►
From YouTube: Kubernetes Community Meeting 20190905
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
A
All
righty
this
is
the
kubernetes
community
meeting.
This
is
a
livestream
event.
That's
going
to
be
posted
publicly
on
YouTube,
so
please
be
mindful
of
anything
that
you
say
is
being
recorded.
I
do
ask
that
everybody
just
give
their
lines.
What
you're,
not
speaking.
We
do
also
have
a
code
of
conduct,
so
please
be
excellent
to
each
other.
Before
we
start
can
I
get
a
note-taker.
I
am
handicapped,
so
I
will
not
be
typing.
Also.
A
C
A
Volunteer
as
note-taker,
oh
thank
you,
sir
thank
you
alrighty.
So
let's
go
ahead
and
get
started.
I
am
going
to
go
over
the
agenda.
Just
so
anybody
doesn't
know
I'm.
This
is
the
first
time
I'm
doing
it
so
I'm
gonna
botch
it
horribly
and
I.
Thank
you
or
giving
me
this
opportunity.
So
you
guys
know
who
I
am.
We
know
who
the
note-taker
is.
We
are
going
to
have
a
demo
from
our
google
Summer
of
Code
student
aza
will
be
going
over
the
release
updates
then
we'll
be
doing
sig
updates
and
then
general
announcements.
A
D
I'm
Elijah
I
I
was
I,
worked
on
the
kubernetes
dashboard
from
me
to
the
handing
of
August
so
had
support
for
custom
resource
definitions
to
the
kubernetes
dashboard,
so
add,
tomatoes,
machine
and
Sebastian,
and
the
demo
work
for
loot
saying
so.
The
what
we
set
out
to
do
was
to
so
originally
the
kubernetes
dashboard
supported
battery
sources,
but
when
they
were
replaced
in
Cuban,
it
is
by
C
by
C
artist
being
sick
did.
Cpr
support
was
removed
from
dashboard,
so
for
a
while
there
wasn't
really
any
support
for
extensions.
D
Then
we
also
wanted
support
to
be
able
to
edit
and
the
let's
see
how
this
right
from
the
dashboard,
then
some
additional
features
that
we
wanted
to
have
was
supporting
sub
resources
or
see
how
this
especially
the
skills
of
a
shot.
So
on
the
key
behind
this
dashboard,
the
previously
you
could
already
scale
with
deployments.
D
Usually
there
is
scale
deployment
so
because
I'm
Sierra
this
also
have
the
ability
to
support
the
scale
sub
resource.
We
also
wanted
to
it's
great
support
for
scaling,
so
the
dashboard,
and
then
we
also
wanted
to
paint
see
artists
to
the
sidebar.
So
we
decided
to
paint
share
this
because
or
me
originally,
the
plan
was
login
to
the
key
branches
dashboard
and
then
all
these
artists
in
the
course
I
would
be
displayed
in
the
sidebar.
But
we
discussed
that
approach
and
discovered
that
some
clusters
might
have
a
large
number
of
Sierra
days.
D
So,
instead
of
displaying
everything
at
once,
we
could
just
show
a
single
many
items
for
customers,
definitions
that
would
allow
users
to
pin
penny
charities,
as
might
wish
so
some
of
the
mock-ups
that
were
originally
designed
for
the
charity
list
page.
You
can
see
this,
who
is
the
charity
showing
a
list,
a
list
of
objects
for
this
charity,
and
this
is
this-
is
an
audacity.
D
D
So
yeah,
you
can
see
some
of
the
actions
that
the
action
section
here
so,
of
course,
during
the
implementation.
They
will
also
think
other
things
that
we
had
at,
for
example,
pinning
sub
resources.
We
had
to
have
a
separate
icon
right
here
and
then,
when
you're
scaling
objects
also,
we
had
a
separate
action
action
icon
here
as
well,
so
now
for
the
demo,
I
have
the
Cuban.
It
is
dashboard
open
in
my
browser
already.
