►
Description
We have PUBLIC and RECORDED weekly video meetings every Thursday at 10am US Pacific Time.
https://docs.google.com/document/d/1VQDIAB0OqiSjIHI8AWMvSdceWhnz56jNpZrLs6o7NJY
A
A
But
in
short,
it's
essentially
one
of
our
key
strategies
for
making
the
release
process
better
and
ergo
the
lives
of
all
the
contributors
and
consumers
are
Korean.
So
it's
a
pretty
big
deal
and
it's
a
great
time
for
us
to
talk
about
what
went
well,
what
didn't
go
well
and
most
excitingly
what
we
plan
to
change
for
one
pot,
noggin!
So
now
further
ado,
let's
pass
the
baton
to
Sara
to
give
us
those
results
and
if
anybody's
got
a
drum
for
a
drum
roll,
please
convinced.
B
Okay,
we
can
do
it
without
the
drum
roll.
You
can
all
hear
me.
Yes,
fantastic
I've
been
having
troubles
with
Mike's,
okay,
no
I'm,
really
just
trying
to
elongate
this
event.
So
I
want
to
give
a
moment
of
context
for
this.
The
blog
post
with
this
information
is
going
to
be
posted
at
10:30
on
the
kubernetes
blog.
So
if
you
are
interested
in
tweeting
about
it,
I
would
I
would
be
sure
that
you
or
I
would
ask
that
you
put
your
tweets
or
reference
your
tweets
at
some
point
to
that
blog
post
as
well.
B
We
have
spent
the
last
two
years
we're
working
toward
a
more
formalized
community
governance,
and
this
is
the
biggest
step
that
we
have
taken
so
far.
So
I'm
super
excited
to
announce
the
winners
for
the
2017
election
for
the
kubernetes
steering
committee.
Now,
as
you
remember,
there
are
rules
around
how
many
positions
can
be
held
by
a
particular
company,
so
I'm
going
to
go
through
the
order
in
which
our
top
our
top
vote
make
Rotech
takers
vote.
B
Our
top
voted
candidates
were,
and
then
one
of
them
unfortunately,
will
not
be
able
to
take
a
seat
because
of
that
company
limit.
All
of
the
data
from
this
will
be
posted
into
the
github
repository
in
community
community
election
2017.
So
people
can
take
a
look
at
it
there
as
well
and
as
asked
and
answer
any
questions
as
you
run
across
them,
so
the
people
who
have
been
awarded
three
year
or
sorry
what
two
years
seats.
B
So
these
will
be
the
2018
and
19
seats
are
Derek
Carr,
Michelle,
Nirali
and
Phillip
Bullock
wha
Philip,
which
Rock
so
many
many
congratulations
to
them.
The
person
who
was
on
the
cutoff
between
the
two
year
seats
and
the
one
year
seat
was
Michael
Rubin,
who,
unfortunately,
because
of
the
minimum
nor
maximum
number
of
company
seats
allowed,
will
not
be
able
to
take
a
seat
because
that
we
can
only
have
one
additional
Googler
be
added
to
the
community
steering
committee.
So
Michael
was
voted
in
but
is
not
unable
to
take
a
seat.
B
The
people
who
are
taking
one-year
seats
are
Timothy,
st.
Claire,
Quinton,
Huell
and
Aaron
tricking
Burke,
so
I
want
to
say
many
many
many
thanks
and
everybody's
clapping
on
mute,
which
is
really
quite
amusing.
Many
many
thanks
to
all
of
you
who
voted
all
of
you
who
have
contributed
all
of
you
who
are
engaged
in
this
community
and
have
given
us
as
as
given
us
as
a
community,
the
opportunity
to
grow,
and
then
please,
please,
please
offer
your
support
to
these
people.
B
B
Everybody
silent
as
I
said:
Oh
read
the
names
again.
Okay,
so
the
blog
post
will
go
out
in
as
I
said
at
10:30,
but
we've
got
Derek
Carr,
Michelle,
Nirali
and
Phillip,
which
rock
for
two
year
terms:
Timothy
st.
Claire,
Quentin,
cool
and
Aaron
Creek
and
Berg
for
one
year
terms,
Creek
and
burger.
Sorry,
you
have
an
extra
earth.
B
But
yes,
please
everyone,
everyone,
everyone
support,
both
your
new
steering
committee
numbers
and
your
bootstraps
steering
committee
members
who
will
be
helping
this
next
two
years
of
governance,
evolution,
okay,
saying
she's,
not
sure
she
got
the
names
I'll
do
them
one
more
time,
but
as
I
said,
this
will
be
published
on
the
internet
for
all
posterity
in
22
minutes,
Derek,
Carr,
Michelle,
Murali,
Philip,
Witt,
Rock
two-year
terms
Timothy
st.
Clair.
That
would
be
Timothy
stay
clear
of
hep
geo,
not
Timothy.
B
B
B
E
All
right
well,
I
guess
we
will
dip
ourselves
right
into
the
1.8
retrospective
thanks
Sarah
for
the
board
winners.
Well,
community
committee
members,
thanks
so
hi
everybody
in
Paris
today
we're
going
to
talk
about
the
1.8
retrospective.
We
have
tons
of
material
to
cover.
This
will
be
broken
down
into
two
sessions
today
and
tomorrow
tomorrow,
at
the
same
time,
same
zoom,
and
from
before
we
get
started
really
quick.
We
do
have
some
guidelines
that
we
need
to
discuss
since
there's
again
since
there's
a
lot
of
material
here.
