
be
a
sig
responsibility.
B
Job
that
should
probably
be
on
contributes
to
be
honest,
so
so
long
as
it's
working
and
you
can
post
that's
all
that
really
matters.
To
be
honest,
if
you
need
anything
from
me,
if
I
want
to
set
up
another
one,
I
reached
out
to
I've
been
having
a
very
long
delayed
response
between
my
thomas
and
myself
and
the
order
of
weeks
where
it's
like.
What
day
did
you
want
it
on
week
passes
by
or
week
and
a
half
it
says
mondays,
and
then
I.
B
I
I
did
have
a
question
with
regards
to
like
there
for
a
while,
like
months
ago,
there
was
this
question:
where
developers
don't
really
care
about
kind
versus
mini
cube
and
they
shouldn't,
and
there
is
this
weird
conflict
between
the
two
subprojects
are
folks
in
a
happy
path.
Nowadays,
where
things
are
trucking
along
on
each
individual
sub-project
or
what's
the
state
of
the
state
now.
D
I
I
haven't,
we
haven't
really
heard
from
from
ben
in
a
while.
As
far
as
I'm
concerned,
it's
fine,
I
think
I
think
kind's
focus
is
very
much
on
like
testing
kubernetes
itself,
which
is
something
we're
not
interested
in
supporting.
D
So
in
that
sense
that
we
have
different
user
bases
and
we
we
are
slowly
moving
away.
So
we,
the
the
the
short
of
it,
is
that
we
use
kind's
base
image
to
support
our
darker
driver
and
we're
slowly
moving
away
from
that
as
our
needs
have
changed,
and
so,
like
our
our
code,
paths
are
diverging
almost
completely
at
this
point,
so
that's
that's
sort
of
where
we're
at
I
don't
I
I
haven't.
We
have
other
many
like
direct
conflict
between
two
projects
in
a
while.
A
D
A
Yeah
in
terms
of
user
bases,
people
are
using
kind
for
all
sorts
of
things,
including
local
website
development.
I've
seen
they
are
also
running
it
in
production,
which
is,
you
know
not
recommended
on
the
side
of,
like
you
know,
running
it
as
an
endpoint
cluster
is
a
bad
idea,
but
yeah.
Definitely
it's
really
hard
for
us
to
resolve,
especially
given
that
this
is
owned
by
a
like
a
separate
sick.
E
Yeah
some
nice
updates
all
pr's
filed
by
justin.
Thank
you
shout
out
to
him.
So
we
have.
We
have
talked
in
the
past
about
you
know
at
the
adm
being
a
cli,
but
then
also
you
know
developing
a
sort
of
higher
level
orchestration.
E
On
top
of
that,
there
is
there's
a
a
long,
a
long
lived
project
that
justin's
been
working
on
that
has
been
used
in
cops,
called
std
manager
with
a
lot
of
that
orchestration
functionality,
and
we
we
had
planned
for
for
a
long
time
to
to
merge
it
in
as
a
you
know,
into
the
stdm
repo
to
to
make
make
development
of
both
the
cli
and
this
piece
easier.
So
that's
that's
going
to
be
merged
in
it's.
E
It's
not
that's
going
to
be
in
the
repo,
but
it's
not
going
to
be.
You
know
it's
not
going
to
be
part
of
the
stadium.
E
You
know
binary
artifact
yet
and
then
also
you
know
something
on
our
roadmap
was,
was
adding
support
for
std
running
in
a
static
pod,
as
opposed
to
or
not
as
opposed
to,
but
in
addition
to
systemd
service,
so
that
initial
support
for
that
is,
is
also
there.
So
those
are
yeah.
Those
are
great
updates.
E
Thank
you
again,
justin,
I'm
I'm
reviewing
those,
and
you
know
if
anybody
is
interested.
Please
please
drop
in
to
the
to
the
pr
and
leave
leave
your
comments.
E
C
Yes,
so
I
opened
a
ticket
with
the
help
desk.
C
This
actually
is
very
related
to
the
thing
we're
talking
about
earlier,
like
the
I
tried
creating
this
std
manager
as
an
associated
project,
which
probably
would
have
meant
we've
had
the
cla
bot
and
I
couldn't
get
it
going
at
the
time,
and
so
now
here
we
are
like
a
year
or
two
later
and
somewhere
in
these
commits
is
hopefully
one
commit
that
doesn't
have
like
the
cla,
and
we
have
to
go
like
chase
that
person
or
decide
that
it's
trivial
but,
like
most
of
these
people,
have
signed
it
like
the
big
committees
have
all
signed
it
like
a
lot
of
them
are
cops
contributors,
but
my
guess
is
that
one
of
the
either
I've
used
the
wrong
email
or
one
of
the
like
one-off
contributors,
has
not
signed
it.
C
I
didn't
I
felt
like
that
might
be
cheating
like
if
that
might
satisfy
the
letter,
but
not
the
spirit
of
the
law
type
thing
the
whether
or
not
we
should
keep
history.
I
wasn't
sure
it
felt
like
it
was
probably
a
good
idea
to
keep
the
history
in
that
then
you
know
you
can
see
the
history
and
this
one
pr
is
monstrous,
but
everything
else
is
is
then
fine
from
then
on?
It's
not
like
it.
It's
not
like
it
lives
on.
A
A
Yeah,
okay:
I
I'm
going
to
subscribe
to
this
if
anybody
else's
interest
just
have
a
look
at
this.
A
This
has
been
a
feature
that
was
requested
for
a
while
I'm
clicking
on
subscribe
and
subscribe
again,
because
this,
where
the
github
notification
arrive
at
my
personal,
like
notification
label,
a
gmail,
otherwise
they're
going
to
arrive
at
the
same
label,
for
that
is
for
a
cdm.