►
From YouTube: Kubernetes 1.13 Release Team Meeting 20181114
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
Okay,
hello:
everyone
welcome
to
113
burndown
I've
linked
the
meeting
minutes
in
the
chat.
So
please
add
your
attendance
if
you
haven't
already
and
the
meeting
is
being
recorded
and
it'll
go
on
to
YouTube
shortly
so
the
standard
disclaimer,
please
be
mindful
of
what
you
say
with
that:
let's
get
started
so
we
are
almost.
A
Steadily
towards
code
freeze,
which
is
slated
for
this
Friday
and-
and
that
is
also
a
beta
one
cut
so
far-
the
release
I
would
say
if
around
the
yellow
status,
mainly
because
we
are
still
waiting
for
major
enhancements
to
merge
and
then
stabilize
and
look
at
a
CI
post
code
freeze
as
of
yesterday,
the
CI
signals
are
looking
good
and
hence
I'm
thinking.
We
will
be
on
track
to
cut
our
beta.
A
That
said
the
main
agenda
for
today
is
we
have
Michelle
from
six
storage
and
patrick
from
windows,
mainly
to
give
an
update
on
the
two
main
enhancements
that
are
slated
to
go
GA.
This
release,
so
I
think
we
I'd
like
to
get
started
with
them
and
once
they
finish,
we
can
go
on
to
a
standard,
rigid
agenda
for
the
day.
So
Michelle
do
you
want
to
give
us
quick
status
on
one
seven?
Eight,
yes
sounds.
B
Good,
so
we
so,
with
only
a
few
days
left
before
feature
freeze,
we
decided
that
we
need
to
focus
all
of
our
efforts
on
getting
on
the
work
to
get
CSI
to
GA
and
all
the
other
CSI
related
work
is
going
to
be
dropped
for
this
release,
so
that
includes
this
596
CSI.
In
my
volume
support.
Oh,
that
one's
going
to
be.
That
was
my
question.
Okay,
yeah.
A
B
B
B
B
A
B
B
C
B
A
A
B
D
A
B
B
A
Makes
sense
and
then
my
only
question
was
once
all
this
is
done
in
master
I
mean
the
spec
moved
to
1.0
and
all
of
that
and
when
sui
falls
forward,
we
expect
those
changes
to
work
in
the
1.13
branch.
There's
nothing
else
that
needs
to
happen
in
terms
of
updating
dependencies
for
that
branch.
Yeah.
That's
correct!
Ok,
ok!
So
from
from
the
CI
signal
side,
everything
should
stay
stable.
Yes,
ok,
so
I'm
going
to
quickly
go
back
to
that
178
where
we
were
tracking
the
list
mission
effect
in
that
I
think.
A
B
A
B
Tracking
yeah,
so
we
added
some
tests
to
the
alpha
suede.
They
ended
up
failing
so
we
temporarily
disabled
them,
but
we
have.
We
have
to
pee
ours
out
to
adjust
two
out
of
the
three
failures
and
we're
making
progress
triaging.
The
third
failure,
so
I
think
we
should
be
able
to
get
a
fix
out
before
code
freeze
and
then
we
can
re-enable
those
tests.
Yes,
for.
A
E
B
A
So,
okay,
so
at
this
point
we
have
about
four
five
one
yeah.
We
have
about
three
and
then
four
five.
Five.
We
are
five
issues
at
least
open
and
then
there's
the
docs
one
is
Docs.
Is
that
is
one
zero,
eight,
nine
three,
the
only
Doc's
PR
yeah?
Okay.
So
how
confident
are
you
I
mean
to
make
a
go/no-go
call
today
or
how
confident
is
the
sake
it
self
to
get
all
these
wrangled
in
by
Friday.
So.
B
A
And
in
case
in
case
there
is
some
delay
in
any
of
these,
going
in
would
be
be
able
to
if
things
spill
into
code
freeze-
and
we
don't
want
that-
is
that
a
way
to
feature
I
mean
will
make
the
entire
thing
stay
in
stable
are
in
beta
versus
going
to
GA.
Is
there
a
feature
gate
yeah?
We
can
always
flip
that
the
future
gay.
