►
From YouTube: Kubernetes 1.10 Release Burndown Meeting 20180319
Description
A
Hello,
everybody
welcome.
It
is
Monday
March,
19th
2018.
This
is
the
kubernetes
1.10
release
team
burned
down
meeting.
We
are
burning
down
the
release.
So
if
you've
got
your
fire
extinguisher
handy,
let's
get
this
thing
put
out
by
a
next
Monday
and
released.
So
today
we're
going
to
run
through
a
few
things,
but
to
respect
time
of
those
people
calling
in
at
ridiculous
hours
in
China
we're
going
to
move
through
the
risk
assessment,
real,
quick
and
we're
going
to
start
with
issue
or
a
pull
request.
A
B
Sure
so,
as
part
of
the
as
part
of
the
updates
required
to
get
the
alpha
CI
test
green,
one
of
the
Alpha
features
is
using
the
scheduler
to
schedule.
Daemon
set
pods
and
we
are
discovering
some
implications
of
that
which
is
good
to
discover
in
alpha
state
and
so
an
effort
to
turn
the
alpha
CI
job
green.
Several
scheduling
changes
have
come
in,
and
that
is
what
these
pull
requests
are
about.
B
C
B
A
B
B
D
My
two
cents,
the
systems,
is
that
the
yama,
since
we,
if,
if
we
were
going
to
go
with
six
one,
three
one
three,
that's
a
code
change.
If
it's
just
going
with
six
one,
three
one
one,
it
is
just
one
yama,
that's
changing
to
make
the
thing
go,
green,
so
definitely
that
would
be
more
preferable
than
six
one.
Three
one
three
and
you
know
it's
not
gonna
affect
that
much
people
can
edit
their
AML
file
if
they
want
to
change
the
way
things
work,
yeah.
B
Adding
a
toleration
yeah
is
about
the
safest
thing
you
can
do.
All
that
does
is
especially
on
a
damon
set
where
you're
not
going
to
change.
The
number
of
things
scheduled
Damon
site
is
going
to
try
to
run
on
every
node
and
adding
this
toleration
will
just
improve
the
chances
that
it
runs
on
every
node.
If
you
don't
have
these
alpha
features
turned
on,
it
won't
affect
anything
at
all.
Actually,
so
six
one
three
one
one
I
would
characterize
extremely
low
risk.
B
A
F
A
I
mean
given
that,
given
the
situation
that
we're
in
I
I
don't
know
that
we
can
justify
six
one
three
one,
three
mm-hmm
I
am
really
leaning
toward
the
ML
change.
Four
six
one
three
one
one,
but
that
I
mean
that
in
a
way,
this
kind
of
kicks
the
can
down
the
road
and
I
think
that
maybe
it
is
part
of
the
retro.
We
talked
about
this,
because
this
is
something
that
is
going
to
continue
to
be
a
problem.
Looking
forward
and.
B
So
actually,
the
signal
that
this
test
gave
us
was
that
switching
to
the
scheduler
and
working
via
taints
and
Toleration
Xin,
this
way
is
not
strictly
backwards
compatible
because
we
see
this
requirement
to
either
change
the
manifest
of
a
component
that
used
to
work
or
start
adding
colorations
and
an
optional
admission
plugin,
and
so
this
is
good
signal
to
get
it's
something
that
will
have
to
be
considered
and
documented
or
worked
around
before
these
features
can
migrate
out
of
alpha.
So
I'm
glad
we
are
finding
out
that
these
changes
are
required.
B
A
What
of
what
is
the
okay
so
just
week,
so
we
can
summarize
to
make
sure
that
I
understand
the
full
scope
of
work
that
needs
to
be
in
here.
So
we
we
get
six
one
three
one
one
merged.
What
is
the?
What
are
the
other
steps?
We
need
to
take
to
make
sure
that
we're
doing
due
diligence
around
this,
especially
in
terms
of
moving
forward
for.
B
There
are
more
considerations
that
need
to
be
written
down
out
of
this,
because
it
could
have
a
customer
impact
on
the
way
Damon
sets
work
if
they
want
to
run
those
if
they're
using
host
network
and
I
want
to
run
on
nodes
that
aren't
saying
their
network
is
ready,
so
ready,
but
but
for
110
if
I
understand
Klaus
correctly,
this
is
the
this
is
six
one.