D
Right
so
just
quit
at
three
different
Sierra
T's,
then
it
displayed
it
displays
the
some
basic
information
for
charities
year,
including
the
full
name,
the
group,
and
whether
or
not
the
name,
space
or
cluster
scoped.
So
going
to
this
year
the
details
page,
we
can
see
the
research
information
section,
the
names
that
have
been
accepted
for
this
charity
and
currently
there
there
is
no
object
for
this
particular
showed
again.
We
can
see
conditions
as
well.
D
D
So
that
works
mine
we
can.
We
can
see
that
there's
no
object
for
this
particular
charity
in
the
mode.
Then
the
hiking
opinion
charities
is
obviously
action
bar
so
clicking
on
this
instance.
30
to
the
background
is
persisted
in
a
config
map,
so
it's
close
to
white,
it
persisted
was
reloads
and
then
a
perform.
Apart
from
the
action
bar
action
icon
subpoena
charities,
we
also
have
a
separate
menu
item
to
pain
and
on
pain,
and
then
skilling
also
works.
We
have
a
skill
in
you:
have
a
soul
food
for
CR.
D
It
is
the
support,
skill,
the
skill
services,
for
example.
It's
going
to
show
up
on
that
users.
Information
and
the
contract,
for
example,
supports
scaling
so
clicking
on
the
menu
clicking
on
the
menu
drop-down
for
this
particular
object.
For
example,
we
can
see
the
skill
pop
up
and
we
can
select
a
particular
skill
replica
number,
so
there
isn't
any
any
controller
power.
It
is
particular
charity.
So
it's
it's.
D
It's
not
spin
anything
up
or
anything,
but
we
can
see
that
it
is
ad,
but
the
desire
of
the
best
section
is
shut
up
there
and
then
we
can
delete
and
perform
a
bunch
of
for
the
actions
here
as
well.
So
that's
that
has
been
disco
of
supporting
charities
and
we
op
the
committee
and
test
out
this,
but
the
graphical
and
give
us
some
helpful
feedback.
We
got
in
future
implementation.
Thank
you.
A
D
Yet
so
my
experience
was
was
really
great.
I
was
communicating
regularly,
with
my
mentors
on
slack
and
for
a
while
I
started,
contributing
to
keeping
it
is
around
February
I
start
opening
a
few
pull
requests
against
the
Cuban.
It
is
dashboard
and
so
I
got
I
got
to
know
some
of
the
cichlids
and
get
familiar
with
the
general
process
of
contributing.
So
when
the
program
it
started
me,
I
was
a
bit
familiar
with
the
process,
so
I
are
getting
from
there.
It
could
be
as
well
so
contributing
was
really
straightforward.
D
My
methods
were
very
helpful
because
I
had
any
problems,
then
I
add
there
were
holdup
people
in
the
community
hall,
so
dad's,
for
example,
Stefan.
It's
so
I.
Could
you
know
just
chat,
shut
up
a
few
people
and
ask
for
help
when
the
vamps
talk
so
so
generally,
the
committee
was
very
helpful
and
I
learned
a
lot
and
it
helped
me
improve
my
skill
and
also
land
the
job,
so
I
started
working
with
with
reflects
that.
I
will
just
you
know
two
weeks
ago,
so
hidden
cabinets
is
monitoring
solutions.
D
A
A
E
E
It's
all
good
I,
just
like
the
hearing
people
struggle
with
it,
it's
good
fun.
My
parents
do
too
that's
why
they
got
to
me
wonderful,
Thank,
You,
Maki,
okay,
so
release
updates
hi,
my
name's
Lachlan
I'm,
the
116
release
lead
I'm,
going
to
give
you
some
updates
on
116,
which
is
the
release
that's
coming
out
in
two
weeks
time.
So
the
the
delivery
date
for
116
is
slated
at
September
16th,
which
is
not
met
next
Monday,
the
Monday.
After
a
couple
of
deadlines
coming
up
that
you
should
be
aware
of.
