►
From YouTube: Kubernetes Community Meeting 20170316
Description
We have PUBLIC and RECORDED weekly video meetings every Thursday at 10am US Pacific Time.
https://docs.google.com/document/d/1VQDIAB0OqiSjIHI8AWMvSdceWhnz56jNpZrLs6o7NJY
Demo Kargo deploy; Release 1.6, 1.5, 1.4; SIG UI; SIG Windows; SIG CLI
A
All
right,
everybody
welcome
to
another
thrilling
installment
in
the
Cuban
I
between
the
meeting
I
am
Jason
singer
de
Mars
and
I
will
be
hosting
today
and
I
am
one
of
the
co-leads
of
cig
closer
up.
So
I
guess
I
said:
cluster
ops
is
taking
a
turn
at
the
wheel
here
so
and
I'm
quite
excited
to
actually
see
a
demo.
Today,
from
my
cohorts,
rob
her
shoulders
got
a
demo
cargo
Rob
are
you
here,
I
am
hoping
so.
A
D
B
A
C
C
I'm
going
to
talk
about
what
the
demo
is
going
to
be
and
then
we'll
try
and
go
through
it,
because
it's
like
six
Cooper
Nettie's
deployments
in
parallel
I,
don't
like
to
do
anything
halfway
and
what
what
we're
doing
so
rack
and
really
believes
in
this
community.
Best
practice
approached
open
source
projects
and
deployments,
and
we
really
want
to
see
a
reduction
in
installer
sprawl,
but
even
more.
C
We
want
to
see
the
operators
coronaries
working
together
and
collaborating
and
feeding
their
changes
back
upstream
and
what
we've
been
doing
is
we
are
we
making
a
project
called
digital
rebar
which
is
underlay
and
it
attracts
all
the
infrastructure
cloud
physical
and
then
it
will
run
standard
things
like
ansible
to
then
do
deployments,
and
so
for
these
demos.
What
I'm
going
to
show
you
is
we're.
Just
using
community
play
boats,
which
is
ansible
cargo,
to
do
this
Coover
Nettie's
deployments
and
in
completely
off
the
shelf
way.
C
That
way,
as
we
find
things,
we
can
feed
improvements
back
to
cargo
as
cargo
improves.
We
can
come
back.
So
it's
a
really
nice
layer,
differentiation,
digital
rebar,
salsas,
its
infrastructure
problems.
So
it's
a
different
problem
space.
It
does
it
by
creating
modularity
within
the
infrastructure.
So
we
break
things
up
in
small
units
and
then
we
reuse
the
units
that
we
can
reuse
like
cargo,
and
then
we
keep
the
units
that
are
different
like
physical
infrastructure
different.
So
you
get
a
lot
of
reuse
out
of
the
approach
and
then
for
the
demo.
C
What
I'm
doing
is
I'm
going
to
show
a
deployment
against
Amazon,
where
I
can
show
an
upgrade
I'm
going
to
build
a
classroom
for
two
students,
then
I'm
going
to
have
an
OpenStack
deployment
that
is
based
on
Cooper,
Nettie's
and
then
helm
odds
of
stuff.
That
is,
as
the
80s
ap
grew
some
of
the
stuff.
That's
moving
into
the
community
and
depending
on
how
we
do
timewise
I've
got
I,
actually
have
a
48
node
infrastructure.
C
Out
of
the
interfaith
groups
that
I
can
kick
off
a
deployment
so
and
since
on
time,
box
I
will
start
doing
these
deployments
right
away
on
my
mouth.
C
C
What
you'll
see
it
do
is
it's
going
to
basically
just
guide
me
through
building
something
I
could
set
the
version
back
a
little
bit
if
I
wanted
to
play
an
older
version.
I
can
do
things
like
turn
on
secondary
integrations
like
helm
deus.
So
even
though
cargo
doesn't
include
league,
yet
we
do
it
as
a
secondary
staff.
The
same
is
true
with
like
prometheus,
so
we
have
a
monitoring.
You
can
click
on
monitoring
and
add
things
into
the
infrastructure
and
there's
no
magic.
C
It's
just
building
adjacent
file
that
describes
the
infrastructure,
the
nose
and
things
like
that.
So
that's
going
to
go
ahead
and
work.
I
shouldn't
have
named
it
amazon
because
that
was
already
named,
so
that
is
going
to
create
ad
opponent
like
this
and
I
will
show
you
what
that
looks
like
in
a
second
cut.