A
D
A
B
So
I
kind
of
gave
an
update
for
that
sort
of
earlier
there's
two
PRS
pending
for
the
couplet
plug-in
registration,
one
is
to
move
the
feature
to
GA
and
the
other
one
is
to
fix
a
bug.
I
think
the
the
PR
to
move
the
feature
to
GA
should
be
able
to
go
in
today.
All
it
needs
is
a
final
approval
and
then
the
bug
fix
I
think
will
should
be
able
to
go
in
by
either
two
tomorrow.
Okay,.
B
B
A
B
A
A
B
A
B
A
B
A
A
B
A
A
Machine
anybody
else
has
any
questions
for
Michelle,
Rock,
triage
enhancements
or
CI.
B
A
A
A
Do
you
wanna
start
off
by
just
giving
ya
the
state
and
then
I
can
follow
up
with
any
questions.
Yeah.
G
A
G
So
basically,
I've
got
a
lot
of
columns
here.
Most
of
these
don't
matter
other
than
these,
these
first
three
here
so
basically,
you
know:
we've
proposed
to
take
our
feature
tab
and
it
enable
Windows
nodes
and
kubernetes
to
stable
for
1.13,
and
so
that's
what
this
top
work
item
is
currently
tracking
here,
underneath
this
I
have
a
list
of
the
other
work
in
progress
tracking
stuff.
That
needs
to
be
done
with
like
an
organizational
and
doc
perspective.
Okay,
and
so
you
know,
one
of
these
is
basically
some
API
related
documentation
that
I'll
be
updating
promised.
G
G
Yeah
and
then
so
this
this
next
thing
we're
looking
at
is
actually
I
mean
be
easier
for
describe
these
two
things.
First
right
now,
so
we've
got
two
pr's
that
are
going
to
hook
up
the
subset
of
Windows
tests
that
we're
going
to
be
running
over
and
tests
for
it,
and
so
I
started
talking
with
with
Benjamin
elder
yesterday
and
I'm
gonna
push
on
him
to
merge
those
one
specific
question
I
have
about
this
is
for
the
testing
for
repo.
Do
you
need
any
sort
of
like
priority
or
other
tags
on
there?
A
G
G
A
G
Handing
that
stuff
off
today
so
with
just
a
little
bit
of
background
here.
This
also
is
going
to
be
scheduling.
Tests
on
the
Azure
cloud
and
I
got
an
account
set
up
for
the
CNC
F
just
yesterday,
and
so
it's
just
like
yeah
I
could
either
sleep
or
not
at
this
point,
so
so
I'm
getting
that
stuff
handed
off
there.
G
We've
already
got
a
separate
prowl
instance.
That's
been
that's
been
running
so
other
than
you
know,
issues
bringing
that
online.
That's
mostly
done.
Ok.
Last
group
of
work
that's
happening
right
now
is
we're
still
continuing
to
triage
a
few
of
the
tests
after
what's
going
on
now,
is
that
originally
we
had
a
proposal?
That's
been
moved
into
the
postponed
column
here,
where
we
were
going
to
have
the
conformance
tests
make
a
decision
on
how
they
were
going
to
run
based
on
whether
or
not
Windows
notes
are
priced
in
the
cluster.
G
Over
the
last
three
weeks
in
discussion
with
people
in
the
conformance
working
group
and
cig
architecture,
they
have
basically
tabled
a
decision
and
they're
not
going
to
accept
any
changes
for
thirteen
and
so
they've,
basically
separated
the
process
of
defining
what
conformance
means
and
the
context
multiple
operating
systems
from
us
defining.
What
the
signal
is
to
make
Windows
stable
and
so
the
general
test
approach
that
were
following
is
or
making
a
best
effort
to
make
all
the
conformance
tests
passed.
G
E
G
G
A
G
A
G
I'm
also
working
with
you,
jus
long
and
Peter
Hornak
over
at
Google,
because
they're
also
doing
the
same
thing
on
virtual
machines
that
are
running
in
GCE,
so
that,
let's
just
expect
it
to
be
the
same
regardless
of
cloud
provider.
So
it's
just
sort
of
a
you
know,
hidden
finish
on
that
on
that
final
list,
based
on
what
code
landed
in
what
code
was
postponed.
Okay,.