E
E
So
they
are
the
main
three
milestones
coming
up
next
week
for
1.16
in
terms
of
patch
releases
for
the
current
supported
releases,
which
is
113,
114
and
115,
we
are
targeting
the
next
set
of
patches
to
go
out
just
after
116
zero.
So
the
same
week
as
916,
we
don't
have
a
date,
yet
it's
probably
the
Thursday
of
that
week,
so
that
would
be
one
dot:
13
11
1,
dot,
14.7
and
1.15
dot
for
going
out
the
week
of
September
16th,
which
means
the
cherry-pick
deadline.
E
E
Tim
I
see
your
cursor
going
for
putting
those
updates
in
there
very
much
appreciated,
and
if
you
want
to
stay
up
to
date
on
the
latest
releases
and
when
they
come
out
and
things
like
CBE's
and
fixes
that
fix
CVS
go
ahead
and
sign
yourself
up
to
the
kubernetes
dead
as
mailing
list,
and
you
will
be
the
first
people
to
know
once
that
happens,
that
is
it
for
release.
Updates.
I
will
flip
it
back
over
to
Marky.
Thanks
for
having
me
thank.
A
You
very
much
that
the
release
team
I'm
gonna,
give
you
a
broken
arm,
clap
here.
The
release
team
is
no,
that's
no
small
feat
there.
So
I
really
appreciate
now
and
I'm
sure
the
entire
community
does
all
the
work
you
guys
put
in
like
I
said.
That
is
its
massive
there's,
a
lot
of
things
to
deal
with.
So
thank
you
very,
very,
very
much
for
all
that
you
guys
do
cool
moving
right
along.
We
are
going
to
go
into
the
sig
updates.
A
F
Hi
everybody
I
mean
go
ahead
and
share
my
screen
here
soon.
You
guys
can
all
see
it.
Yes,
all
right
cool,
so
my
name
is
Michael
Michael
I'm,
one
of
the
co-chairs
for
sig
windows,
along
with
Patrick,
Lang
and
I'll,
give
you
guys
a
quick
update
today
and
also
do
a
demo.
So,
first
things
off.
We
want
to
basically
recognize
and
congratulate
two
individuals.
F
I
have
had
significant
contributions
to
the
sig
Windows
community
for
the
past
one
year,
plus
that's
deep,
deploy
and
Ben
Moss
recognizing
them
and
making
them
as
technical
liaison
sig
windows,
so
they
can
start
reviewing
and
approving
pull
requests
for
each
other
as
well,
as
you
know,
continuing
their
technical
excellence
within
our
community.
Congratulations
guys!
Thank
you.
F
So,
let's
talk
a
little
bit
about
1.16.
Ever
since
we
won
GA
with
1.14
have
been
working
on
improving
the
end-user
experience
around
windows,
and
that
includes
a
wide
variety
of
areas,
starting
with
enabling
Windows
users
to
leverage
better
identity
options,
so
that
containers
can
access
other
external
resources
within
the
network,
communicate
databases,
commuity
file,
servers.
What
is
in
other
services
within
that
environment,
and
in
order
to
be
able
to
do
that,
we
had
to
enable
two
specific
key
technologies.
The
one
is
called
group
managed
service
accounts
which
were
graduating
to
beta
with
1.16.
F
So
congratulations
to
the
team
for
moving
that
along
the
further
down
the
path
and
then
also
introducing
a
new
capability
called
Rani's
username.
That
will
enable
the
initial
process
that
kicks
there's
the
container
to
carry
a
specific
identity
as
well.
So
both
of
those
are
Windows
identity
options.
Windows
developers
are
expecting
these
capabilities.
They're
used
to
using
Active
Directory
identity
when
they're,
creating
workloads
and
applications
so
enabling
this
in
kubernetes
is
a
key,
is
a
key
capability.
F
We're
also
announcing
alpha
support
for
cube.
Adm
join
the
Windows
community,
along
with
significant
contributions
from
the
C
cluster
lifecycle.
Luba
Mir
made
significant
contributions
here,
along
with
Kayla
and
Ben
mass.
