►
From YouTube: Kubernetes 1.12 Release Team Meeting 20180831
A
B
What
what
conference.
B
Snap,
yes,
that
thing,
but
I
really
wish
I
could
have
been
at
I
was
telling
I
was
talking
with
Sarah
Novotny
earlier
this
week
and
like
I,
would
have
much
rather
been
up
there
instead
of
the
internal
Google
summit,
that
I
was
at
and
she
kind
of
wanted
to
be
at
the
Google's
on
it.
Although
she
got
to
know
it's
a
cool
thing
up
there
and
I
got
the
blog
about
the
cool
things
so
yay,
that's
a
very.
B
Look
forward
to
more
of
that,
but
I
think
the
short
unsurprising
version
is
I.
Think
we
all
collectively
agree
that
it's
more
worth
focusing
on
the
lower
level
infrastructure,
especially
stuff
related
to
releasing
stop
related
to
testing
and
automation
of
the
project,
will
take
a
little
bit
more
time
to
do
and
by
a
little
bit
we're
talking
on
the
order
of
months
yeah.
C
All
right,
I
think
I
want
to
go
ahead
and
get
started.
We've
got
probably
a
critical
mass
of
folks
on
the
line.
As
always,
this
is
being
recorded.
It's
a
community
meeting.
I
will
be
posting
it
to
you
too,
right
after
we
are
done
so
be
excellent.
People
as
I
know
you
are
so
we
are
going
to
focus
on
issues
first,
because
Guinevere
has
delicious
hot
pancakes
waiting
to
have
breakfast.
A
All
right
so
we've
had
compared
to
the
previous
weeks.
Quite
a
few
more
bugs
come
in
and
all
of
the
ones
except
for
one
that
are
coming
in
our
people
are
super.
On
top
of
it,
one
of
them
actually
open
and
closed
while
I
was
watching
it.
So
with
a
pull
request
and
everything.
So
that's
pretty
good.
There
is
one
III
cut
it
the
pull
requests
now
the
red
ones
basically
mean
red
status.
That
green
ones
have
an
open,
pull
request
and
are
actively
being
worked
on.
A
A
Then
there
is
a
new
one
that
just
popped
up
today.
That's
awaiting
triage
at
six,
eight
one,
one,
five
I
reminded
cichlids
that
code
freezes
next
week.
So
if
this
is
urgent,
then
I
want
to
see
no
sorry
yeah
yeah.
If
that
is
urgent
and
I
want
to
see
some
action
on
it.
So
I
don't
know
that
I
shouldn't
have
put
that
into
critical
urgent
I'm.
Sorry
I'll
be
doing
that
after
the
meeting
and
then
I
have
one
candidate
that
hasn't
seen
any
action
since
June
6,
4,
5,
3
0,
and
it's
priority
important
soon.
A
So
I
again
reminded
people
on
the
issue
and
I
might
escalate
to
slack.
If
I
don't
see
any
action
on
it,
it
would
get
kicked
out.
The
only
other
one
that
I'm
kind
of
worried
about
is
6
5,
3,
9,
2
right
above
it,
because
there
seems
to
be
a
lot
of
discussion
going
on
and
I,
don't
really
see
a
whole
lot
of
result
fomenting
out
of
it
update
on
on
66
376
Aaron.
Thank
you.
Those
tests
are
now
well.
Thank
you.
I
mean
thank
you
for
passing
along
the
message.
A
C
Yeah
I
don't
see
anything
particularly
like
horribly
worrying
and
there,
the
the
one
you
caught
out
for
the
active
discussion
on
the
his
taints
for
shutdown
I
think
was
the
kind
of
topic
an
attach
detach,
yeah.
C
C
Okay,
thank
you
for
the
update,
we'll
keep
an
eye
on
those,
and
it's
also
a
little
heartening
to
see
that
we
are
getting
a
little
bit
higher
issue
volume
right
now.
It's
still
extremely
low,
but
it's
a
good
sign
that
people
are
paying
a
little
bit
more
attention
on
things
right
now,
starting
to
look
in
and
record
things
as
opposed
to
just
I've
got
three
months
and
I'll
get
a
PR
in
there
somewhere.
A
One
thing
that
I've
been
a
little
bit
concerned
about
is
that
the
issue
that
I
saw
open
today,
basically
had
somebody
say:
hey
sick
leave,
can
can
we
triage
this,
which
is
following?