E
One
please
mute
when
you're,
not
speaking,
we
will
ask
for
tons
of
comments
and-
and
you
know
how
you
feel
about
certain
things,
so
there
will
be
a
lot
of
people
talking
at
the
same
time.
So
please
mute
when
you're
not
also,
please
be
respectful,
keep
in
mind
that
keep
in
mind
that
you
should
be
talking
about
processes
and
tooling
that
possibly
didn't
work
out,
not
necessarily
persons
or
or
people,
and
also
please
be
mindful
time
as
as
we're
going
here.
E
If
there's
any
concerns
at
all,
please
private
message
me
and
we'll
get
those
taken
care
of
so
now
on
to
the
show
really
quickly,
though,
I
just
wanted
to
give
a
huge
shout
out
to
Jace
the
release
team
lead.
He
did
an
awesome
job,
I've
heard
from
so
many
people
about
a
wonderful
work
that
he
did
so
if
we
can
also
give
him
a
clap
as
we
are
all
muted,
that
would
be
awesome
and
everybody
else
to
that.
Did
such
hard
work
for
this
release.
E
Now,
let's
go
to
right
into
it
and
we'll
talk
about
follow-up
from
the
last
one
point:
seven
retro
did
we
do.
We
said
we
did
and
it
looks
like
for
the
most
part,
that's
a
pretty
accurate
statement.
We're
not
gonna
go
line
by
line
with
these,
but
just
some
really
quick
things
that
that
we
did
that
we're
really
well.
We
did
the
code
freeze
across
all
communities,
repos
that
depended
on
the
main
repo.
E
E
C
So
my
role
on
the
release
team
was
sort
of
like
bug
triage,
but
also
as
a
representative
of
state
testing,
I
was
trying
to
make
sure
that
testing
for
structure
looked
good
and
test
grid
was
tremendously
helpful.
For
me,
in
this
regard,
the
kubernetes
pre
submits
test
grid
dashboard,
which
was
added
during
this
release
cycle.
Let
me
quickly
look
at
like
are
all
of
the
pre
submit
jobs
working
generally
well
across
all
PRS
I.
C
Couldn't
expect
them
to
be
passing
100%
at
a
time
because
some
people
check
in
broken
code
and
their
pull
requests,
but
it
was
a
really
really
quick,
heads-up
and
also
just
having
a
little
icon
next
to
each
job.
That
said,
you
know
passing
in
flaky
or
failed
gave
me
a
quick
heads-up
on
which
jobs
I
wanted
to
focus
on
and
then
further.
C
D
E
A
Gosh,
so
this
was
this:
was
Eric
fada,
stepping
up
and
organizing
the
tech
crew
and
into
daily
stand-ups,
so
I,
don't
know
how
how
much
people
know,
but
Google
Google
testing
is
putting
in
a
tremendous
amount
of
effort
on
this
stuff.
All
the
time
and
Eric
organized
a
stand
up,
because
I
was
having
a
hard
time
wrapping
my
head
around
all
the
different
threads
that
were
taking
place
in
a
given
moment,
and
it
was
really
informative
to
be
there
and
to
hear
the
stand-ups
ands
just
the
fact
that
they
were
willing
to
organize
that
ceremony.
A
Basically
because
I
asked
was
a
amazing,
truly
great
and
just
truly
in
the
spirit
of
community.
So
also
I
wanted
to
say
that
the
release
team,
specifically
I
thought
worked
really
well
together
and
our
meetings
felt
productive
to
me.
Oftentimes
meetings
can
feel
like
a
slog
and
ours.
Never
really
did
people
were
up
bead
and
focused
on
solutioning,
and
it
was.
It
was
good.
They
were
very
effective,
so
I
was
thankful
for
that.
Because
time
is
insurance
apply
for
everyone.
A
I
thought
extending
the
code.
Freeze
was
great
because
it
allowed
us
to
get
a
lot
more
time
to
dig
into
some
of
the
technical
that
around
upgrade
testing
whatnot
and
hopefully
also
laid
some
groundwork.
That
1.9
will
be
better,
and
that
was
it
was
good
too,
because,
amazingly
the
the
schedule
didn't
slip
because
of
any
of
that
activity.
Nor
does
it
seem
like
the
velocity
of
1.9
was
impacted,
so
that
was
that
was
a
really
nice
opportunity
and
it
was
nice
to
see
a
tweet.
A
A
So
sick
themes
superseding
the
release,
themes,
I
thought
was
great
and
allowed
us
to
expose
the
governance
model
to
our
community.
That
is,
is
good
for
people
to
start,
seeing
that
the
real
power
is
in
its
sakes
and
not
necessarily
some
some
collection
of
people
in
a
room
somewhere,
that's
making
decisions
and
secret.
It's
really.
Everything
is
transparent
and
true
to
the
community
spirit.
So
that
was
good,
not
that
there
were
groups
of
people
in
secret,
but
there's
a
opacity
has
been
a
problem
in
other
some
other
open-source
projects
in
the
past.
A
Lots
and
lots
of
people
were
regulars
in
the
release
process,
even
though
not
on
the
formal
release
team.
These
people
Jordan
mehdi
lukas
young
and
a
lot
of
others
holy-moly
every
day.
I
would
go
in
and
and
get
greeted
by
people
and
get
the
rundown
of
what
had
me
happened,
the
prior
night
or
or
what
was
going
on,
and
people
would
raise
hey
this.
This
PR
needs
tension
or
that
whatnot
it
was
just
so
amazing
to
see
people
stepping
up
and
and
just
pitching
in
all
the
time
it
was
like.
How
is
this
happening?
A
This
is
so
amazing.
