►
From YouTube: Kubernetes Community Meeting 20190912
Description
The Kubernetes community meeting is intended to provide a holistic overview of community activities, critical release information, and governance updates. It also provides a forum for discussion of project-level concerns that might need a wider audience than a single special interest group (SIG).
See this page for more information! https://github.com/kubernetes/community/blob/master/events/community-meeting.md
Like what you see here? Continue the conversation on https://discuss.kubernetes.io
A
All
right
welcome
everybody
to
your
weekly
kubernetes
community
meeting
I
will
be
your
host
today,
George
Castro.
We
have
no
demo
today,
but
we're
gonna
do
a
quick
release,
status,
update
with
Jessica
from
the
release
team
and
then
we're
gonna
go
into
two
sig
updates.
Today,
it's
gonna
be
six
scalability
with
Wojtek
and
sig
network
and
sig
multi
cluster
sends
the
regrets
they
will
be
doing
their
updates
at
a
future
date
and
with
that
we're
gonna
move
on
to
this
week's
release,
update
Jeff,
take
it
away.
B
Hello,
everyone
I'm
here
to
give
your
lovely
1/16
release
updates.
So
this
week
on
Tuesday
we
cut
our
released
candidate
and
code
saw,
has
occurred
so
yay
to
that
the
cherry-pick
deadline
for
116
is
today.
So
if
you
have
something
you
need
to
cherry-pick
into
116,
please
get
that
in
today
and
if
they,
otherwise
it
is
not
going
to
be
in
the
release.
Everything
is
looking.
Green
CI
is
looking
great,
so
we
are
scheduled
to
release
on
Monday
September
16th
and
it
looks
like
we're
going
to
meet
that.
B
But
if
not,
we
will
be
letting
everyone
know
that
things
are
delayed,
but
things
are
looking
good.
As
for
patch
releases
for
our
site,
September
13th
is
the
cherry-pick
deadline
for
113,
11,
114,
7
and
115
for
and
they
have
target
release
dates
of
September
18th,
which
I
have
no
idea
what
day
that
is,
but
it
is
some
day
next
week
just
a
reminder.
All
of
these
things
pending
data
everything
we
try
to
blast
onto
kubernetes
dev
mailing
list,
and
that
is
it
for
me.
Does
anyone
have
any
questions.
C
Can
you
see
it
now
yep,
looking
good
okay,
so
so
I'm
Wojtek
I'm,
one
of
the
chairs
of
six
scalability,
so
I
will
give
a
quick
update
of
what
happened
in
last
quarter
last
release
or
this
release.
Actually,
that
is
getting
to
two.
We
are
getting
to
the
release
and
what
we
are
planning
to
do
next
obviously
feel
free
to
interrupt
me
if
you
have
any
questions
so
what
we
did
in
last
cycles.
So
there
are
like
three
main
group
free
right
top
to
our
themes
that
we
were
focusing
on.
C
One
was
like
extending
kubernetes
scalability
definition
like
we
are
basically
kubernetes
scalability.
We
are
defining
kubernetes
scalability
or
we
are
saying
that
the
cluster
skates,
if
a
certain
set
of
SLO
SL
SL
O's,
is
satisfied.
So
we
defined
in
the
previous
quarter's,
together
with
Signet
working
for
SL
ice
around
networking
I
should
probably
have
put
a
link
to
those
I
will
add
it
after
the
meeting
and
this
quarter
we
were
basically
working
on
implementing
those
and
measuring
those
in
tests.
It's
almost
done,
but
it's
not
fully
done
yet.
C
I
will
talk
like
what
we
are
planning
to
do
with
it
next
in
in
a
couple
seconds.
The
second
theme
was
like
basically
extending
scalability
testing
or
improving
scalability
testing
in
general.
So
one
of
the
things
that
is
happening
as
we
speak
is
extending
our
to
exercise
couple
more
objects
or
couple
more
kubernetes
concepts
like
things
that
we
will
we
didn't
test
at
all
before,
like
stateful
sets
or
persistent
volumes
and
stuff,
like
that,
we
also,
together
with
six
storage,
extended
or
created
a
bunch
of
tests
that
are
purely
like
storage,
related
and
measuring.
