►
From YouTube: Kubernetes 1.12 Release Team Meeting 20180828
A
A
A
A
A
All
right:
it's
three
minutes:
past
I'm
gonna
go
ahead
and
get
started,
I'm
going
to
need
to
be
a
little
bit
short,
a
probably
about
a
half
hour
max,
but
should
be
able
to
get
through
things.
So
as
normal
we've
got
the
Google
Doc
and
that
looks
like
everybody's
dropped,
their
name
in
there.
So
you
see
where
we
are
on
the
agenda.
There's
only
a
couple
of
things.
I
wanted
to
talk
about
in
detail
today.
I
think
we
have
a
lot
of
discussion
yesterday.
A
So
the
one
of
them
we
discussed
yesterday
Windows
container
stable,
G,
a
status
that
basically
was
aggressive
and
then,
as
they
went
out
to
the
other
SIG's,
to
propose
that
they
were
ready,
a
number
of
other
stickers
reviewed
and
found
some
things
where
they
had
some
a
little
bit
more
work
to
do
so.
That's
moving
forward
to
1.13
release,
but
there
is
still
a
lot
of
stuff
that
is
landing
in
1.12,
so
like
from
a
communications
perspective,
release
notes,
documentation,
I
do
think
they're
still
work
there.
A
It's
not
like
that's
completely
falling
away,
but
it'll
be
important
to
make
sure
that
we're
not
calling
out
is
reaching
a
stable,
G
a
milestone.
Specifically,
then
around
CI
signal
and
test
infrastructure.
We
we
still
have
a
whole
bunch
of
failures,
and
these
aren't
reflected
completely
in
the
button
count.
So
there's
a
number
of
things
that
are
failing
in
tests
for
it
and
have
been
for
a
while.
We
we
haven't
had
issues
open
for
them,
so
Mohammed
is
supposed
to
be
opening
some
more
issues.
A
There,
we've
been
trying
to
pull
in
more
sick
folks,
specifically,
look
at
some
specific
symptoms,
not
a
whole
lot
of
traction.
There
are
change,
but
actually
one
thing
landed.
Yesterday,
the
most
of
the
tester
test
cases
happen
to
run
in
the
US
West
Zone
on
GCE,
and
it
was
actually
getting
to
be
oversubscribed
for
the
set
up
that
we
have,
and
then
you
move
some
stuff
over
to
his
a
different
zone
and
suddenly
quite
a
few
of
the
tests
are
passing
now.
So
it
was
just
test
pods
weren't
reliably
starting
up.
A
A
Probably
so
I
feel
like
it's
important
signal
and
then
we'll
probably
end
up
doing
the
same
as
what
we
did
in
1.11
looking
at
them
case-by-case.
If
we
still
have
failures
and
we
can
convince
ourselves
for
the
specific
failure
that
the
code
in
question,
we
know
we
have
covered
elsewhere
and
it's
working,
then
we
probably
won't
block
on
the
gke
failure,
but
we'll
have
to
play
the
fight
here
as
we
go
forward.
A
What
else
in
here
so
I
kind
of
mentioned,
Doc's
things
falling
away.
We
we
have
had
kind
of
an
aspirational
goal
of
getting
PR
is
open
for
Docs
and
we
were
hoping
to
see
a
lot
of
them
as
placeholders
in
it
to
give
us
a
sense
of
what
work
was
coming,
but
does
that
kind
of
admit
it,
even
though,
like
37
are
missing
their
PR?
Is
that
that's
kind
of
good?
A
In
a
way
we
did
get
a
good
want
it
to
open
things
up
and
give
us
a
heads
up,
and
then
the
other
thing
that
he
mentioned
was
that
this
is
kind
of
a
leading
indicator.
So
we
have
this
big
list
of
features
that
were
proposed
if
somebody's
not
willing
to
commit
to
writing
their
documentation.
That's
probably
that's
potentially
an
early
sign
that
their
code
isn't
contracted
completion.