This
is
just
kind
of
process
observed
in
the
wild
and
I'd
like
that.
That
happened
right,
pinging
signals.
Hey,
can
reach
reach
this
into
possibly
the
miles
they.
It
was
labeled
with
a
milestone
by
the
PR
creator,
but
the
triage
on
how
critical
this
is
was
asked
from
the
cig
leads,
but
that
is
not
an
established
process.
A
A
A
C
In
the
communication
path,
I
feel
like
is,
is
less
clear
and
not
not
that
it
was
super
effective.
But
going
back
quite
a
few
months
we
had
more
of
the
Google
Groups.
Where,
from
our
lease
team
perspective
you
could
say
or
not
the
Google
Groups,
the
on
the
github
teams
or
you
could
say,
cig
foo,
reviewers,
sig,
few
sig
fou.
What
bugs
could
you
all
take
a
look
at
this?
Not
that
that
led
to
human
IC
and
necessarily
because
people
filter
so
much
of
those
types
of
notifications?
C
B
E
B
B
You
act
kubernetes,
/,
sig,
usually
I
do
sig
foo
test
failures.
If
it's
a
test,
failure
or
sake
foo,
PR
reviews,
if
they
need
to
take
a
look
at
PR
or
you
can
try
a
cig
foo
bugs
yeah,
it's
just
a
mention.
So
that's
that's
what
I've
done!
That's
what
I
codified
in
the
CI
signal
guide
for
what
it's
worth
I
didn't
I
personally
didn't
see
a
whole
bunch
of
responsiveness
from
that
I
as
a.
C
C
Okay,
so
then,
let's
transition
from
there
into
CI
signal,
a
lot
has
happened.
This
week
we
had
a
lot
of
additional
failures
and
a
lot
of
things
have
actually
gotten
fixed
as
well
in
the
house
that
some
fixes
known
and
fight
but
Mohammed.
What
would
you
like
to
say
any
specific
failures
where
you
need
help
getting
attention.
F
Because
I'm
definitely
getting
like
good
responses
now,
especially
after
I
started,
not
just
pinging
on
the
channel,
but
also
pinging
on
IBM
in
them
and
stuff.
So
right
now,
as
far
as
I
can
see,
while
there
are
a
lot
of
about
the
Figgis
left,
there
are
both
of
them
have
fixes
in
progress
like
they
did
there
a
move
to
fix
it,
I'm
kind
of
attacking
them
and
I'm
Beverly
pinging
them.
But
of
course,
like
I
can
only
call
my
timezone.
So
please
guys
go
ahead
and
try
to
cover
our
timezone
I.
F
Also
try
to
get
more
help
like
there's
a
couple
of
guys
over
it.
Just
want
to
go
to
community,
so
I
have
told
them
to
count
how
pendant
try
to
help
me
off
people
and
stuff
as
well,
so
yeah
I'm
doing
what
I
can
so.
Issues
are
up
and
I
am
also
gonna.
Do
things
that
I
think
what
this
is
something
that
needs
to
even
document
it.
So
I'm
really
looking
to
disappear
to
the
see
a
signal,
people
to
add
the
fact
that
paying
people
personally
you
know
the
other
stuff
and
I'm.
F
Also
something
that
that
is
missing
is
5v.
I
took
the
format
for
the
issues
from
I
showed
you
this
not
stiffly
documented,
like
the
infamous
there
should
be
there
inside
the
issue,
so
I'm
gonna.
Take
that
as
well.
That's
what
I'm
looking
at,
but
currency,
I
signal
status
is
I
would
say,
yellow
red
it
is
based
on
test.
Students
read,
but
fact
that
there
are
fixes
in
flight,
makes
it
the
sort
of
yellowish
red
yeah.
B
A
F
B
F
C
Unless
help
folks,
they
should
have
a
little
more
overlap
with
you,
timewise
I,
believe
yep,
and
one
thing
that
I
would
really
ask
at
the
end
of
your
day,
if
you
can
drop
a
message
and
slack
saying:
hey
folks
for
issue,
X
I
could
use
help
bringing
folks,
XY
and
Z
for
this
other
issue.
I
could
use
somebody
to
try
and
run
down
these
other
people.
C
So
if
you,
if
you
give
a
couple
of
explicit
issues
and
who
that
who
is
next
in
line
to
to
try
to
reach
out
to
help
for
them,
then
we
can
do
that
with.