What
I
will
do
is
I'll
switch
over
and
do
the
Cooper
Nettie's
deployment.
C
I
want
to
create
tenants
of
students
can
see
each
other
and
then
I'm
going
to
run
those
same
deployment
templates,
and
so
in
this
case
it
will
go
ahead
and
start
that
work
going.
We'll
see
that
in
the
background,
at
from
from
the
deployments
as
we
build
them,
what
gets
super
interesting
is
when
you
then
look
at
just
came
to
you
between
students
view.
So
this
is,
let's
see,
have
to
login
as
another
user.
In
this
case
this
is
student.
C
One
student
one
is
getting
the
deployment
and
you'll
see
they
have
a
cluster
kubrick
Nettie's
cluster
being
built
for
them
dynamically.
On
the
fly
it's
just,
they
could
see
if
I
jump
back
over
to
the
primary
view.
It
should
hear
what
you'll
see
in
that
view
is
that
I
actually
can
see
all
the
clusters.
So
this
would
be
like
an
instructor's
view,
and
one
of
the
things
that
we
really
feel
like
is
important
here.
C
C
Not
where
to
find
working
in
those
cases,
you
could
actually
run
the
deployments
with
different
configuration
options.
We
could
test
we've
in
calico
and
flannel.
You
could
then
bring
in
different
operating
systems
and
mix
and
match
right.
The
goal
is
not
to
restrict
it
to
really
help
the
community
find
ways
to
bring
all
of
those
useful
variations
into
a
standard
practice.
C
So
right
really
worked
for
its
best
practices
so
that
we
can
reuse
code
and
share
things
Wow,
all
right,
I'm,
almost
out
of
time
on
that,
and
then
the
physical
deployment
going
to
look
exactly
the
same.
This
is
the
physical
cluster
with
a
whole
bunch
of
nodes
in
it,
and
I
could
walk
through
that
deployment.
So
that's
that's
that
a
deployment
there
oof
everything
so
nicely
laid
out
on
my
other
desktop
to
a
physical
deployment
laid
out.
There
would
basically
cool
in
a
whole
bunch
of
these
nodes.
C
It's
not
just
an
OpenStack
thing,
so
we
can
take
pieces
of
that
and
move
it
back
into
cargo
or
into
other
helm.
Chart
community
places
our
goal
with
all
this
is
to
create
an
environment
where
people
can
interact
and
we're
looking
for
people
who
want
to
sort
of
work
with
us
to
build
the
whole
stack
we're
not
creating
a
distro.
This
is
really
something
where
you
know
rack
and
can
provide
support
for
it.
A
A
A
C
You're
very
bullish
on
the
I
was
incredibly
skeptical
about
OpenStack
under
on
top
of
Cooper
Nettie's
just
six
months
ago,
and
the
progress
that
has
been
made
in
the
community
with
the
helm
charts
has
been
really
really
really
strong,
there's
still
a
long
ways
to
go
and
we're
getting
to
the
hard
parts
yeah.
But
my
expectation
is
that
this
will
be
the
way
to
a
book
fuzzums
back
inside
a
year.
Excellent,
alright.
F
Dancer
currently,
it
said
we're
not
really.
There
is
still
a
night
earlier.
There
were
seven
Bubba
milestone
still
in
place,
mimosa
and
we're
having
some
issues
with
a
protesting
heating
bill
system.
If
you
live
test
in
turn
house,
so
that's
basically
anything
that's
in
the
milestone
is
preventing
extra
blocking.
Some
people
have
put
non-release
blocking
but
still
blocked
release
once
everything
else
is
so
if
everyone
keeps
focusing
on
a
query,
therefore,
the
1.6
milestone
and
just
trying
to
get
everything
out
of
it.
F
F
A
F
Really
just
going
down
the
list,
the
biggest
here's
an
easy
flip
over
them.
I
think
most
the
things
are
being
done,
but
really
just
if
you're
did
you
get
england
and
he
just
being
responsive,
is
the
biggest
thing
so
we're
going
to
keep
burning
down
on
these
issues.
So
if
things
come
up
and
we
discover
that
we've
set,
this
issue
is
not
in
the
certain
area
of
NBC
news
and
somewhere
else
being
responsive
everything
accurately,
that's
the
biggest
thing
act
in
general
and
just.
A
A
E
G
G
Let's
go
up:
yes,
we
got
it
awesome.