B
G
A
G
It's
all
gonna
depend
on
if
I
can
get
enough
time
to
get
those
two
PRS
I
mentioned
merged.
Otherwise
I
could
share
logs
that
are
already
running
from
the
for
another
branch.
I
competed
point
you
at
the
one
that
you
jus
has
running.
So
you
can
see.
You
know
multiple
test
results
there
and
I
can
give
a
copy
of
logs
from
from
a
juror
it's
just.
They
won't
be
on
test
grid
until
those
two
PRS
are
merged
and
are.
A
G
A
A
Okay
and
I
will
follow
up
on
that
and
Cole
here.
The
call
is
our
test
and
Frawley
for
113.
So
if
you
need
any
any
help
with
with
the
test
grade
or
any
of
interest
off,
he
can
also
help
you
there
yeah
cool
I've
added
those
to
be
ours,
that
to
test
entropy
has
to
do
the
make
minutes
here
in
case
you
need
to.
E
A
A
D
Absolutely
we're
talking
about
one
offs
right,
I
mean
given
you
know
we're
introducing
a
new
optional
picture
into
kubernetes.
So
you
know
Windows
is
not
enabled
by
default.
Mm-Hmm,
that's
a
big
thing
for
our
for
our
specific
use
case,
but
you
know
you
know
like
Patrick
mentioned
I,
don't
think
he's
sleeping
much
lately
trying
to
get
it.
I've
been
running
and
migrating
them
all
into
the
new
test
infrastructure
and
getting
everything
green
chillis
to
bother
who
may
stumble
on
a
couple
of
things.
D
So
you
know
we'll
work
with
you
guys
to
make
sure
the
amount
of
changes
is
super,
meaning
I'm,
just
hoping
that
if
we
do
have
to
fix
a
couple
of
test
cases
post
the
16th
that
we
are
gonna
give
in
a
little
bit
of
latitude.
Just
because
of
the
fact
that
these
features
is
a
know
making
code
changes
are
making
stabilization
test
changes.
Yeah.
A
G
The
other
thing
I
want
to
mention
is
that
the
stuff?
That's
in
a
test,
triage
column.
My
board
are
currently
things
we
believe
will
be
fixed
their
bugs
in
the
Windows
version
of
the
test.
Containers
are
not
bugs
in
the
cubelet
or
kubernetes
code
itself,
okay,
so
just
a
little
bit
of
background
here,
you
know
the
current
test
approach
that
that
we
fought
was
trying
to
make
windows
containers
act
like
Linux
containers
and
so
in
many
cases
we're
using
things
like
busybox
to
give
output
that
look
like
bash
on
Linux.
G
That's
not
actually
something
that's
built
into
Windows
and
so
over
time,
rather
than
trying
to
you
know,
do
things
that
sort
of
simulate
the
Linux
behavior.
One
of
the
expectations
that
I'm
expecting
for
conformance
is
that
we're
going
to
actually
rip
those
out
and
replace
them
with
the
same
tools
at
a
Windows
customer
would
years
so
the
test
case
says
secrets
should
be
read
only
then
we'll
test
that
with
the
Windows
command.
So
that
way,
when
an
admin
looks
at
the
test
output,
it
matches
their
expectation.
G
A
So
this
I
think
they're
already
over
time,
but
I
can
sync
with
you
offline
too,
because
this
dashboard
is
more
of
conformance
I
thought
you
had
another
dashboard
which
ran
the
feature
test
itself,
which
gave
you
signals
that
were
not
integrated,
but
I
can
follow
up
with
the
offline
to
see
if
those
test
results
are
exported
anywhere.
Yes,.
G
So
we
were
you'd,
we
were
using
the
conformance
tests
as
the
bar,
but
now,
based
on
the
recent
discussions,
we've
been
told.
Let's
change
the
approach
here
and
so
that's
a
relatively
new
development
that
happened
to
close
to
two
code
freeze
for
us
to
go
through
and
copy
and
paste
these
and
mark
and
feature
tests
instead
of
conformance
test,
we're
going
with
the
subset
of
performance
approach.
Okay,.
A
G
A
Thank
you
I
think
we
will
I'll
synch
up
offline
and
then
update
on
the
on
the
burndown.
