►
From YouTube: Kubernetes 1.12 Release Burndown Meeting 20180914
A
B
Alright
I
see
the
one
person
who
I'm,
especially
looking
to
get
some
status
from,
is
join
so
I'm
gonna
go
ahead
and
get
started.
It
is
Friday
September
14th.
This
is
the
112
release
burned
down
meeting
of
the
day,
I'm
Tim
pepper
your
release,
Lee
this
meeting
is
recorded
and
will
be
posted
to
YouTube,
so
the
world
will
see
our
breakfasts.
B
Quinn
smirks
I
barely
just
finished
my
breakfast.
Let's
see
what
what
to
say
so
still
should
have
RCE
on
Tuesday
released
target
is
officially
September,
25th
I'm,
starting
to
particularly
worry.
We
have
had
a
lot
of
good
fixes
coming
in
this
week,
but
we're
not
seeing
those
reflected
in
tough
status
and
I'm,
not
understanding.
Why
entirely
yesterday,
we
knew
we
had
a
few
issues,
but
those
being
clearly
resolved.
B
Let's
see
the
the
main
thing
that's
worrying:
there
then
is
the
underlying
provider
seems
to
be
being
a
problem
and
as
much
as
we're,
we
can't
see
much
of
anything
we're
having
on
slowness,
that's
running
longer
than
expected
a
bunch
of
things,
timing
out
and
it
feels
like
either
we've
got
a
hosting
issue,
or
else
we
have
an
undiagnosed,
really
major
pod
scheduling
and
dispatching
issue.
I'm
inclined
to
not
think
that
and
as
much
as
we
do
have
lots
of
other
tests
elsewhere,
an
test
grid
that
are
consistently
successful,
so
I
don't
know
Barney.
B
D
Yeah,
hey
we're
still
trying
to
figure
out
what
is
the
root
cause,
so
they
were
trying
to
also
understand
what
might
be
the
mitigation
to
it.
Turn
off
jizya
decay,
Eve
deaths
St
as
a
blocker
to
remove
it
from
here.
If
your
blockages
are
that's
the
we're
trying
to
understand
that,
but
we
don't
know
at
because
the
tests
failure
seemed
to
be
random
and
except
for
the
DNS
and
a
couple
of
Pete
storage,
related
failures,
others
seems
to
be
random.
D
B
It
definitely
does
not
feel
consistent
in
it
and
it's
a
classic
inconsistency
that
makes
it
really
hard
to
debug
so
I'm
curious,
other
people's
thoughts.
I
I
think
some
GCE
things
have
gotten
better,
so
a
good
sign
but
I
don't
feel
like
it's
completely
gke.
We
have
quite
a
few
things
still
in
GCE
that
aren't
clearing
up
and.
E
B
I
feel
like
the
symptom
is,
is
pretty
much
the
same
as
the
gke
one
in
terms
of
things,
just
kind
of
failing
things,
I
should
have
been
running,
aren't
running,
don't
know
if
it's
resource
starvation
or
or
what
but
I
I
I
feel
like
if,
if
we
just
drop
gke,
it's
not
clear
that
that's
gonna
be
a
an
answer.
Okay,.
D
B
So
that's
again,
what
kind
of
makes
me
feel
like
hosting
so
some
of
these
things
we're
not
seeing
the
original
symptom
after
the
fix,
but
we're
still
not
getting
a
green
test
result:
the
HP
a
debugging
as
of
like
an
hour
ago,
Sally
thinks
he
might
have
a
potential
fix,
so
we'll
watch
that
space.
That
could
be
a
positive
change
and
then
yeah
the
mate.
The
main
list
of
things
there
to
be
watching
is
is
around
GCE
and
gke.
I
guess
that's!
B
I
skipped
over
on
a
skipped
over
scalability,
so
a
bunch
of
change
there.
Yesterday,
we
we
believe,
based
on
six
availability,
input
that
they've
cracked,
the
the
main
scalability
issues,
but
for
one
remaining
thing-
and
that
is
with
the
switch
from
cube,
pianist
accordion
s,
Cordia
neces
pods
were
being
killed,
and
this
is
strange
because
core
DNS
was
generally
expected
to
perform
better
in
terms
of
its
throughput
on
on
satisfying
requests
and
do
so
at
a
lower
memory
demand
so
that
the
initial
assessment
was.
B
If
so,
that's
that
at
least
gives
us
a
basis
to
move
forward
on
scalability,
but
it
still
leaves
the
core
DNS
question
open
and
the
core
DNS
locks
and
scalability
are
talking
together
about
how
to
devise
an
experiment.
If
we're
needing
to
delay
the
release
anyway,
they
could
kick
off
something
on
Monday,
maybe
get
some
data,
maybe
debug
and
fix,
and
still
have
time
for
another
run,
to
prove
greenness
again
on
stability
on
scalability,
with
core
DNS
ahead
of
a
delayed
release.
B
A
This
actually
ties
into
most
of
what
I
was
going
to
talk
about
I.