So
again,
super
awesome,
community
and
I
thought
adding
dims
to
the
release
team,
Aaron
Aaron
Cohn
oversaw
that,
because
James
was
covering
a
lot
of
the
bug,
triage
awana
great
right,
I
mean
this
is
this
is
a
good
recognition
for
somebody.
Who's
spent
countless
hours
and
trying
to
make
things
better.
This
feels
like
a
huge
monologue.
We
did
a
good
job
balancing
the
needs
of
expedience
to
the
needs
of
stability.
A
I
I
feel
really
proud
about
this,
because
I,
we
I
think
we
had
really
rich
discussions
around
all
of
the
things
that
were
balancing
community
needs
versus
the
needs
of
the
release
and,
and
so
stability
was
never
cast
aside
in
terms
of
expediency
and
I.
Think
that's
a
huge
thing
for
us
to
continue
doing
and
hopefully
release.
Teams
of
the
future
will
continue
that
that
pattern,
because
I
think
it's
important
for
us
to
have
those
discussions.
A
Even
though
I
tell
you
at
times
they
were
difficult
and
there
were
times
her
I
actually
had
to
as
to
release
lead
play.
The
the
devil's
advocate
for
an
alternative
viewpoint
that
I
didn't
even
agree
with,
because
I
felt,
like
those
opinions
needed
to
be
in
the
room,
so
I
think
that
that's
that's
super
important
and
hopefully
something
will
see.
Continued
Erin
looks
like
you've
got
a
couple
other
things
and
then
off
tomorrow,
yeah.
C
Sure
so
I
I
felt,
like
the
sake
release
lacked
channel,
was
pretty
active.
I,
don't
know
for
me.
I
do
like
a
constant
stream
of
stuff
happening
in
the
slack
channel
is
a
sign
that
people
are
engaged
and
stuff
is
actually
happening.
Long
periods
of
silence
can
be
a
little
unsettling
at
times
so
and
I
I
know
this
is
gonna
sound
stupid,
but,
like
seriously,
all
of
the
emojis
and
slack
reactions
that
were
being
thrown
around
health.
C
A
lot
to
let
people
know
that,
like
yes,
I
saw
your
message
or
yes,
I'm
doing
a
thing
or
yes,
I
think
burritos
are
delicious,
like
that
was
all
really
really
helpful
and
it
the
activity
there
made
it
like.
If
you
had
a
question
about
what
was
going
on
with
the
released,
you
knew
that
could
be
a
place
to
go,
find
some
engaged
people
and
figure
out
where
to
go
from
there.
C
Helpful
and
sort
of
back
to
Jason's
point
about
the
the
daily
testing
stand
up
like
I
I,
just
wanna
give
a
shout
out
so
the
tests
you
have
and
proud
team
at
Google
and
the
folks
who
are
in
charge
of
testing
for
inside
testing,
because
we
try
to
really
be
very
mindful
of
the
release
this
time.
You
know,
there's
this
big
there's
a
lot
of
attention
paid
to
code
freeze
and
making
sure
that
the
code
itself
is
stable.
C
We
thought
maybe
it
would
be
a
good
idea
to
make
sure
the
tests
infrastructure
is
also
stable
and
that,
like
maybe
we
don't
try
changing
the
test
infrastructure.
Well,
the
code
isn't
changing
all
that
much.
So
we
really
try
to
evaluate
you
and
everything
we
were
doing
with
like.
How
much
is
this
going
to
change?
What
is
familiar
to
the
release
team?
What
is
the
potential
risk
of
deploying
this
change?
C
Can
it
easily
be
rolled
back
if
we
find
out
that
something
is
broken
and
I'm
sure
I'm
biased
since
I'm,
like
Kalitta's
a
testing
but
I,
think
we
did
a
decent
job
of
keeping
things
relatively
stable
compared
to
last
time?
There's
some
big
changes
for
planning
and
rolling
out
on
1/9
that
we
could
have.
We
could
have
started
rolling
out
now,
but
we
decided
to
you
know,
hold
off
to
make
sure
that
the
release
actually
got
cut
and
tests
passed
as
much
as
possible.
So
with
that
I
will
hand
off
to
you.
F
F
So
hopefully
that'll
just
encourage
both
issue
owners
and
the
SIG's
that
are
associated
with
those
issues
to
take
action
in
a
timely
fashion,
so
that
when
we
come
to
the
end
of
the
release,
development
cycle,
we'll
actually
be
able
to
just
focus
on
things
that
are
actually
actionable
versus
just
flailing
at
issues
that
we
have
no
idea
what
their
actual
and
Cortis.
That's
all.
E
C
Okay,
so
the
the
biggest
one
for
me
above
anything
else,
was
that
we
have
these
dashboards
in
test
grid
and
they
have
the
word
blocking
in
them
almost
as,
if
we're
going
to
like
not
cut
a
release
for
going
to
block
the
release
until
all
of
the
tests
and
those
dashboards
pass.
And
what
did
we
do?
We
cut
builds
when
the
tests
were
failing,
we
cut
alpha,
builds
when
the
tests
were
failing,
nobody
seems
to
care
about.
C
Alpha
builds
and
that's
fine,
that's
a
separate
topic,
but
we
also
caught
our
beta
builds
when
tests
were
failing,
and
then
we
had
a
last-minute
highly
engaging
90
minute
conversation
over
whether
or
not
it
was
appropriate
to
cut
the
final
release
and
ship
it
out
the
door,
because
the
test
was
failing
and
it's
a
test
that
had
been
failing
for
the
entirety
of
the
release.