So
if
I'm
understanding
correctly
164
communities
is
largely
kind
of
stabilization
released
and
that's
the
same
approach
that
the
dashboard
team
took
with
our
162
new
features
that
are
facing
to
users
are
we've
added
storage
classes,
so
those
were
introduced
in
14
and
now
you
can
see
those
in
dashboard
and
also
third-party
resources.
G
G
We
are
working
on
our
2017
plan.
We've
got
Dan
Pike,
helping
out
with
that
effort.
We've
decided
to
pull
together
some
some
road
map
details
there
and
then
finally,
one
when
you
feature
that
I
wanted
to
mention
this
on
the
horizon
is
the
exact
into
a
pod
functionality
into
working
on,
and
that
will
be
hopefully
coming
soon
that
doesn't
ring
any
questions.
A
A
D
B
D
H
D
No
wrong,
hey
guys,
the
quick
updates
from
us.
You
know
we
took
the
opportunity
of
the
stabilization
released
for
what
six
weeks
to
kind
of
work
on
on
other
things,
so
we
did
a
complete
rewrite
of
the
networking
box
or
plastic
windows,
and
now
one
of
the
capabilities
of
the
demos
at
the
next
conference
last
week
was
the
ability
to
usual
bnn
OBS.
D
As
then,
there
was
an
overlay
for
for
windows
notes,
so
that
gives
us
the
dbol
networking
main
shapes
that
would
be
both
across
windows
and
linux
for
containers
cause
we're
not
completely
done
with
the
work.
This
was
just
an
alpha
previewed
as
we
showed
that
the
Google
next,
we
feel
something
like
expensing,
not
like
an
Indian
s,
work
properly
and
equality,
but
it
was
a
great
sign
of
progress
that
were
able
to
get
network
connectivity
inside
the
containers
right
now.
D
So
all
in
all,
all
the
demos
will
be
done
in
the
past
were
able
to
do
them
again
now
with
a
much
more
improved
thermal
connectivity,
including
internet
access,
and
we
don't
even
need
the
cool
box
animals.
We
use
the
obeah
novia
player
as
well
as
a
gateway
to
specifically
has
control
the
traffic.
We
did
plant
in
going
thunderpig
networks
very
soon
and
presents
for
Intel
Architecture,
and
then
it's
my
people
is
working
on
native
tech-
overlay.
D
I
I
Movies
for
the
screens,
on
the
same
with
my
camera,
all
right,
so
this
quarter,
we
spend
a
lot
of
focus
on
stabilizing
the
KU
controlled
via
feature.
Let's
leave
this
best
shout
out
to
Jordan
and
ming-chi
for
a
lot
of
work.
They
did
there,
but
we
got
a
coup
control
edit
now
works
properly
with
apply
third-party
resources,
have
a
bit
work,
they're,
not
perfect
and
not
totally
feature
complete,
but
they
work
a
lot
better.
Instead
of
being
totally
unsupported,
we
have
support
for
explicitly
leading
values
and
apply.
I
We
have
merging
primitive
lists,
now
works
and
we
added
some
nice
porcelain
command,
but
allow
you
to
actually
view
and
set
the
annotation.
That
tells
that
you,
the
last
two
days
that
was
applied
so
in
the
past
of
the
documents,
would
say.
If
you
want
to
see
the
last
configuration
that
was
applied,
go
look
at
this
annotation
on
the
object,
which
is
a
single
line
of
JSON.
That's
not
formatted
for
human
eyes,
so
we
got
some
nice
porcelain,
turn
and
flare,
and
we're
going
to
update
the
docs
in
one
dot.
I
Six
to
use
those
commands
wind
makes
explaining
absolute
control,
apply,
works
for
more
than
odd
seven.
What
we're
looking
at
is
really
rounding
out
and
trying
to
finish
as
much
as
we
can
with
the
cout
control
fly
open
issues,
we're
aware
of
a
number
of
open
issues.
Specifically
the
strategic
merge
patch
keys
are
not
correct
on
the
logic
unions,
any
type
objects
aren't
handled
properly
and
they
are
in
many
cases,
but
there's
some
things.
I
J
K
Okay,
no,
nothing
to
me
thanks
I
vieira,
okay,
sure
so.
I
also
wanted
to
mention
the
word.
Ccli
has
been
doing
around
the
Service
Catalog.
The
greater
project
under
you
are
an
edit,
so
we've
been
working
on
a
base
around
making
sure
that
cube
CTL
can
play
well
with
a
Service
Catalog
API,
which
it
doesn't
today.