A
So
we
we
already
knew
that
our
list
of
features
was
large
and
some
of
our
kind
of
expecting
to
fall
away
because
they're
aggressive
to
reach
this
goal,
but
that
that
gap
in
the
docs
may
be
a
sign
of
that.
So
these
will
be
ones
that
we
want
to
start
watching
over
the
coming
weeks.
Are
they
are
they
rushing
in
right
at
code?
Freeze
are
they
low
quality
than
as
they
do
that?
This
is
just
a
something
that
we
can
use
as
a
heads
up.
A
So
that's
a
kind
of
a
handy
thing,
I
guess
to
compare
Caitlin
mentioned
that
the
the
communications
team
is
meeting
with
the
features
folks
to
start
assessing
that
feature
status
and
preparing
for
the
blog
post
Docs
in
communication.
Anything
anybody
wants
to
save
their
and
any
questions
so
far.
I'm
just
kind
of
rambling.
C
A
Something
that
I'm
not
sure
if
this
is
called
out
in
the
notes.
Yesterday,
oh
yeah,
it
is
at
the
end
so
I'll
get
to
it
still
couple
process
changes.
We've
talked
previously
about
the
status
approved
from
a
stone.
The
Lazy
consensus
date
went
by
yesterday
for
that
and
basically
no
objections,
one
question
and
maybe
half
dozen
thumbs
up
on
it.
A
So
we're
gonna
go
ahead
and
do
that
and
I've
got
a
couple
bits
of
work
to
just
clean
up
some
some
documentation
and
things
around
that
and
the
the
team
on
github
the
membership
of
it.
But
is
we're
talking
about
that
yesterday?
It
came
up.
This
is
a
good
time
for
everybody
on
the
team
to
look
through
the
release,
team
role,
handbooks,
and
that
was
one
of
the
the
main
things
that
I
wanted.
To
mention
to
you.
B
A
I'm
putting
a
link
in
the
minutes,
there
glance
through
those
for
your
role
and
if
there's
anything
that
you
notice,
that's
maybe
changed
during
the
cycle
relative
to
what
you've
observed,
make
a
PR,
update
the
documentation
or
start
a
conversation,
and
we
can
keep
these
dogs
upstage
and
then
also
since
the
set
of
folks
today
are
shadows.
This
is
a
chance
also
to
think
about
talking
to
your
lease
lead.
If
you're
interested
in
taking
on
the
lead
role,
it's
a
possibility.
A
B
A
You're
interested
in
taking
on
a
role
as
a
lead
or
another
shadow
in
the
release
team
role
know
somebody
who
is
we're
kind
of
getting
into
the
phase
where
a
little
bit
of
time
should
be
spent
on
thinking
about
that
succession
and
there's
opportunities.
If
you
all
want
to
do
more
and
then
I
want
to
give
a
shout
out
to
young
Li
and
ask
if
he
has
anything
he
wants
to
say
today,
so
our
branch
manager
has
been
on
vacation
since
Thursday
and
the
two
shadows
on
us,
coral
and.
B
A
A
Yeah
so
far,
everything
looks
good
and
I
really
want
to.
Thank
you
because
one
of
our
questions
has
been
how
and
this
is
like
across
the
kubernetes
community.
How
can
we
do
more
to
get
involvement
from
folks
in
China,
and
we
know
that
there
are
certain
issues
like
the
usage
of
certain
Google
tools,
make
it
harder
for
people,
and
it's
been
really
awesome
to
see
you
engage
and
you
also
be
able
to
successfully
run
some
of
these
things
so
that
that
experimentation
is
really
helpful
to
us
and
in
any
suggestions.
B
B
Think
not
everyone
here
has
just
access,
so
I
have
just
read
it,
so
the
business
model
are
related
to
open
source
is
not
measure
in
China,
so
many
companies,
participation
in
open
source
project
is
limited
by
the
profit
pressure,
so
so
that
employees
are
not
encouraged
to
participate
in
open
source
projects
during
the
working
hours.