Now,
like
we
sort
of
like
started
to,
we
have
started
to
coalesce
around
humans.
C
Looking
at
these
things,
but
like
we
also
have
our
tools
looking
at
the
release
blocking
test
lists
and
we
actually
had
to
change
the
amount
of
time
that
Anagha
waits
or
attempts
to
find
a
build
that
actually
passes
all
of
the
tests
it
pulling
into
basically
an
infinite
loop,
because
I
couldn't
find
a
build
of
kubernetes
that
passes
all
of
the
tests
that
are
supposed
to
define
that
kubernetes
is
working.
I
think
we
seriously
need
to
actually
use
those
tests,
as
blocking
and
I
think
we
need
to
prune
the
blocking
lists
down
a
little
bit.
C
We
sort
of
set
out
one
a
saying
like
okay,
this
time
for
real
we're,
gonna
pay
attention
to
these
tests.
They
have
the
word
blocking
in
them,
we're
totally
gonna,
do
it
and
then
we
didn't,
we
didn't
do
it
early
on.
So
to
me,
it
was
a
red
flag
that
we
started
ignoring
us
for
the
Alpha.
But
again
nobody
cares.
We
started
ignoring
this
for
beta,
because
well
I
guess
we
had
to
keep
marching
on
schedule
like
not
cutting
a
bill
and
actually
blocking
the
release.
C
I
thought
was
supposed
to
be
the
steak
that
you
know
having
blocking
tests.
Failing
was
supposed
to
provide
us,
so
I
feel
like
we
ignored
one
of
our
tools,
which
led
us
to
situation
we're
yet
again.
Humans
have
to
interpret
test
results
and
it
takes
a
lot
longer.
It
increases
the
feedback
cycle
and
it
reduces
confidence
and
the
stability
of
the
build.
C
We
also
found
that,
apparently
we
don't
just
you
just
generally
don't
exercise
and
I'll
go
a
whole
lot,
so
we
ran
into
another
like
weird
corner
case-
that's
probably
covered
by
tribal
knowledge,
but
apparently
when
cutting
the
first
data
I'm.
Sorry
when
cutting
beta-1
Onaga
was
now
set
up
to
look
at
tests
in
the
release.
C
1/8
board
instead
of
the
master
board
and
like
not
all
of
the
tests
had
been
ported
over
in
time,
and
so
it
seems
like
there
was
a
little
bit
of
coordination
that
could
have
helped
to
making
sure
all
the
tests
were
ported
over
from
master
over
its
1/8
there's
a
bit
of
tribal
knowledge
in
what
tests
actually
belong
in
a
release,
blocking
a
release
board
versus
the
master
board.
If
I
look
at
the
tests
in
release,
1/8
vs.
release,
1/7
they're
slightly
different
and
I,
don't
know
why.
C
Let's
see
next
item
I've
seen
this
happen,
pretty
consistently
every
release
the
release
team
doesn't
seem
to
know
how
and
why
code
gets
merged.
This
time
it
seemed
like
nobody
really
understood
the
use
of
the
approved
for
a
milestone
label
or
exactly
how
it
worked.
I
have
faith
that
we're
gonna
get
significantly
better
at
this.
C
Now
that
we
have
the
milestone
Munzer
in
place-
and
I
I
regret
not
being
johnny-on-the-spot
with
the
fact
that
if
you
go
to
the
submit
queue
and
click
on
a
tab
called
info
right,
there
is
the
set
of
criteria
that
this
makes.
You
use
this
to
decide
if
a
pull
request
is
ready
to
merge,
there's
also
a
tab
that
explains
the
sort
order
for
what
labels
and
milestones
and
due
dates
and
stuff
to
make.
C
So
I
feel
like
we're
in
a
situation
where,
if
we
can
socialize
how
to
point
to
live
Docs
that
are
updated
as
the
code
changes
we'll
be
in
a
slightly
better
situation,
but
I
almost
questioned
whether
or
not
the
release
team
as
a
whole
needs
to
rally
at
the
beginning
of
the
release
process,
or
certainly
before
code
freeze
is
enacted
to
understand.
Like
do,
we
understand
how
what
are
the
criteria
for
code
getting
merged
and
all
that
stuff,
because
we
all
seem
to
have
to
reverse
engineer
the
mechanics
of
this
whole
process.
C
C
That's
why
I've
attended
so
many
releases
is
to
try
and
help
carry
that
forward,
but
sometimes,
like
my
knowledge,
might
be
based
off
of
a
release
that
happened
three
quarters
ago
and
it
might
be
antiquated
and
that
might
cause
us
to
lurk
around
as
we
try
to
figure
out
the
way
things
are
today
versus
the
way
things
used
to
be
versus
the
way
things
should
be
versus
the
way
we
want
them
to
be
and
just
like.
Having
that,
all
documented
in
a
place
would
be
helpful.
C
C
So,
like
I
said,
I
feel
like
increased
socialization
of
the
existing
places
that
are
updated,
live
as
we
like.
If
there
are
places
that
are
in
charge
of
merging
code,
that
don't
have
documentation
sitting
right
next
to
them
that
explain
how
they
work,
we
should
make
sure
such
pages
exist
and
for
the
things
that
do
have
that
we
should
socialize
those
links
and
make
them
more
applicable
and
I
think
we're
we
aren't
getting
there.
But
that
sounds
like
a
good
action
item
to
just
like
enforce
that.
C
There's
probably
some
documentation
and
the
community
repo
that
could
be
updated
and
I
probably
have
some
takeaways
to
add
to
the
testing
for
repo.