So
it
means
a
lot
of
stabilization
movie
team
server
side
and
also
trying
to
make
sure
these
are
discoverable
on
the
client
standoff
hard
coded
like
they
are
and
keep
CTL
today.
K
So
we've
been
doing
that
and
we
are
also
going
to
propose
these
are
working
Fergus.
It's
going
to
have
a
written
proposal
very
soon,
a
new
set
of
commands
for
interacting
with
a
source
catalog
service,
especially
for
end
users
like
service
customers,
so
they
so
that
they
can.
You
know,
request
services
and
buying
them
to
their
applications
easily
and
I
in
a
user-friendly
way.
Really,
instead
of
having
to
you,
know,
handle
a
binding
or
you
know,
a
service
instances
based
on
an
EML
files.
All
other
than
that,
like
Philip
mentioned
these
looks
best.
K
Release
has
been
for
a
lot
of
about
stabilization,
so
we
worked
a
lot
of
on
a
bug's,
picking
bugs
and
review
stuff
from
contributors.
We
also
finished
a
new
document
for
onboarding
new
contributors.
Deaths
are
interested
to
contribute
special
for
cube
CTL,
so
the
outcome
was
really
nice,
so
I
think
we
are
going
to
advertise
that
on
the
public
mailing
list
very
soon,
and
we,
along
with
a
call
for
contributors,
you
know
anyone
want
that
wants
to
contribute
to
keep
CTL.
We
now
have
this
document.
The
links
to
you
know
some
issues.
K
L
Question
on
until
you
mentioned
that
you're
adding
commands
for
the
Service
Catalog
260
li
now
the
service
catalog
is
broken
out
as
an
extension
to
the
core
components.
Why
aren't
we
doing
the
same
thing
on
the
command
line
tools
and
it
and
I
think
you
know
this
is
probably
a
longer
discussion,
but
we
need
to
have
a
scalable
way
where
we
don't
start
building
everything
into
take.
One
kitchens
think
here
in
terms
of
the
command
line
tool
set
yeah.
We
talked
about
this
before.
L
L
H
L
I
think
I
think
everybody
should
sort
of
like
roll
up
their
sleeves
and
help
build
the
extension
points.
I
mean
this,
isn't
on
the
Service
Catalog
supposed
to
actually
both
like
blazing
trail
and
use
it
right.
We
can
all
help
sort
of
like
blazed
that
trail,
but
I
think
we
should
you
know
we
should
do
that
in
a
you
know
and-
and
you
know
somewhat
related-
is
you
know
getting
keep
control
out
of
the
main
repo,
which
is
a
whole
nother
ball
of
wax,
but
but
we
need
to
start
breaking
the
stuff
apart.
Yeah.
I
H
So
there
are
most
definitely
multiple
things
that
way.
I
mean
nothing
that
has
been
said.
The
60
or
I
actually
disagrees
with.
We
just
need
more
people.
Work
out,
like
literally
three
people
show
up
to
see.
Oh
I
right,
so
there's
been
ongoing.
Work
for
several
quarters,
all
ready
to
move
more
functionality
from
the
client
to
server.
So
there's
claim
has
a
nope,
NPR,
I,
think
moving
gets
and
describe
functionality,
for
example
the
server.
H
So
once
we
have
that,
then
the
API
aggregation
mechanisms
that
Service
Catalog
you
eating
would
work
for
that
additional
resource
by
resource
sort
of
functionality
and
keep
control
and
then
for
special
commands
the
audience
pension
proposal.
Do
you
think
it
it
for
that
and
I
believed
it
was
eating?
A
prototype
is
a
line.
We
just
need
more
people,
helping
I,
think.
L
It's
a
larger
discussion
but
I
think
it's
also
a
bad
idea
to
blindly
move
stuff
from
client
to
server.
I
am
a
total
integrated
of
that,
and
if
we
do
move
into
the
server,
we
should
put
it
as
separable
components
that
are
alpha
that
are
essentially
helper
and
out
of
court
right,
I
I
know
we've
been
going
on
for
a
while,
but
the
three-way
merge
that
developing
to
apply,
I
think
is,
is
is
very
complicated
from
from
typical
users
point
of
view.
H
I
mean,
I
would
argue
with
the
term
blindly,
but
things
like
cascading
deletion,
where
you
can
even
delete
resources
properly
without
extending
the
clients.