If
it's
not
explicit,
it's
hard
for
us
to
know
the
we've
got
to
go
troll
through
all
of
the
the
open,
failing
tests
and
current
status
on
test
grid
and
try
to
reach
riaj,
so
any
triage
that
you've
done.
If
you
capture
that
into
sort
of
a
message
to
us,
then
we
can
continue
on
with
it
explicitly.
G
Movement
again,
I'm,
sorry,
if
I'm
asking
this
again,
is
there
a
spreadsheet
where
you
have
the
title
of
the
issue
and
then
the
issue,
the
body
itself,
it's
kind
of,
at
least
for
me,
it's
tough
for
me
to
find
the
tracking
bug
for
each
failure.
Yeah
there
is
actually
a
sheet
in
the
chat
or
in
the
slide
channel
yeah.
F
F
C
E
Actually
it's
disabled
I
cannot
share
anyway.
I
wanted
to
say
that
if
you
go
to
release
massive
walking,
we
are
deprecated
into
a
comedian,
GC
self
hosting
tests.
This
one
in
particular,
is
a
feature
that
Canadian
no
longer
support
and
the
other
one
is
Cuba
dmg
see
now.
This
is
problematic,
I
kind
of
fixed,
the
kubernetes
anywhere
issue.
That
was
not
a
like
with
help
from
Ben
from
six
testing,
but
the
problem
seems
elsewhere,
because
some
of
the
notes
are
not
reading
them
and
I
cannot
figure
out
what
the
problem
is.
E
C
B
B
Have
my
historical
context
is
cube.
Adm
has
generally
always
continuously
failed
until
right
before
release
it's
an
anti-pattern,
but
that's
nothing.
New
scalability
same
thing,
anti-pattern,
but
nothing
new
upgrade
tests
aren't
always
the
while
list
of
rides
because
they
usually
almost
always
fail,
except
for
a
few
there
and
weeks
as
we
focus
on
them
during
burned-down.
C
There's
not
a
lot
of
signal.
The
signal
gives
the
impression
of
unhappiness,
but
then,
if
all
if
all
we
are
focused
on
during
the
burndown
is
those-
and
we
do
have
a
track
record
of
managing
to
trick
things
back
that
fix
and
get
that
stabilized
at
the
last
minute
that
it's
not
a
good
pattern,
but
there's
a
reason
to
think
we
could
pull
that
together
again.
C
So
yeah
I
was
gonna
also
say
so,
in
addition
to
kind
of
documenting
criteria,
process
for
adding
and
removing.
So
it's
clear
like
if
we
wanna,
like
historically
I,
feel
like
we've
decided
as
a
team.
Okay,
we're
not
going
to
block
on
this
thing,
that's
in
the
blocking
bucket,
but
to
make
it
to
make
a
decision
on.
E
So
I
can
explain
a
little
bit
about
the
self-hosting
removal
in
particular,
so
the
the
feature
is
not
maintained,
nobody
uses
it
and
the
test
is
like
red
and
even
if
we
fix
it
somehow
I
mean
we
probably
won't
be
able
to
fix
it
because
of
the
code
base
is
broken
right
now
when
we
are
removing
the
code
behind
the
self-hosting
feature.
So
this
is
no
not
going
to
work.
B
B
But
the
thing
is
it's
now
out
of
date,
the
thresholds
that
were
mentioned
in
it
no
longer
apply
because
time
has
passed
so
I
kind
of
need
to
revisit
it
and
then
I
need
to
figure
out
who
I
shop
it
around
to
you
to
make
it
official
policy
whether
or
not
like
this.
Do
you
in
committee
needs
to
be
involved,
I,
don't
think
they
do
Greeley's
team
definitely
does
a
cigar.
Detector
need
to
be
involved
so
on
and
so
forth.
So
I
have
a
pretty
good
first
pass.
B
G
C
C
Okay,
continuing
backwards,
then
through
the
agenda
talk
Nick
could
not
make
it
today,
but
heaping
me
earlier
today,
letting
you
know
that
he
had
played
around
so
I'm
with
hack
MD
and
is
liking
it
and
there's
a
link
there
for
a
draft
he's
going
to
try
and
use
I
think
hack
MD
for
drafting
the
release.
Notes
then
documentation
we're
looking
do
we.