For
like,
for
example,
you
can't
deep
link
to
the
the
sort
order
and
the
merch
requirements
page
you
have
to
like
tell
somebody
to
go
click
on
the
info
thing
and
then
click
on
this
other
thing.
That's
a
little
confusing.
C
The
things
of
that
nature.
So
what
we
ended
up
doing
was
going
in
just
like
removing
those
tests
from
the
older
release,
which
kind
of
reduces
the
fidelity
of
the
earlier
releases
testing
against
the
earlier
release.
But
hey
got
it
passing
in
the
upgrade
test.
So
that's
cool,
and
maybe
that's
maybe
that's
appropriate
for
for
where
we're
at.
But
it
seems
like
there's
like
a
mechanism
missing
to
gate
what
tests
should
be
run
against
which
version
of
kubernetes
and
I
think
Tim
st.
Clair
even
opened
an
issue
about
this
somewhere,
but
I.
C
C
Is
this
McHugh
blockage
around
basal
failing?
That
was
just
something
that
happened
completely
outside
of
our
control
and
it
was
something
we
were
responsive
enough
to
fix.
Maybe
we
could
have
fixed
it
a
little
bit
sooner,
but
just
an
example
of
sometimes
the
Siddiqui
breaks.
It
just
happens:
okay,
I
guess
to
Jason-
or
at
this
point.
A
That
was
all
really
great
here
and
thank
you.
Yeah
I
would
+1
everything.
You
said
it
was
those
phone,
so
so
this
the
next
couple
of
bullet
points
are
really
I.
Think
indicative
of
where
the
project
is
at
in
terms
of
our
governance
model,
because
it's
shifting
I
mean
obviously
we
elected
a
steering
committee
now
and
that
we're
looking
at
ways
in
which
we
can
arbitrate
decision
making
at
a
high
level
and.
A
A
Is
that
what
gets
into
career
Nettie's
is
not
going
to
be
quote-unquote
a
feature,
but
more
about
increments,
of
value
that
we're
delivering,
because
value
can
be
defined
as
something
that's
user
facing
or
something
that's
internal
or
something
that
is
is
going
to
increase,
contributor
experience,
velocity
or
whatever.
That
is,
but
essentially
making
it
less.
Like
people
think
of
a
feature,
as
in
hey
I
can
do
this
thing
in
the
UI
or
I.
Can
you
know
now
the
play?
That's
gonna
work
load
it.
A
It's
less
about
that
and
about
how
do
we
enrich
the
ecosystem
in
a
way
that
it
works,
and
so
what
I'd
love
to
see
is
more
synchronization
between
product
management
and
and
what's
actually
delivered
in
the
release,
because
ultimately
sakes
are
driving
that
and
not
the
the
PM
group
necessarily
so
the
PM
group,
in
my
opinion,
should
pivot
toward
more
of
a
collecting.
You
know
valuable
feedback
from
the
user
community
in
and
how
are
we
serving
the
needs
of
all
of
our
various
appliances?
A
You
know
internally,
extremely
the
companies
that
sponsor
a
lot
of
development,
whatnot,
so
I
think.
That's
that
that's
a
big
takeaway
from
this
for
me,
you
are!
Is
there
anything
you
want
on
that
because
you,
you
know
you're
in
CPM
and-
and
you
were
along
with
us
and
you
managed
release
features
so
I'd
be
very
interested
to
hear
what
you
have
to
say.
There
yeah.
G
First
of
all,
I'd
like
to
lose
taken
to
Everson,
you
said
first,
the
consent
decree
folder
is
before
the
release
process
has
been
started,
but
it
was
really.
It
was
like
deep
Surma.
Not
so
many
people
were
in
line
and
we
have
put
most
puzzles
on
hold
until
the
release
has
been
cut.
So
I'm
sure
we've
started
are
working
on
enhancing
our
job,
to
make
it
more
efficient
in
terms
of
the
product
management,
and
we
actually
had
several
discussions.
G
How
can
we
enhance
that
more
from
the
purely
technical
value,
but
also
to
the
persistent
failure
and,
and
so
on,
speaking
about
the
issues
that
we
had
exactly
in
this
release
cycle,
it
was
mostly
about
our
total,
a
synchronous
work
in
terms
of
the
project
management
side
in
terms
of
the
features
tracking
officious
management
ways,
the
technical
implementation
administer.
So
we
even
we
have
even
faced
a
few
features
that
have
been
submitted
and
committed
in
their
last
few
days
before
the
release
has
been
cut.
G
G
And
next
next
item,
so
if
we
are
done
with
this
one
Jace,
the
Nexus
yeah
I,
am
speaking
about
the
action
item,
so
that
did
what
should
I
say
it
like
a
few
seconds
ago.
Did
you
go
away
yeah,
like
primary
action
item
for
for
the
next
release,
so
to
elaborate
deeper
on?
Well,
how
do
we
do
the
product
management
stuff?
How
do
we
do
the
high
level
product
management
of
kubernetes
is
with
implementing
some
functionality
into
it?
G
A
Yeah
and
I'm
I've
told
a
few
people
is
but
I'm
working
on
a
proposal
to
to
how
do
we
enhance
that
that
connection
between
six
and
product
management
too
so
stay
tuned
on
that
okay
cool?
So
then,
I'm
gonna
go
to
move
on
to
the
next
bullet
point,
which
is
the
missing
consensus
model
around
who's
in
power
to
make
the
final
call
on
the
go
no-go.
A
As
Aaron
said,
there
was
a
90
minute
discussion
on
this.