C
We
are
mostly
focused
on
like
startup
time
of
pods,
but
for
let's
call
them
stateful
pots,
so
those
that
are
mounting
pd's
or
other
kind
of
volumes
or
persistent
disk
or
other
kind
of
volumes
we
are
in
order
to
like
save
a
bunch
of
like
a
lot
of
money.
We
are
trying
to
move
migrate,
all
our
tests
to
private
clusters
so
that
not
every
node
has
a
public
IP
address
that
it
like
I,
can't
remember
the
exact
numbers,
but
it
translates
to
like
hundreds,
thousands
of
dollars
per
year.
Basically,
so
this
is
a
lot.
C
C
We
spend
a
lot
of
time
with
golilong
fault,
because
there
seem
to
be
another
regression
in
go
113
that
was
just
released
or
performance
regressions.
We
knew
it
before
they
released
like
roughly
two
months
before
they
released,
and
we
consciously
made
this
decision
with
them
to
release,
and
we
are
because
we
still
don't
fully
understand
the
problem,
but
we
are
pretty
close.
I
think
like
basically,
the
last
experiment
for
from
literally
yesterday
seems
to
be
very
promising.
We
are
trying
to
confirm
that
we
we
are,
that
it's
it's.
C
What
exactly
we
need,
but
I
believe
like
within
hopefully
week
or
two
weeks,
which
we
will
have
like
a
go,
verse
or
pot
release
version
of
go
on
13
that
we
will
be
able
to
use
yeah
and
if
the
third
theme
that
we
were
working
on
is
like
basically
improving
the
scalability
per
se
or
like
making
the
improvements
and
detecting
bottlenecks
and
stuff
like
that.
And
there
are.
There
are
couple
things
here
that
are
happening
or
happened.
C
So
from
the
theme
of
like
scaled
up
extending
scalability
definition,
we
are
going
to
finish
implementing
the
SL
eyes
that
we
have
for
networking
and
set
up
fresh
calls
to
make
the
SLS
from
them
and
promote
them
to
be
like
official
ones
for
on
which
will
be
also
which
will
be
enforcing
in
our
scalability
dense
test
and,
like
those
tests
are
in
the
end,
blocking
kubernetes
releases.
So
so
we
will
basically
introduce
those
those
new
s
allows
we
will.
C
C
I
was
already
discussing
that
with
like
couple
folks
and
it's
still
unclear
what
exactly
it
will
be.
What
I
my
mind,
though,
is
to
like
make
that
ask
hey
a
slight
weight
as
possible,
so
that,
on
one
hand
like
everyone
will
have
to
think
about
it.
On
the
other
hand,
like
the
overhead
from
that
will
be
really
necessary,
will
be
minimal.
C
So
that's
something
we
are
going
to
think
about
it.
Also.
We
are
also
going
to
extend
work
on
like
extending
our
scalability
test
and
performance
test
to
cover
more
concepts
and
introduce
other
types
of
tests,
like
cows,
testing,
testing
of
pathological
scenarios
and
stuff
like
that
and
a
bunch
of
work
in
terms
of
like
improvements
and
detecting
more
bottlenecks,
which
is
mostly
finishing
what
we
are
currently
doing
and
like
looking
for
more,
which
I
don't
have
anything
specific
on
my
mind.
Yet.
C
Anything
that
effects
other
six
I
think
there
are
like
three
main
things
here.
To
mention.
One
is
the
scalability
approval
process
that
I
just
mentioned,
like
once,
we
figure
out
some
proposal
that
we
will
think
it
makes
sense.
It
may
make
sense.
We
will
be
looking
for
like
actual
cat
owners,
ideally
in
new
caps,
so
that
we
will,
with
whom
we
would
like
to
like
validate
if
that
is
really
like
lightweight
process,
and
if
it
really
brings
some
value
both
for
the
cat
owners
and
for
sick
scalability.
C
So
we
will
probably
be
looking
for
such
people,
but
it's
not
something
we
are
ready
for.
Yet
the
other
thing
worth
mentioning
is
like
this
self
index
brocation
selling
is,
it
doesn't
seem
to
be
widely
used.
I
believe,
like
we
figured
out
all
the
places
and
it
would,
it
doesn't
seem
to
be
anything
that
is
that
may
that
it
that
this
duplication
make
us
problem.
C
Yeah
and
that's
that's
roughly
what
I
had
like
here,
two
links
with
like
issues
that
where
you
can
find
issues,
if
you
want
the
country,
contribute
in
this
caliber
DT
area.