Three
one
one
is
the
extent.
A
E
E
E
G
G
That's
what
I
was
hearing
yeah,
but
we
can
take
that
one
up
later.
I
just
want
to
say
for
110
I
mean
just
the
relevant
piece
for
right.
Now
is
six
point
three
one
one,
because
it
looks
really
straightforward
to
dock
really
seems
like
just
fine
on
this
side
to
right
indicator
at
this
point.
In
anything.
B
H
G
And
the
only
reason
I
flagged,
that
was
that
that
Klaus
had
offered
to
update
the
dock.
Oh,
Klaus
and
I
could
do
this
a
synchronously
sort
of
determine
the
final
state
of
that
dock
pull
request
and
and
get
it
merged.
We
don't
need
that.
We
have.
You
know
we
have
a
24
hour.
Turnaround,
add
more
for
that
one.
So
yep
doesn't
need
to
wrangle
that
one
immediately
okay.
C
A
A
E
A
H
A
A
A
Right
great
well
get
that
get
that
over
line.
Hopefully,
today,
before
we
end
code
freeze
them,
then
we
have
to
go
through
creating
trade
picks
for
these
things,
which
is
a
different
animal,
all
right
Klaus.
Thank
you.
So
much
anything.
Does
anybody
need
to
anything
with
cost
before
we
release
him
to
asleep.
I
B
A
B
Request
so
there's
this
is
against
master
and
and
thinking
about,
like
interactions
between
110
and
111.
So
by
default,
the
interactions
between
110
and
111
will
be
fine,
which
is
why
I
don't
think
this
particular
one
is
released.
Blocking
there
are
two
polls
open
against
one
nine
one
is
the
test
fix
that
I
had
and
one
is
a.
B
B
So
I
had
my
test
change,
open
and
tagged.
The
six
storage
leads
for
their
input
and
I
guess
we
could
get
their
input
on
the
controller
back
port
as
well.
I
would
defer
to
the
sick
storage
leads
which
they
want
to
do,
unfortunately,
until
what
we
don't
have
a
good
way
to
run
the
downgrade
tests
against
the
pull
request
and
the
method
I
use
for
running
e
DS
I,
don't
actually
have
it
set
up
to
be
able
to
install
during
the
test.
So
I
can't
run
the
downgrade
test
locally.
A
A
J
B
B
H
H
Think
I
know
but
I
never
run
it
locally
as
well.
I
think
Chevron,
stick
storage,
that's
wrong!
Friday
I
mean
if
Ellen
had
someone
run
at
Friday
and
croci
gives
them
instructions.
B
A
D
A
Feeling,
okay
and
then
we
just
need
to
document.
It
is
right
under
known
issues,
I
feel
bad
that
we're
kicking
the
can
down
the
road
on
that.
But
I
mean
yes,
there's
a
limited
amount
of
time
for
us
to
get
stuff
done.
So
if
something
come,
if
a
pull
request
comes
in
in
the
next
24
hours,
I'll
have
you
look
at
it
dims,
but
otherwise,
let's
assume
it's
10.1,
not
10,
that
one
yeah.
D
K
A
K
K
K
A
Some
reason
I
did
not
see
any
of
that
communication,
which
is
weird
because
I
should
have
so
I
can
I
can
copy
you
directly
on
that
thread.
Maybe
that
could
you
that'd
be
great?
It's
J,
Dumars
JD
ARS
at
gmail,
okay,
I'll!
Do
that
right
now.
That's
super
helpful
yeah,
so
I'm
gonna
open
the
floor
to
any
any
objections,
because
this
isn't
I
mean
I'll
probably
make
the
call,
but
I
want
to
know.
If
anybody
has
any
objections
to
bumping
this.
J
A
Yeah
I
mean
for
me
just
just
as
a
general
rule
if
we
can
get
off
the
Alpha
image,
I'm
Way
happier
to
get
off
the
Alpha
just
from
a
stability
standpoint,
but
the
the
introduction
of
the
rate-limiting
stuff
felt
like
a
relatively
significant
change
with
possible
unknown
impacts.