Their
ovens
were
introducing
a
certain
set
of
scripts
that
would
allow
that
will
be
maintained
by
the
cylinders
community
and
they
will
allow
end
users
to
install
all
the
prerequisites
on
CNI,
so
they
can
stand
up
in
prepare
a
node,
so
it
is
ready
to
be
joined
to
the
environment
and
then
we
other
support
for
cube
ADM
joints.
F
You
can
add
the
windows
note
to
the
cluster.
You
can
get
it
to
the
ready
state
and
it
was
ready
to
start
accepting
workloads.
This
is
a
significant
step
forward
towards
making
it
easier
for
users
to
stand
up
hybrid,
heterogeneous
cluster,
we're
also
introduced.
As
for
support
for
CSI,
you
know
a
lot
of
users
ask
us
about
what
are
storage
options
around
windows
and
so
far
we
didn't
really
have
a
good
story.
F
We
supported
flex
volume
and
a
couple
of
other
entry
providers
with
alpha
support
for
CSI,
allow
end
users
to
leverage
persistent
storage
options
for
Windows
containers
as
well,
and
we've
been
able
to
do
this.
We
had
to
bypass
it
the
capabilities
and
use
a
host
OS
proxy,
because
windows
containers
don't
have
privilege
a
privilege,
execution
privileges
on
the
windows.
F
Node,
so
we
had
to
use
a
host
frogs
or
as
proxy
to
kind
of
bypass
the
requirement,
but
great
work
by
the
community
on
all
three
areas
here
were
very
excited
for
one
to
sixteen,
and
we
look
forward
to
advancing
some
of
these
capabilities
to
beta
and
GA
quality
to
understand
the
line
so
I'm
gonna
sit
it
over
too
deep.
The
boy
who
is
an
engineering
manager,
a
doctor
and
he's
gonna,
give
us
a
quick
demo.
G
Great,
so
what
we
have
here
is
a
very
simple
kubernetes
cluster
get
nodes
here,
shows
you
that
we
have
a
one
Windows
worker
node,
a
standard
master
and
a
Linux.
So
what
I'm
going
to
do?
What
I'm
going
to
demonstrate
here
is
how
GMs
a
credential
specs
one
of
the
features
that
just
when
beta
works
within
kubernetes.
G
So
here
we
have
the
Windows
node
and,
as
we
see
here,
it's
connected
to
a
domain,
it's
called
to
an
80s
GM
sa.
Just
for
the
sake
of
this
demo
and
there's
also
something
called
a
managed
service
account
a
GM
sa
created
within
this
domain.
We
are
calling
it
GM
SAE
to
be
from
this
GM
sa.
We
create
something
called
a
credential
spec
which
is
displayed
here
and
for
which
we
have
a
PowerShell
script
that
queries
the
necessary
item
and
produces
this
potential
spec
document
on
the
kubernetes
IDE.
G
What
we
did
is
we
take
this
and
we
put
it
inside
the
cluster
as
a
CR
D.
So
if
you
look
at
the
CR
DS
that
have
already
been
initialized
in
this
cluster,
we
have
some
thing
called
the
GMs,
a
Prudential
effects
that
windows
are
kubernetes
at
I/o
and
specifically,
we
have
created
a
GM
SAE
to
each
spec
that
corresponds
to
what
I
just
showed
you
here.
These
are
the
same
elements
pretty
much
but
presented
to
the
kubernetes
cluster
as
a
CRD.
G
G
Wherever
it
comes
up,
you
do
not
need
to
worry
about
making
sure
that
the
identity
needs
to
follow
the
pod
in
some
ad
hoc,
out-of-band
fashion,
I'm,
seeing
a
lot
of
popups
here,
not
sure
if
those
are
addressed
towards
me-
probably
not,
but
okay,
so
jumping
right
back
in
okay.
So
now,
let's
try
to
create
a
pod
that
uses
this
GM
s,
identity
and
comes
up
within
the
domain.