If
there
isn't
anything,
you
are
really
interested
in,
but
you
want
to
do
something
like
feel
free
to
also
ping
us
or
me
personally,
on
slack
yeah
and
Kirk.
Here
is.
The
here
are
some
links
for
slack
channel
mailing
list
and
stuff
like
that
soon
yeah,
that's
what
I
had
do
you
have
any
questions.
D
Hello,
yeah,
okay,
you
all
right,
hi,
I'm,
Bowie
I'm,
giving
the
update
for
a
sick
Network.
So
we
had
all
two
big
features
that
came
in
this
cycle
on
116,
so
the
first
one
is
dual
stack
support,
so
now
kubernetes
is
done
with
phase
two.
You
can
find
the
details
in
the
cap,
but
basically
we
made
a
lot
of
API
and
command-line
changes
in
classic
computer
science
fashion.
Much
of
this
work
was
taking
a
singular
thing
and
making
a
plural
and,
of
course,
this
results
in
a
lot
of
changes.
D
The
other
major
thing
that
happened
in
116
is
the
addition
of
a
new
discovery
API.
So
one
of
the
motivations
for
this
new
API
is
that
the
current
endpoints
API
does
not
scale
above
a
certain
amount.
We
found
around
1k
depends
on
your
usage
or
a
really
endpoint
slice
allows
you
to
scale
to
massive
services
way
beyond
this,
and
actually
there's
a
blog
post.
D
So
network
policy
and
ingress
have
now
been
moved
out
of
the
networking
API
group
and
please
update
your
yeah
muls
uyy,
because
these
things
I
think
in
116
Network
policy
will
stop
being
served
so
that
will
just
stop
working
and
then
ingress
will
follow
suit,
probably
in
117,
so
please
whereat,
deprecation,
117
and
then
stopping
service
sometime
afterwards.
So
please
update
your
configurations.
D
Finally,
we
had
a
bunch
of
version
updates
and
bug
fixes.
For
example,
services
now
have
finalized
errs
for
the
resources,
various
proxy
or
fixes
and
on
PBS,
and
we
have
a
new
core
dns
version
in
this
release,
and
many
many
thanks
to
all
the
contributors.
The
cycle
awesome
work,
especially
the
heavy
stuff
like
Ivy
via
ipv6,
and
then
all
the
bug
fixes,
of
course,
so
like
kind
of
reviewing
the
major
themes
for
2019,
we
at
the
beginning,
I
think
January.
D
We
said
we're
going
to
look
into
ingress,
l7
revamp,
we're
definitely
doing
dual
stack,
we're
looking
into
topology
service
endpoint
to
revamp
windows
and
scalability,
and
we
have
made
progress.
Although
116
we
I,
don't
think
we
got
to
a
state
of
wrapping
up
in
the
GA
for
ingress
dual
stack,
awesome
phase
2
is
done
onwards
to
the
next
two
phases.
I
think
that
still
remain.
We
made
progress
in
topology,
but
nothing
in
116,
and
then
we
have
made
progress
in
scalability
and
endpoints
with
the
new
discovery
API
and
the
Windows
support
it
works.
D
We
went
I
think
we
went
through
all
the
tests
and
tried
to
figure
out
which
ones
for
Windows
applicable
and
which
ones
were
not
looking
forward.
We,
the
city,
has
not
done
planning
for
2020,
but
given
the
sort
of
conversations
that
have
been
on
these
sake,
two
things
pop
out.
One
of
them
is
that
cube
proxy
thanks
to
Valerie
for
driving.
A
lot
of
this
is
key
proxy
component
cessation,
which
is
taking
cube
proxy
breaking
up
into
its
constituent
pieces,
some
pieces.
D
They
actually
don't
interact
with
the
other
but
they're
in
the
same
repo
movie
into
a
separate
repo,
and
this
will
allow
faster
velocity
and
development
on
that
sentiment.
And
then
we
have
a
lot
of
conversations
about
networking
and
multi
cluster,
so
that
one
is
an
interesting
one.
We'll
see
what
happens
in
2020
in
terms
of
work
with
other
SIG's
I
think
definitely
was
ironic.