Just
I
just
saw
that
we
reverted
the
core
DNS
enablement
just
but
it's
basically
four.
It
doesn't
fix
anything
as
for
analysis
purpose,
which
is
why
all
the
related
issues
which
are
mostly
which
I
basically
is
it
six
eight
four
nine
zero
and
six
eight
two
six
zero
they're
still
open
because
they're
basically
related
to
that
feature
and
not
and
we're
still
waiting.
But
people
are
definitely
talking
about
yeah.
A
Well,
maybe
we,
this
feature
is
not
going
to
make
it
into
the
release,
and
in
that
case
those
issues
would
be
closed
and
deferred
to
113.
They
are
in
general,
I.
Think
four
out
of
the
seven
issues
are
in
danger
of
being
rolled
back,
so
the
breaking
the
bug.
Will
you
know
the
break
that
the
thing
that
caused
the
breaking
change
will
be
rolled
back
and
the
fix
will
make
it
into
one
thirteen
which
is
unfortunate
but
I'm
leaving
them
open.
A
The
only
one
I'm
really
worried
about
is
the
scalability
one,
because
that's
time
boxed,
but
the
other
ones
I'm
just
keeping
them
open,
because
there's
an
understanding
that
well
if
it
doesn't
get
fixed,
it's
not
it's
not
really,
stalking.
That
would
be
six
seven,
eight,
two
three
and
six,
eight
six
one
three
is
currently
also
just
being
reverted
and
debug.
A
So
if
they
come
up
with
a
fix
great,
if
not
it's
not
release
blocking
and
then
I
have
to
open
PRS
that
are
being
that
one
is
really
close
to
closing
needs
an
LG
Tia
and
the
other
one
people
are
still
discussing.
So
if
nothing
happens
by
Monday
or
Tuesday
I'm
going
to
have
to
come
with
the
banhammer,
and
there
is
one
PR
that
looks
like
it's
an
easy
fix
and
someone
said
they
were
going
to
do
it
over
the
weekend.
So
since
again,
it's
not
released
blocking
I'm,
okay,
letting
it
pass
which.
A
B
A
A
B
A
A
Wrong
out
around
words
that
didn't
mean
what
they
thought.
They
meant:
okay,
cool
yeah.
It
does
feel
like
to
me.
We
need,
like
some
sort
of
in-between
I,
don't
know
some
sort
of
in
between
states,
because
right
now,
I'm
hearing
a
lot
like
yeah.
We
would
like
to
get
this
until
112,
but
it's
okay,
if
we
don't,
which
so.
B
Just
glancing
it
what's
there
for
PRS
that
are
labeled
release,
blocker
I
think
we
stopped
using
it
and
there
it
would
make
sense
to
since
priority
critical
urgent
is
one
of
the
metrics.
That
would
allow
things
to
merge
to
go
ahead,
and
if
people
are
saying
that
it's
not
going
to
be
112
milestone
blocking
that
we
drop
priority
critical
urgent
too
important
soon
and
then,
if
the
fix
is
still
needed
for
112
it'll,
be
a
cherry
pick
back
I'm
in
a
patch
release
in
the
future.
A
B
A
B
E
B
C
Up
and
do
all
that
inside
of
that's
all
docks
related
and
all
asynchronous
for
our
test.
So
as
far
as
I'm
concerned,
with
the
features
listed
on
the
spreadsheet,
we
are
good
to
go.
It
would
be
nice
to
know
with
the
net
accordion
SS
drop
out,
though,
because
I
have
to
back
those
commits
out
of
the
branch
at.
B
F
F
The
mute
button
so
I
will
I
will
talk
until
she
gets
unmuted,
so
basically
Lyndsey
put
in
a
fix
to
our
generation
routine.
That
will
have
the
the
release
notes
and
then
the
sings
that
they
come
from
to
make
it
a
bit
easier
for
the
for
us
and
for
the
cig
leads
to
define
where
these
things
actually
go.
F
B
F
B
A
B
F
B
Alright
futures
Doc's
released
its
communications.
Similarly,
their
Kaitlyn
I
dropped
a
mention
because
I
had
mentioned
Cordia
Ness
and
one
of
the
the
doc
drafts
so
I
mentioned
that
I
all
recommended
there,
and
actually
we
don't
have
Kaitlyn
today,
so
I
made
sure
that
I
propagated
the
information
there
that
we
may
have
to
drop
that
wording.
B
B
You
all
right
Barney,
if
you
find
anything,
hear
anything
figure,
anything
out
need
any
help
or
anything
today,
I,
don't
know
that
I
can
do
as
much
as
your
own
internal
network
at
Google,
assuming
we've
got
issues
that
are
google-specific,
but
but
let
me
know
sure
all
right
have
a
great
day.
Everybody
have
a
great
weekend
and
we
will
see
what
we
can
get
for
improvements
today
and
what
test
grid
shows
us
on
Monday.
Why.