So
I
think
this
is.
A
Here
is
to
see
how
that
evolves,
I
seen
similar
here
and
in
other
locations
to
some
extent
it's
sort
of
a
pendulum
that
goes
back
and
forth.
There's
big
interest
in
open
source
and
then
maybe
a
little
more
focus
of
profit
on
times
and
often
then,
companies
that
mature
a
little
bit
longer
start
to
realize
that
their
success
they
owe
partly
to
this
open
source
and
they
do
need
to
to
give
back
to
ensure
its
vitality.
A
But
but
then
maybe
there
are
things
that
some
of
us
existing
in
the
community
can
do
to
to
help
them,
whether
it's
advocating
with
from
business
to
business
or
or
finding
things
that
are,
even
though
some
of
the
subtle
things
like
having
github
issues
that
are
labeled
as
good
first
issue,
because
they're
a
little
bit
more
bite-sized
that
somebody
could,
if
they
were
trying
to
and
they're,
not
quite
over
time,
they
can
still
squeeze
in
and
make
meaningful
contribution
and
I
start
to
get
engaged.
I'm
interesting
to
see
how
it
evolves.
A
Looking
at
this,
the
notes
from
yesterday
on
cherry-picks,
though
this
is
actually
maybe
a
good
transition.
This
trim
pick
process
so
part
of
branch
management.
At
the
end,
just
before
we
release
is
going
to
be
cherry,
picking
commits
off
of
the
master
branch
and
into
the
release
branch
and
that
terrific
process
is
kind
of
clunky.
So
there's
some
work
underway
right
now
to
do
some
little
bits
of
tweaks
to
improve
it.
A
If
you
see
things
there
that
you'd
like
to
improve
but
I
definitely
encourage
you
to
give
it
a
try
and
I've
reached
out
to
some
of
the
and
suggested
the
wind
also
who'd
been
involved
in
some
of
it
suggested
that
she
reach
out
to
some
of
the
prior
branch
managers
and
patch
managers
for
ideas.
If
we
get
a
little
bit
more
documented
in
the
issues,
maybe
we
can
figure
out
some
so
small
tweaks
that
we
could
changed
to
tools
or
processes
and
then
the
documentation
as
well.
A
A
We
get
some
some
major
codes
at
lands
and
those
always
have
the
the
risk
of
destabilizing
things
so
having
a
good
CI
signal
than
is
important,
so
we
we
notice
and
understand
the
particular
commit.
The
particular
feature
was
the
one
that
destabilized,
though
CI
signal
is
really
going
to
be.
The
focus
for
the
next
week
or
two
I
would
expect
for
a
number
of
us
and
then
docks
release,
nodes
communications.
All
of
that
is
ramping
up
as
well,
and
if
that
happens,
in
parallel.
A
A
A
Okay,
then,
the
final
thing
is
a
reminder.
Our
meeting
cadence
is
changing
so
this
weekend
next
week
we
go
to
meeting
Monday
Wednesday
Friday
for
the
the
core
team
I'm,
keeping
the
the
Tuesday
meeting
as
is,
but
have
added
a
Wednesday,
Friday
and
then
Monday
coming
is
a
u.s.
holiday,
so
I
canceled.
That
meeting.
A
And
hopefully,
that's
all
reflected
on
your
calendars,
but
free
y'all
come
in
on
Tuesday.
It
won't
matter
too
much
and
I
would
encourage
you
if
you
can
and
then
in.
If
you
also
I
guess
I
would
say
if
the
Wednesday
Friday
times
are
particularly
bad
and
you
would
like
to
attend.
Let
me
know
and
I'd
be
willing
to
see
about
Quisling
the
time
a
little
bit,
but
over
the
next
month,
I'd
encourage
you
a
shadows
to
extend
a
few
of
those
additional
readings,
because
the
the
pace
of
things
is
gonna
wrap
up
a
bit.