G
So
for
that
I
have
something
called
iis
with
G
Emison,
it's
basically
a
IAS
for
the
Internet
Information
Services
pod,
where
we
have
configured
a
reference
to
the
GMs,
the
e
to
e
that
I
just
showed
above
what
this
means.
So
this
is
kind
of
like
the
key
part
of
the
work
we
did.
We
made
the
GMs
a
credential
spec
name
and
a
JMS,
a
credential
spec.
G
As
far
as
parts
of
the
the
core
pods
fact
security
option,
and
within
that
we
have
a
stanza
called
the
windows
options
where
you
can
specify
these
DMS
a
credential
details
for
ease
of
the
user.
You
just
need
to
specify
a
JMS,
a
ete,
just
the
name
of
the
GMs
a
and
we
have
a
vent
hood
which
which
does
a
couple
of
couple
of
different
things.
One
is
validation
of
the
fact
that
the
service
account
that's
using
the
GMS.
A
does
have
access
rights
to
it.
G
In
this
case,
we
have
already
already
configured
a
cluster
role,
binding
that
says
that
the
default
service
account
has
access,
has
used
permissions
for
this
GM
si
and
what
it
also
does
is
when
the
part
is
launched.
It'll
mutate,
the
pod,
spec
and
put
the
actual
Prudential
spec
details
that
we
showed
here
within
the
within
the
pod
spec.
So
let's
go
ahead
and
launch
this
iis
GMS,
adagiamo
and.
G
G
What
we
see
here
is
the
webhook
by
mutating
webhook
we
have
configured
did
his
job
and
the
GMS
a
each.
The
GM
suspect
name
was
actually
expanded
to
based
on
the
CR
d
that
was
already
configured
within
the
cluster
into
the
actual
spec.
So
the
whole
thing
is
there
now
also
there's
a
command
called
NL
test
that
allows
you
to
verify
whether
the
pod
is
correctly
joining
to
the
domain.
So
what
we
can
do
here
is
cube
cuddle
exact
into
the
pod.
G
G
And
that
shows
that
we
indeed
joined
the
domain,
that
we
had
configured
on
the
on
the
window
side
and
just
to
double
check
that
the
cube
LED
through
the
dark
regime.
There
is
job.
What
we
can
do
here
is
look
at
docker
PS,
and
we
see
that
our
pod
came
up.
So
we
can
do
a
doctor
inspect
on
the
container.
And
here
what
you'll
see
is.
G
So
basically,
what
happens
is
docker
assigns
the
credential
spec
details
within
the
within
the
docker
container
that
it
came
up
with
as
well
I'm
not
able
to
find
the
entry
right
now.
But
essentially,
what
we
do
see
here
is
the
container
was
correctly
configured
with
bib
the
GMs
image
results
in
the
domain,
the
domain
being
correctly
and
merited
as
part
of
a
null
test.
G
G
G
You
well
the
no
art
case.
We
will
see
that
pretty
much
no
service
is
available,
which
means
that
part
did
not
yet
the
moon.
So
what
that
shows
is
a
domain
details
are
getting
correctly
configured
within
the
pod
as
part
of
the
cue
cuddle
integration
that
we
did
with
docker
and
that
same
work
is
also
part
of
CRI
overall
and
we're
working
with
Microsoft
to
make
sure
that
say,
mediation
wants
the
container
gee.
G
F
A
H
H
Okay.
So,
what's
new
in
horizontal
path,
autoscaler
recently
we
have
been
fixing
various
our
cases
around.
How
we
are
scaling
up
on
some
metrics
are
missing
how
we
are
scaling
down
when
some
metrics
are
missing,
because,
for
example,
pods
are
unschedulable
due
to
cluster
being
too
small
and,
like
various
needs
around
what
come
differently
since
participating
deployments
and
HPA.
We
also
increase
test
coverage
and
stability.
H
H
Hopefully
it
will
be
completed
in
near
future.
We
added
support
for
setting
both
pottery
quest
and
what
limits
it
was
frequently
requested
by
the
community.
We
added
cruncher
to
support.