So
I
just
wanted
to
fully
understand
any
risks
associated
with
that.
But
if
you've,
if
you
vetted
those
and
those
seem
like
you've,
got
buy-in
and
Tim's
okay
with
it,
then
I
I
have
no
no
qualms
on
that
myself.
A
A
A
That's
great
yeah
cuz
we're
eventually,
if
we're
gonna,
hopefully
have
these
in
an
markdown
file
somewhere
in
github.
So
we
know
what
exceptions
are
approved.
Okay,
so
all
right
moving
along.
So
today
is
the
end
of
code
freeze.
In
theory,
I
would
like
to
have
the
fine
test
and
for
finding
coal.
If
you
want
to
weigh
in
on
that
thoughts,
and
if
anybody
else
send,
you
want
to
weigh
in
on
that.
H
A
B
A
That's
good
question.
Thank
you.
Many
any
other
questions
concerns
what
about
cherry-picks
cuz.
Obviously,
now,
once
we
open
the
the
master
back
up,
we're
gonna
be
in
the
business
of
cherry
picks
for
things
that
means
that
any
open
PRS
are
gonna.
We're
gonna
have
to
get
those
either
I,
don't
know
what
the
hell
we
do
sexually.
A
H
B
B
A
A
A
A
A
A
A
D
A
D
Mean
we
found
one
and
that
was
already
reverted
and
so
now
we're
trying
to
find
another
one
right.
The
problem
is
so
one
question
I
had
for
Sen
here
is
sent.
Can
you
tell
us
if
there's
any
other
jobs
where
this
part
startup
is
likely
to
you
know,
show
so
I
can
go.
Look
there
to
see
when
it
actually
started
other
than
the
bisection.
Of
course,
I.
H
J
J
B
J
B
Didn't
reference
the
flu
nd
issue,
so
it's
a
little
unclear
to
me
the
heritage
of
the
issue
if
it
was
a
flake
and
there
were
potentially
multiple
causes
for
the
flake,
one
of
which
was
the
flu
in
descaler
and
the
other
of
which
was
this
wait
test
code,
it's
unclear
to
me.
It
would
be
good
to
get
Shyam
to
verify
that
this
is
resolved.
B
B
B
B
The
test
already
actually
improved
significantly
after
the
fluency
fix
went
in.
It
was
going
red,
maybe
50%
of
the
time,
and
it
went
down
to
like
one
out
of
seven
or
eight,
and
so
we
saw
a
noticeable
improvement
after
the
fluency
fix
and
the
test
code.
Six
this
morning,
I
expect
to
take
care
of
it
permanently.
D
B
B
A
A
Yeah,
let's
see
what
so
there's
yeah,
there's
three
and
actually
I
Caleb.
We
were
talking
about
this
I,
don't
know
if
you're,
but
just
making
sure
with
Kobe
diem
folks
that
that
at
the
bumper
110
it
is
that
something
we
needed
to
do
before
we
cut
our
C
or
do
we
do
it
before
we
actually
cut
the
release
itself
on
Monday.
He
did.
D
A
Right
so
no
problem,
so
so
before
you
get
the
get
the
thing
going:
the
RC,
let's
merge
this
okay
and
this
show
looks
free
to
go
so.
F
A
Okay,
so
all
right
today,
well
we'll
get
all
that
sorted
out.
We
don't
really
need
to
go
through
metrics
I.
Think
we
pretty
much
did
all
this.
Is
there
anything
the
name
that
you
want
to
add
Josh
in
terms
of
where
we're
at
I
mean
you're
writing
some
stuff
from
the
channel
I
didn't?
If
you
want
to
talk
about
those
well.
G
J
Of
their
questions
was,
is
it
acceptable
for
a
downgrade
fix
to
go
into
one
that
9
latest
and
I
didn't
know
the
answer
to
that,
because,
as
far
as
I
know
our
perspective
and
downgrading
is
the
main
reason
for
some
downgrade?