Basically,
you
know
it
was
kind
of
funny
because
we
were
dealing
like
everybody
is
in
agreement
on
the
course
of
action,
and
then
somebody
added
to
do
somebody
got
added
to
the
meeting
sort
of
late
and
they
had
a
diametrically
opposed
viewpoint
on
on
whether
we
should
release
or
not,
and
so
it
it
kicked
up.
What
I
thought
was
an
absolutely
phenomenal
discussion
about
what
what
are
the?
A
It
was
a
group
decision,
but
essentially
I,
think
I
was
on
the
hook
to
to
make
the
call
or
not-
or
at
least
it,
to
adjudicate
all
the
various
forces
to
try
and
come
with
a
consensus
answer
which
I
think
we
did,
but
there
was
I
couldn't
go
and
look
at
a
policy
that
says
you
shall
not
put
out
communities
with
blocking
test
red
I
mean
there,
there's
just
nothing
like
that,
so
it
was
a
very
you
know
to
Aaron's
point.
This
is
a
tribal
knowledge
thing.
So
as
a
release
leader
I
did
the
best.
A
I
could
and
I
tried
to
be
take
the
the
owners,
responsibility
of
managing
everybody's
viewpoint
equally,
but
it
was
hard
and
I
would
love
to
see
us
codify
that
more
and
also
understand
what
is
it,
release
lead
actually
what
what
are
the
the
powers
that
are
invested
in
there
really
to
make
those
kind
of
decisions
and
what
are
the
guidance
or
on
that?
So
one.
D
Thing
that
we
might
want
to
think
about
here
is
is,
if
you
look
at
the
Apache
process,
there's
a
very
formal
process
for
signing
releases
that
involves
you
know
the
essentially
lers
leadership
group
for
the
for
the
top
level
project.
So
so
there
is
sort
of
a
very
formal
like
we
are
blessing.
This
release
in
the
name
of
the
Apache
foundation,
type
of
thing
the
CNCs
doesn't
have
any
sort
of
you
know.
D
Official
type
of
thing
like
that,
but
it
feels
like
kubernetes
is
getting
to
the
point
where
we're
mature
enough
that
we
might
want
to
have
more
formalism
and
some
sort
of
like
Kabuki
theater
around
signing
their
releases,
which
might
be
one
way
to
sort
of
you
know
at
least
sort
of
you
know,
put
a
pin
in
it
when
we
do
release
order
discussion.
Oh
okay,
yeah.
A
H
D
I,
don't
want
to
derail
the
meeting
here.
I'm
apologize
I
mean.
Maybe
this
is
something
that
we
can
take
up
in
another
forum,
but
my
understanding
is
that
is
that
there's
some
sort
of
limited
liability
sort
of
like
your
do
on
behalf
of
the
legal
entity
versus
people
over
time,
but
but
it
also
would
be
a
way
to
actually
you
know,
have
have
sort
of
some
cross
checks
on
sort
of
you
know.
Is
this
release
ready
and
a
way
to
sort
of
surface
to
to
to?
H
Totally
think
that
we
need
to
come
up
with
a
formalized
criteria
and
process
and
some
something
like
approvers
or
shipping,
the
release,
in
addition
to
the
criteria,
because
their
criteria
will
never
cover
or
never
try
to
cover
every
situation
that
comes
up
so
I
do
think.
We
need
to
formalize
that,
but
I
think
the
burden
is
on
this
project
to
formalize
that
and
in
particular
I
think
sig
release
should
come
up
with
a
proposal
for
it,
and
then
we
can
just
iterate
on
that
going
forward.
We
can
investigate
the
liability.
H
If
you
and
I
agree
that
CN
CF
is
kind
of
our
legal
arm,
we
do
have
a
software
license
that
discusses
liability,
that
ships
with
the
release-
I,
don't
know.
If
there's
more,
we
need.
We
can
ask
CN
CF
for
a
legal
opinion
on
that
and
that's
sort
of
how
the
relationship
works.
We
we
ask
for
things
and
CN
CF
can
give
us
an
answer.
I
will.
C
D
Some
room
for
things
to
go
wrong.
There's
always
gonna
be
some
level
of
question
on
this
stuff.
Are
the
tests
valid
or
not
right?
So,
like
you
know,
ideally,
it's
gonna
be
as
all
automated
and
obvious
and
trackable,
and
you
know
written
the
blockchain.
Is
you
know
as
we
want,
but
there's
always
going
to
be
some
some
wiggle
room?
Well,
so
just
as
one
example,
somewhere.
H
We
need
to
write
down
what
are
the
test
Suites.
This
should
actually
pass
exactly
well.
We
run
a
bunch
of
tests
that
we
don't
consider
to
be
blocking
for
this
in
a
queue
or
for
releases
or
whatever
so
somewhere.
We
need
to
write
down
principles
behind
well,
it
has
to
be
in
the
list,
but
we
actually
need
a
concrete
list.
Well,
come
up.
Like
you
baby,
the
question
came
up.
Cube.
Atm
tests
were
not
on
the
list
right.
We
there
was
a
list
and
cube.
H
I
From
Cynthia
yeah
thanks
good
just
30
seconds,
because
I
saw
you
jump
in
like
Rumpelstiltskin
I
just
wouldn't
agree
that
I
actually
don't
think
CN
CF
has
a
role
to
play
here,
we're
very
available
to
the
kubernetes
project.
If
there's
any
kind
of
manual
review
that
you
want
on
something,
but
you
should
think
of
it
almost
as
a
sort
of
automatable
thing
that
you
haven't
gotten
around
to
automating
yet
and
then
we
would
follow
the
directions
that
you
gave
for
it,
but
I
definitely
would
not
recommend
delegating
any
sort
of
important
judgment
issue
on.