So
you
know
you
can
use
vertical
without
the
scale
every
fall
deployment
States
will
set
jobs
and
cron
jobs,
which
were
previously
a
little
bit
problematic
and
we
increased
our
support
for
getting
metrics
from
primitives
in
cluster
autoscaler.
We
added
three
more
cloud
providers.
So
now
you
can
auto
scale
your
cluster
with
open
shift.
H
If
you
are
using
Magnum,
you
can
do
it
with
digital
also
in
packet,
and
we
added
support
for
ADA
Louis
spot
instances.
We
also
improved
performance
around
bin
packing
simulation
inside
of
the
court.
Last
of
cholesterol,
scalar.
We
hopefully
increase
the
stability
and
we
passed
direct
responsibility
for
individual
cloud
providers
to
appropriate
companies.
So
each
cloud
provider
is
now
responsible
for
their
own
fault.
H
A
B
So
this
is
just
a
repeat
of
last
week:
basically
they're
just
three
reminders
here:
the
if,
in
order
to
vote,
you
need
to
be
in
voters
MD,
that's
the
file
that
we
use
to
ensure
that
you
can
vote.
So
it's
kind
of
up
to
you
to
go
in
that
file
and
see
if
you're
in
there,
if
you
are
not
in
there
and
you
feel,
you've
made
enough
contributions
to
be
able
to
vote
but
you're,
not
in
voters
MD,
because
we
used
F
stats
for
that.
B
The
second
link
there
is
to
the
forum
where
you
can
write:
hey
here's
all
the
stuff
that
I've
been
doing
that
doesn't
really
get
counted
in
github.
Please
consider
me
for
voting
and
then,
if
you're
planning
on
running
for
steering
committee,
Wednesday
September
11th
is
actually
the
same
deadline
for
PRI
in
your
400
word
or
300,
word
biography,
with
like
your
little
platform
and
all
that
kind
of
stuff,
and
after
that,
it'll
take
us
a
week
to
prepare
and
then
the
election
begins,
which
should
wrap
up
almost
exactly
a
month
from
now.
B
I
A
Thank
you
very
much.
Okay,
next
on
the
announcements
is
a
proposals
for
sessions
at
the
San
Diego
computer.
Our
sorry
contributors
summit
for
cube
con
those
are
open
until
the
9th
of
September
the
form
is
in
the
agenda.
That's
gonna
be
a
really
super
awesome
thing,
so
please
do
propose
four
sessions.
If
you'd
like
to
speak
in
there
George
did
you
want
to
add
anything
to
that?
C
Who
is
it
is
Paris
or
someone
else,
I
think
that
would
be
me
digging
fire,
but
you
pretty
much
covered
it.
If
you
have
content,
you
want
to
see
it.
The
contributor
summit
please
submit
forms,
we're
gonna,
leave
it
open
for
a
little
while
and
then
we
will
have
the
content
group,
try
and
pare
it
down,
but
we
want
to
hear
everyone's
ideas,
so
thank
you.
I
didn't
mean
to
steal,
thunder
there
I
apologize
deeply.
A
Okay,
let's
see
we
have
the
last
part
is
the
shoutouts,
so
I'll
read
through
those.
We
have
a
shout
out
to
I'm
in
a
bunch
of
these
names,
but
legit
I
think
I
think
that
for
tirelessly
deflating
our
test,
there
is
a
link
to
that
shout
out
in
there.
We
also
have
shout
out
to
Ben
the
elder
and
Katherine
for
their
work
on
unblocking
the
queue
for
merges.
A
A
We've
already
agreed
that
I've
tapped
out
my
joke
limit,
so
we
won't
go
there
alrighty.
Well
then
that
concludes
this
meeting.
We
will
upload
the
video
to
YouTube
as
soon
as
we
can
and
I.
Thank
everybody
for
joining
on
my
first
ever,
which
I'll
be
doing
more
to
the
remainder
of
this
year.
Great
job.
Thank
you.
Hopefully,
work
out
the
jokes
look
they're
better!