Later
with
that,
we
are
almost
over
time
going
back
to
the
standard
agenda.
Let's,
if
any
of
I'll
quickly
go
through
the
sections
and
if
any
of
the
leads
you
have
anything
critical
to
call
out,
go
on,
go
ahead,
Kendrick,
anything
to
add
for
enhancements.
A
A
H
E
Awesome
so
things
are
looking
a
little
bit
less
bonkers
compared
to
Monday.
The
the
main
white
theme
is
the
112
to
113
upgrade
downgrade
tests,
but
they
are
successful
in
master.
So
the
hope
is
that
once
the
you
fracturing
that's
related
to
testing
from
makes
its
way
to
113,
we
should
be
seeing
good.
E
A
I
think
at
least
the
BBC,
if
attached
flakes
mission
spoke
about
and
they
are
looking
into
it
and
the
flake
me
up
greatest
yeah.
Things
have
been
green,
at
least
since
yesterday,
but
I
yeah
I
understand
the
concern
that
it
could
flake
at
any
time
throwing
us
off.
Let's
follow
up
offline
on
those
issues
to
see
if,
if
the
owners
can
do
anything
to
deflate
them
in
the
meantime,
and
talking
about
the
upgrade
I
think
cool,
you
already
are
on
top
of
the
test.
Infra
PR
force
political
impairing
job.
C
Yeah
so
I've
been
started
doing
that
yet,
but
I
took
a
look
I'm,
assuming
that
all
the
arguments
for
the
the
113
jobs
are
gonna,
be
pretty
similar
to
the
master.
Jobs
haven't
actually
looked
at
that
yet,
but
if
they're
not
then
I
know
the
people
to
contact,
so
that
should
be
pretty
straightforward.
Awesome.
A
Yeah
to
just
to
give
context
for
others,
we
had
the
master
upgrade
the
GC
master,
retreat
job
timing
out
on
on
both
the
master
and
113
branch.
Last
I
mean
beginning
of
this
week.
We
kind
of
route
cost
it
to
being
the
HPA
chests
the
horizontal,
auto
scaling
that
got
added
increased
the
time
it
took
to
be
able
to
be
more
than
15
hours.
So
one
approach
just
ended.
A
He
split
the
the
job
itself
into
parallel
and
non
parallel
jobs
so
that
we
could
buy
some
more
time
for
the
non
parallel
test
to
pass
and
after
that,
both
the
job
started
passing.
So
we
kind
of
want
to
do
this
similar
split
for
the
1:13
branch
as
well,
so
that
we
can
see
if
those
turned
green
or,
if
not,
it
has
any
other
issue
that
the
timeout
is
masking
so
yeah
moving
on
to
issues.
Anyone
else
I
mean
I'm,
Nico,
quick,
any
update
there,
hello.
F
G
G
F
G
F
A
F
A
D
F
Yeah,
all
of
these
are
pretty
much
close
to
merging
oh
yeah,
there's
a
last
one
which
hasn't
been
updated
for
several
days
and
well
helping
one
more
once
more
to
the
owner.
There's
several
yards
quite
big
I'm,
not
sure,
and
they
have
some
API
changes
and
it's
not
so
I'm,
not
sure
if
it
would
make
it
this.
A
A
Awesome
and
then
release
notes
Mike
is
out
today,
but
his
update.
We,
we
sent
the
release,
notes
out
to
Leeds
on
Monday
and
folks
have
been
started,
filling
some
major
themes
and
the
he
Mike
and
shadows
plan
to
follow
up
throughout
this
week
and
next
week
and
I
don't
know
if
we
have
Caitlyn
on
the
call
for
any
concept
dates.
A
If
not
yeah,
the
only
announcement
is
the
retro
talk
is
up.
It's
been
up
for
a
while
now
so
as
we
kind
of
come
to
the
rear
end
of
the
whole
release
cycle.
If
folks
want
to
take
some
time
to
add
what
went
well,
what
you
think
could
have
been
improved,
that'll
be
great,
so
I
just
wanted
to
raise
that
here
and
I've
added
a
link
to
the
retro
doc
cool
with
that
I
think
we
are.
We
are
at
the
end
of
today's
one
down.