E
A
That
was
this
was
another
interesting
thing
where
basically,
we
had
a
lot
of
churn
at
the
last
minute
about
the
blog,
the
official
blog
that
went
on
on
career
days,
essentially
that
the
marketing
team
was
trying
to
sort
of
say
well
what
a?
What
are
the?
What
are
the
high-level
themes
and
you
know,
sort
of
follow
the
PlayBook
that
we
had
four
prior
releases
and
earlier
in
the
release
process
sort
of
at
a
lower
level.
A
So
if
a
Sikh
had
a
theme
that
was
going
to
supersede
anything
that
we
did
at
a
project
level
so
and
I
thought
that
was
really
really
great,
because
essentially,
what
we're
doing-
and
we
tried
to
call
out-
was
we're
trying
to
address
the
richness
and
depth
of
kubernetes
as
opposed
to
hey,
there's
this
giant
gaping
hole
of
things
that
Kerber
Nettie's
doesn't
do
that.
We
need
to
get
on
and
create
a
feature
for
him
or
not.
A
So
so
it's
really
a
statement
about
the
maturity
of
the
project
in
a
number
of
different
ways,
and
so,
but
what
happened
was
that
this
this
sort
of
battle,
manifested
at
the
last
minute
in
the
blog
or
where
we're
you
know
we're
used
to
giving
reporters
hey,
here's
our
release,
themes
and
here's
the
things
you're
trying
to
do.
It's
like
well
those
those
those
ideas
about
hardening
and
feature
depth
and
all
those
things
don't
make
as
easy
of
a
sell
to
people
asking.
Why
should
I
have
created
my
cognize
cluster
21.8?
A
So
we
we
had
a
lot
of
back
and
forth
on
it.
I
think
what
we
came
up
with
the
end
was
good,
but
there
was
a
lot
of
turnaround
that
and
again
really
saying
that
her
Minetti
is
somebody
said
it
in
one
of
the
beings.
I
wish
I,
remember
who
said
it,
but
they
said
Wendy
when
you
hear
people
talking
about
you
know
releasing
the
next
version
of
the
Linux
kernel
and
saying
here's
a
new
feature:
it's
like
no!
A
G
Yeah,
that's
that's
not
a
real
issue.
That's
like
some
concern
and
equation
that
we
have
this.
Also
it
it's
is
the
continuation
of
the
same
question
that
we
have
with
general
product
management
and
project
management
and
cooperation
with
with
low-level
items.
At
the
same
time,
I
would
compare
our
marketing
items
industries
and
our
marketing
activities
in
this
release
compared
to
the
reverse,
winds
and
I
would
say
that
was
probably
the
most
successful
release
in
marketing
terms.
It
was
all
the
marketing
activities
really
centralized.
G
We
had
a
dedicated
person
on
the
release
team
and
we
had
Natasha
boots
from
CNCs
my
natural
coordinate
in
all
this
marketing
codes
from
there
from
the
outer
world.
So
I
would
say
that
we
did
you
job
good.
We
can
do
it
better
and
I
hope
we'll
do
it
better,
Internet
release
and
that's
again,
it's
more
the
question
not
from
the
marketing
itself,
but
more
high-level
questions
to
come.
C
Think,
probably
starting
marketing
activity
sooner,
adding
marketing
deliverable
dates
to
the
overall
release
schedule
all
going
to
be
important
and
then
maybe
consider
creating
press
briefing
materials
separate
from
the
blog
post
itself.
That
I
think
that
certainly
there
was
some
confusion
among
among
the
press
about
how
to
talk
about
this
release.
There.
H
Also
was
not
a
coherent
marketing
plan
available,
even
at
the
blog
publication
time,
so
making
decisions
about
what
you
become
in
the
release,
blog
versus
in
separate
blogs
or
later
blogs
couldn't
really
be
made,
because
no
one
had
really
thought
about
that,
or
at
least
it
wasn't
communicated
and
documented
anywhere
I
mean.
Fundamentally,
this
issue
was
escalated
to
me.
H
So
that's
a
result,
an
obvious
symptom
of
a
process
breakdown
right
there,
so
I
think
there
needs
to
be
a
discussion
about
like
Jay,
said
about
this
strategy
like
what
are
we
trying
to
do
with
the
release
blog
post?
What
is
a
story
we're
trying
to
tell
so
we
need
to
create
a
different
approach
to
writing
those
blogs
and
doing
the
release
marketing
generally,
that
is
decided
and
communicated
ahead
of
time
and
not
at
literally
in
the
last
24
hours.
So.
C
I
apologize
if
this
has
already
been
stated,
but
I
believe
this
was
sort
of
the
intent
for
adding
a
feature
freeze
date
to
this
part
of
the
name:
release
that
I
wanted
to
be
able
to
say
at
least
using
ratios
like
roughly
speaking
by
this
date,
which
is
a
month.
I,
would
freeze
here
all
the
things
here:
the
big-ticket
items
we
plan
on
bragging
about
and
from
that
I
wouldn't
expected.
We
be
able
to
like
draft
a
blog
post
or
at
least
some
kind
of
coherent
marketing
plan,
so
I'm
not
higher.
B
He
was
not
as
engaged
and
so
I
think
we
found
a
hole,
a
single
point
of
failure,
as
we've
found
in
many
of
our
evolutions
that
we
really
need
to
pull
together
a
group
that
is
in
charge
of
making
a
marketing
effort
to
make
this
go
forwards,
though
many
thanks
to
Bob
for
all
the
work
he
has
done
in
all
of
this,
and
we
need
to
pull
together
a
broader
group
to
do
this
next
time.