C
H
So
I
we
have
I
have
a
fix
that
we
could
implement
and
I
have
another
more
temporary
fix
that
we
could
implements
to
have
things
working
later
today,
but
it
I've
been
talking
with
Steve
from
Red
Hat
in
the
sig
testing
chat
right
now
and
it's
sounding
like
it
might
be
preferable
just
to
wait
until
we
have
the
like
more
permanent
fix
in
and
I.
Don't
know
that
I
can
finish
that
in
a
single
day.
So.
H
C
Have
no
strong
preference
as
long
as
we
we
have
a
mechanism,
that's
enforcing
and
I'm
I'm,
pretty
I
feel
like
we're,
making
a
lot
of
changes
in
around
mechanism
and
this
release
and
that
I've
kind
of
always
expected
that
we'd
have
a
few
hiccups
like
this,
so
I
would
rather
be
pragmatic
and
go
ahead
and
keep
moving
forward
versus
one
of
the
other
alternative
you
dimension
would
be
to
say:
well,
let's
put
this
off
to
the
next
cycle.
I
would
really
prefer
to
not
do
that.
C
H
It
should
be
really
easy,
too,
especially
way
of
this,
to
make
you
still
set
up
that
shouldn't
reveal
for
me
to
switch
back
so
that
there's
no
risk
of
like
having
March
blocked
or
anything
like
that.
It's
more
really
a
question
of.
Is
it
okay
to
mine,
early
change,
the
bed
experience
and
the
experience
from
OPR
shouldn't
really
change
like
you're
still
gonna
see
a
status
will
be
just
like
this.
Thank
you.
C
I,
don't
have
a
strong
objection
there,
and,
and
also
no
I
mean
it's
having
some
sort
of
a
B
comparison
is
always
nice,
so
to
know
that
we
were
successfully
moving
forward
with
one
thing
and
switched
over
to
the
other,
as
opposed
to
to
having
the
transitions
at
the
major
points,
but
just
just
within
freeze
to
do
a
transition
and
see
the
freeze
would
still
freeze
I,
think
that's
useful
information,
so
yeah
you,
you
I,
would
say
thumbs
up
for
me.
Anybody
else
on
the
call,
thoughts,
okay,.
C
So
one
thing
that
I
would
ask,
though,
depending
on
how
it
goes
and
recognizing
that
Monday
is
a
holiday.
If
you're
not
ready
earlier
on,
Thursday
I
would
say:
maybe
don't
don't
turn
it
on
on
Friday.
H
Sure
yeah,
that's
that's
totally
reasonable
and
yeah.
So
I'll
wait
until
we
have
like
the
full
workaround,
completely
implemented
that
I'll
probably
have
that
early
next
week.
The
other
piece
of
info
that
I've
been
working
on
is
the
plug-in
to
require
kind
in
stick:
labels,
that's
merged
and
deployed
now,
but
just
needs
to
be
configured
so
I'll
do
that
this
morning
and
we
should
start
having
those
enforced.
C
G
Greek
thing
was
there
going
to
be
a
demo
after
before
Cole
you
turn
it
on
or
I,
couldn't.
B
B
Yeah,
like
I,
used
to
be
all
all
ranty,
that
I
thought
the
precondition
for
switching
everything
to
tide.
That
was
to
have
really
awesome,
docks
and
then
shop
it
around
and
do
a
whole
Roadshow
and
give
people
a
time
to
prepare
for
it
and
then
flip
the
switch
and
now
we're
kind
of
flipping
a
switch
and
then
doing
that
right
at
about
the
same
time.
But
again,
I
I
trust
that
you
and
Cole
have
enough
perspective
on
stability
that,
if
you're,
confident
I'm
confident
right
along
with
you.
C
C
C
All
right
and
that
case
reminder
no
meeting
Monday
we'll
have
the
Europe
Asia
friendly
time
on
Tuesday
and
then
Wednesday
and
Friday
next
week
as
well
and
I'll
be
looking
to
to
chat
with
coal,
in
particular
as
the
week's
go
by
on
slack
and
touch
base
on
things
and
see
what
we
can
get
into
the
community
meeting,
hopefully
on
Thursday
as
well
for
heads
up
to
folks
all
right
have
a
great
long
weekend,
everybody
who's
having
a
long
weekend
and
a
great
weekend
otherwise
for
everyone.
Thank
you
happy
weekend.