Is
they
upgrade
to
one
that
10
stuff
starts
failing
and
they
want
to
downgrade
Beneful
case
they're,
not
necessarily
gonna,
be
downgrading
to
the
latest
release
of
1.9
I.
H
A
B
B
B
A
J
H
J
A
I
may
kick
myself
for
saying
this,
but
I
believe,
with
the
exception
of
the
scalability
testing.
We
know
everything,
that's
in
the
release
is
stable
and
the
downgrade
path
is
documented
enough
that
it's
actually
can
be
released
in
people
if
they
need
to
to
deal
with
that
problem
within
the
storage
not
allowing
the
there
the
names
space
cleanup
or
whatever
it
is
that
stuff
is
already
taken
care
of
the
manual
steps.
So
I
think
we
are
down
to
the
one
thing
yeah.
B
So
the
decision
before
us
for
the
PRS
that
are
open
are,
do
we
merge
none
of
them
and
go
with
the
documentation,
fix
and
manual
inspection
of
our
downgrade
tests,
which
we've
done
and
we
verified
that
the
rest
of
the
tests
are
passing
or
do
we
merge
the
downgrade
tests
to
change
so
that
we
get
actual
automated
green
downgrade
tests
or
do
we
merge
the
code,
change
that
we're
looking
to
do
in
1/9
and
I?
Think
the
timing
on
that
is
not
ready
yet
so
I
think
our
choices
are
really
for
today.
H
F
H
J
A
A
Okay,
so
all
right,
so
what
I'm?
What
I'm
hearing
is?
If
we
had
to
pull
the
if
we
had
to
pull
the
go
no-go
lever
right
now,
I
would
be
leaning
toward
ago,
even
with
the
the
scalability
issues.
As
it
stands,
I
mean
I
want
to
have
those
fix,
but
right
now
it
sounds
like
it
sounds
like
the
release
is
in
relatively
good
shape.
B
B
G
I
added
to
the
notes,
as
you
can
see,
and
the
PR
that
I'm
still
mildly
concerned
about
I'm
is
not
there.
It's
the
one
for
the
one
that
nine
downgrade
issue
and
the
only
reason
I'm
a
little
concerned
is.
There
still
seems
to
be
a
little
bit
of
Chur
I'm
a
bit
of
back
and
forth
on
in
tech
review,
and
what
I
want
to
do
at
this
point,
I
haven't
had
a
chance
to
look
carefully
this
morning,
I'm.
G
So
between
now
and
tomorrow,
morning's
meeting
I
want
to
put
my
version
of
the
trunk
together
and
see.
If
I
could
just
stick
it
in
front
of
people
and
say:
look
have
we
got
all
of
these
bits
that
you've
been
adding
to
the
dock
or
taking
out
of
it
clearly
in
place
because
we're,
depending
on
this
I
realized.
You
know
it's
a
bit
of
a
niche
case,
but
we
need
to
get
it
right
and
I'm
not
satisfied
with
the
discussion
in
the
in
a
PR
right
now.
G
So
this
is
one
of
the
reasons
why
I'm
happy
to
hear
the
decision
to
go
with
not
updating
the
1.9
tests.
I
mean
it's
not
a
huge
big
deal,
but
if
indeed
work
around
that
has
to
get
documented
is
a
piece
of
the
fix,
automated
tests
to
sort
of
circumvent
that,
even
in
a
niche
case,
that
seemed
to
me
like
quite
the
right
thing
to
do
so.
G
Anyway,
the
pull
request
is
in
there.
We
couldn't,
you
know,
gun
to
my
head.
We
could
merge
it,
it
would
be
okay.
We
got
a
little
bit
of
time
this
week
to
get
it
a
little
bit
cleaner,
so
trying
and
to
give
it
a
bit
more
love.
Okay,.
G
It's
looking
pretty
good
I
am
in
Seattle
this
week
and
I'm
meeting
with
Steve
Harry
this
afternoon,
and
one
of
the
things
I
want
to
do
is
go
over
the
automated
dot
generations,
so
that
should
he's
I
bought
I
did
have
it.
I
took
a
few
stabs
at
it
last
week,
so
I
waited
because
I
knew
I
was
gonna.