Yeah.
H
And
some
new
people
were
brought
in,
but
they
weren't
adequately
ramped
up
so
I
think
making
sure
that
the
people
involved
are
able
to
understand
the
what
we're
trying
to
do
earlier
and
then
set
some
earlier
deadlines
like
the
future.
Complete
deadline
is
great.
Maybe
we
need
some
marketing
materials
deadlines.
It's
ahead
of
the
actual
release
date,
yeah.
G
I
would
say,
I
would
say
about
Aaron's
concerns.
So,
first
of
all,
yes,
we've
had
the
feature
of
freeze,
add
a
button
of
what
you
do.
We've
had
lots
of
features
that
have
been
excluded
from
the
daytime
delivered
later.
That's
again,
that's
a
digression
that
we
have
discussed
with
chase
a
few
minutes
ago.
The
second
item
is
that
we
we
were
good,
who
is
the
first
deadlines
of
submitting
the
blog
post
so
they're.
A
J
G
Answering
the
question
on
the
marketing
secret
problem
can
have
it,
but
we
may
have
this
exploring
its
unlikely
to
add
absolutely
new,
seek
without
any
any
pros
discussions,
but
we
did
stir
all
CN
CF,
so
we're
covering
marketing
activities
for
our
projects
and
for
killin
address,
especially
and
as
also
departed
responsibilities
for
the
sick
p.m.
so
we
also
people
we're
not
covering
marketing
itself,
but
we
are
cooperate
and
uncoordinated
a
surprise.
G
A
A
What
that
looks
like
in
print
and
do
the
amazing
work
that
Jennifer
and
Radek
I
did
you
know
the
word
smithing
and
all
that
stuff
that
that
maybe
shouldn't
be
the
six
responsibilities,
but
we
should
be
empowering
and
so
that
we're
not
trying
to
come
up
with
a
centralized
group
to
handle
that
stuff.
Because
if
that
the
centralized
group
is
always
going
to
be
a
bottleneck
and
it's
never
going
to
scale
all.
E
J
Documentation
comments,
yeah
I,
only
added
the
release
notes,
content
thing
on
Steve
was
the
one
who
bore
the
brunt
of
the
documentation.
Requests
coming
in
late
I
mean
some
stuff
is
gonna,
come
in
hot,
especially
at
the
level
of
release
notes,
but
there's
really.
E
And
then
it
looks
like
there
was.
A
anonymous
comment,
looks
like
better
test
coverage
for
features
being
characterized
as
beta
several
issues
discovered
in
Kulick
clients
or
rotation,
not
enabled
in
any
tests.
Does
anybody
have
any
comments
about
that?
I?
Don't
know
it
or
Aaron,
or
anybody
want
to
comment
on
that
all
right.
If
not,
then
we
move
to
more
test
coverage,
said
Jordan.
E
A
So
this
is
interesting.
We
so
there
are
things.
They've
got
punted
out
of
the
release
as
far
as
being
issues
that
were
important,
but
not
blocking,
we
need
to
figure
out
how
to
to
get
the
patch
work
flow
in
place.
So
do
those
automatically
get
set
up
as
cherry
picks,
you
know
whatever
the
fixes
are
for
those
issues
or
how
do
we?
How
do
we
make
sure
those
don't
get
lost
because
essentially,
when
things
trail
off
from
from
one
release
and
into
the
next,
that
doesn't
really
sloppy?
A
It's
time
right,
stuff
can
get
lost
in
the
churn
there
and
I
want
to
make
sure
that
there's
there's
some
way
of
both
tagging
or
something
to
make
sure
like
hey
this.
This
issue
needs
to
be
addressed
in
the
next
patch
release,
so
it's
may
not
going
to
1.80,
but
this
absolutely
needs
to
be
fixed
by
the
end
of
1.1
and
it
becomes
essentially
a
blocking
issue
for
the
next
patch
release,
but
there's
really
nothing
in
place
for
that.
So
we
need
to
figure
out
how
to
handle
that
I.
E
A
To
release
people
work
like
relentlessly
on
this,
and
not
only
that
but
I
also
worked
relentlessly
on
this
so
and
then
people
from
sig
Docs
worked
relentlessly.
So
this
was
a
the
release.
Notes
was
a
huge
time
sink
for
a
lot
of
people
and
I
would
say
that
for
the
release
team
on
1.92
people
at
minimum
and
ideally
more
than
that-
and
hopefully
we
get
more
participation
out
of
Six.
J
C
Yeah
maybe
filed
that
under
things
I'm
well,
people
went
tremendously
well,
I,
don't
know.
If
anybody
here
was
has
the
institutional
knowledge,
like
you
actually
tried
to
move
the
docs
deadline
way
way
up
for
this
release
compared
to
the
prior
releases.
To
attempt
to
solve
this
release,
no
conundrum
and
Doc's
coming
in
last
minute.
Does
anybody
feel
like
that
actually
worked,
or
what
would
we
have
done
better
to
actually
keep
the
blocks
deadline
about
two
weeks
further
out
where
it
has
been
in
the
past.
K
So
I
I
felt
like
it,
it
could
have
worked,
but
that
I
I
didn't
find
the
proper
forms
of
communication
to
let
people
know
about
those
deadlines,
so
that's
well,
that's
what
I
would
need
to
do
better
next
time
is
make
sure
I
know
all
the
email
groups
that
I
need
to
use
and
other
forms
of
communication.
That
would,
you
know,
really
communicate
to
people
that
we
have
these
deadlines.