D
That
Wojtek
went
before
me,
but
we
work
a
lot
with
scalability
for
some
of
these
things
and
I
think
ipv6,
probably
adults
that
will
have
some
impact
on
users
in
general,
because
you
know
IP
addresses
are
pernicious
across
all
sorts
of
configurations
in
terms
of
where
to
find
us,
that's
the
zoom
link
and
we
have
meetings
every
Thursday
at
2
p.m.
PST
and
then
those
are
links
to
our
slack
and
our
community
pages
and
that's
any
questions
for
networking.
A
Alright
great
thanks
a
lot
if
you're
joining
late
Sigma
fleet
cluster
Paul
sends
his
regrets,
he
won't
be
able
to
go
today,
so
they'll
go
at
a
future
date.
Okay,
moving
on
to
announcements,
real
quick
election
updates,
the
candidates
platforms
have
all
been
checked
into
github
I,
put
a
link
in
the
notes.
Please
feel
free
to
read
all
their
BIOS
there
to
determine
who
you
want
to
vote,
for,
we
will
be
kicking
off
the
election.
Next
Wednesday
and
I
will
be
sending
the
status
update
mail
to
the
kubernetes
dev
mailing
list.
A
So
make
sure
you
do
that
we
have
about
a
two-week
voting
period
where
voting
will
be
open
and
you'll
in
the
status
emails
you'll
have
the
rest
of
the
little
schedule
there
how
the
election
is
going
to
go.
We're
doing
a
call
for
demos.
For
this
meeting,
we
like
to
do
a
quick,
10
minute
demo
at
the
beginning
of
the
meeting.
So
if
you're
working
on
something
cool,
you
want
to
do
a
demo
see
the
instructions
at
the
top
of
the
notes
document
for
how
to
sign
up
the
contributor
summit.
Registration
is
live.
A
This
is
the
event
that
happens
the
day
before
cube
con
and
cloud
native
con
that
has
a
current
contributor
track
and
a
new
contributor
track.
Those
registrations
always
sell
out
quickly.
So
we
encourage
you
to
sign
up
and
I've
put
the
link
there
for
you
and
you
could
check
out
hash
contributor
summit
on
the
kubernetes
slack.
If
you
have
questions
about
that,
just
a
note
about
this
meeting
next
week
will
be
the
release
retrospective
with
with
the
release
team.
A
Quick
announcements
here,
new
users
I,
need
your
help
on
discussed
at
kubernetes
I/o,
getting
a
ton
of
new
users
asking
a
lot
of
questions.
If
you
go
in
you
help
out,
you
do
something:
cool
I'll,
send
you
a
kubernetes
t-shirt
it'll,
be
great
and
you
can
help
me
out
by
RTing
the
announcement
all
that
and
helping
spread
the
word
through
your
social
circles.
I
would
appreciate
that
in
two
weeks
after
the
release,
retro
six
CLI
working
group
policy
and
working
group
components
standard
will
be
given
their
status
and
Tim
pepper
will
be
your
host.
A
This
is
the
first
time
we're
having
workgroups
kind
of
give
a
status.
So
we're
gonna
see
how
that
works
out.
Quick,
shout
outs
this
week.
This
is
shut
up
channels
on
the
kubernetes
slack.
If
you
see
someone
doing
something
really
great
above
and
beyond,
the
call
of
duty
feel
free
to
mention
them,
and
that
will
make
sure
they
get
a
shout
out
this
week.
A
So
Vince
would
like
to
do
a
huge
shout
out
and
thank
you
too
Catherine
berry,
Tim,
Hawking
and
Jason
D
siberius
for
spending
hours
together
to
get
closer
API
container,
build
automation
up
and
running.
Zachary
Sarah
would
like
to
shout
out
to
Junyi
for
adding
a
helper
script
for
localization
teams
to
discover
changes
to
source
stocks
between
development
branches,
and
we
put
the
link
to
the
pull
request
there
and
mathy
x
would
like
to
shout
out
to
Joe
Beda
Brendan
burns.
A
Tim
Hawking
and
app
are
enough
for
the
nice
TechCrunch
panel
session
and
there's
a
little
link
to
their
YouTube
video
and
bendy
Eldar.
Lastly,
would
like
to
close
it
out
by
shouting
out
the
Jordan
Liggett
for
reviewing
all
the
things,
because,
of
course
he
is,
does
anybody
have
any
other
announcements
or
things
that
they
would
like
to
tack
on
to
the
meeting.