Have
this
meeting
I
think
that
we
are
in
good
shape
overall,
okay,
letting
y'all
know
where
my
little
bits
of
remaining
concern
live
and
also
with
that
Damon
set
scheduler
thing,
there's
still
a
you
know.
G
A
B
J
D
A
Yeah
that
makes
that
makes
a
ton
of
sense
honestly,
and
the
projects
I've
been
around
with
the
level
of
care
we
give
to
alpha
documentation
and
and
socialization
is,
is
much
higher
the
most
so
so
I
think
that's
good
that
we're
doing
that,
but
yet
I
think
Jordan.
If
you
can
help
give
that
a
reliant
I
would
be
great.
Thank
you,
yep,
okay,
Jennifer
anything
else.
You
wanna
add.
A
All
right,
let's
move
to
marketing.
Okay,
so
I
can
start
this
conversation
out
for
this
we
need
we
need
to
get
the
blog
in
much
better
shape
than
it
is
right
now,
with
with
the
level
of
changes
in
terms
of
things
not
shipping.
This
time
the
blog
feels
a
little
bit
like
like
a
some
sort
of
balloon
animal
that
has
had
part
of
the
the
air
sucked
out
of
it.
Now
it
doesn't
look
like
a
giraffe
anymore,
so
it
looks
like
a
droopy
giraffe,
so
we
need
to
to
fix
that
Natasha.
C
Definitely
so
bodily
from
the
storage
big
hat
went
in
last
week
and
he
updated
his
section
for
what's
going
on
with
storage
and
what
are
the
features
and
things
like
that.
So
that
section
does
not
look
deflated,
but
networking
and
security-
and
you
know
all
those
other
sections
are
very
very
bare
and
the
upfront
intro
paragraph
really
can't
be
revived
until
the
other
sections
are
filled
in
a
little
bit,
because
this
is
really
introducing
the
release
to
everyone
and
soon
with
the
title
that
will
need
to
be
revised.
C
A
C
Would
I
would
like
to
add
on
deadline
for
this?
So
if
we
could,
we
could
get
something
in
today
or
tomorrow.
Then
I
would
still
have
time
before
the
release
to
go
to
media
Wednesday,
Thursday
and
Friday.
With
this
and
do
a
few
interviews
worst
case
scenario,
is
we
get
the
vlog
in
a
really
good
shape
on
Thursday,
and
we
only
have
time
to
conduct,
maybe
like
one
or
two
interviews
on
Thursday
Friday
and
a
lot
of
the
press
write
directly
from
the
blog
as
their
source
for
their
stories.
A
A
C
M
M
A
M
Else
may
be
severe
mission,
your
because
we
have
one
alpha
feature
that
is
related
to
C
Ernie
and
would
like
to
understand
if
it's
extremely
available,
specifically
in
the
current
status,
okay,
and
if
you
should
mention
it
or
not,
okay,
that
again
are
the
most
the
most.
The
most
important
for
me
personally
would
like
to
see
are
here
in
the
block:
are
people
from
significance?
M
A
Alright,
and
if
anybody
else
on
this
call
can
look
at
that,
that
would
be
super
helpful
I
mean
the
more
eyes
on
this.
The
better-
and
this
is
really
this
is-
gets
a
lot
of
traction
so
for
better
for
worse
any
eyes
on
this,
so
they're
gonna
make
us
look
better
in
the
media,
which
is
always
nice,
so
yeah,
yeah
and.
A
N
We
are,
we
are
proceeding.
Well,
we
have
been
weeding
this
down
quite
steadily,
we've
gone
from
90
some
pages
to
70
some
pages,
and
we
expect
to
have
this
pretty
much
edited
down
by
Wednesday
Noah's
going
to
pull.
Do
one
last
pull
on
any
PRS
that
didn't
get
into
the
last
one
and
then
we'll
just
track
the
last
you
know
free
manually,
so
we're
doing
pretty
well,
the
known
issues
section
will
still
need
to
be
written
and
you
know
obviously
the
TLDR,
but
we're
we're
proceeding
nicely.
Everything
is
green.