Coordinate
resources
is
not
really
a
long-term
viable
position
so,
and
we
can
talk
about
which
is
addicting
should
stay
in
the
client
which
specific
things
should
go
on
the
server,
but
things
we
need
for
every
single
resource
having
it
in
the
clients.
A
problem.
L
So
Scarlett
no
I
just
want
to
say
that
you
know
I.
Think
I
know
that
that
I'm
going
to
put
alex
over
here
was
it
seeps
CLI
last
time
I
want
to
start
getting
involved
there.
So
you
said
that
there
were
only
three
people
there
I
think
we
need
to
change
that
for
sure
so
I
just
want
to,
like
you
know,
get
support
for
the
folks.
Who've
been
you
know,
making
this
stuff
happen.
You
know,
regardless
of
lack
of
sort
of
wider
attention,
so
I
don't.
A
So
because
that's
very
disempowering
from
a
project
standpoint.
If
you
have
people
working
hard
on
something
that
is
never
going
to
seek
a
light.
So
I
think
that's
something
as
a
community.
We
should
probably
look
into
seniors
and
later
just
to
have
that
discussion.
I
know
pretends
very
passionate
about
this,
so
there
may
be
some
some
leadership
there
and
there's
also
the
proposals
that
are
getting
linked
in
the
chat.
So.
L
And
I
think
we
need
to
do
a
better
job
of
a
disease.
Things
come
up
driving
them
through
the
feature
process
and
using
that
as
an
opportunity
to
get
early
feedback
into
this
I'd
love
to
see
us
have.
You
know
some
sort
of
regular
feature
review
meeting
where
everybody
can
get
involved
and
become
aware
and
learn
about
this
stuff
and
give
us
feedback
I
feel
like
there's
a
sort
of
Architectural
Review
component.
That
is
also
missing.
L
M
H
H
A
M
These
places
where
we've
had
done
a
bad
job,
where
cigs
have
bubbled
things
and
then
all
of
a
sudden
people
who
are
not
necessarily
paying
close
attention
to
that
sig
find
out
about
it
and
there
is
a
secondary
reaction
and
I
think
that
leads
to
the
leads
to
frustration
on
both
sides
and
that
I
think
that's
the
biggest
tension.
We
need
to
figure
out
what
they
learn:
stuff,
yeah.
M
B
You
did
successfully
join
in
to
suggest
government's.
Google
group
are
sorry,
I'm
not
suggest
Cooper
Nettie,
stash
governance,
google
group,
and
so
it
appears
to
be
open.
I
will
go
double
check.
All
the
permissions
on
its
own
is
also
publicly
beautiful,
but
it's
going
to
be
where
we
send
the
invite.
A
Great,
thank
you
so
much
for
that
doing.
Oh
great,
there's
a
link
in
the
chat
I'll
put
that
in
the
transcript
and
if
you
are
interested
in
being
part
of
the
governance
discussion,
this
is
your
opportunity
and
I
will
reiterate
it.
If
you
don't
have
an
opinion
spoken
in
these
meetings,
then
that
the
same
is
not
having
one
so
to
be.
Definitely
if
you
feel
passionate
about
these
things
be
present,
show
up,
and
definitely
let's
get
a
really
good
discussion
going
on
on
something.
A
So
Rob
and
I
are
going
to
be
hosting
that
and
we're
going
to
be
doing
our
regularly
scheduled
Sid
quest,
drops
meeting
at
that
time
and
will
be
transmitting
via
zoom
as
per
normal,
and
we
would
love
to
have
anybody.
That's
coming
to
glue
con,
be
a
part
of
that
and
if
you
are,
if
you've
never
been
in,
one
of
our
big
boss
drops
meetings
which
you're
missing
out
because
I'll
tell
you
they're
there,
the
back.
A
We
have
the
best
with
very
this,
so
yeah
come
to
come
to
our
meeting
and
you'll
definitely
get
to
see
some
in-person
discussions
and
really
our
goal
is
to
try
and
enthuse
and
integrate
the
operator
community.
So
we
can
get
more
feedback
and
and
information
from
real-world
cluster
operators
and
integrated
into
all
the
rest
of
the
great
work
that
you're
doing
so
and
last
up
it
looks
like
we
are
at
the
notices
section,
so
I
will
give
the
floor
to
sarahs
at
yoga
the
Canaries
day
thing.
A
B
J
B
A
Thank
you,
SI
no
community
meeting
on
330
because
we
decided
that
we
wanted
to
follow
this
trend
of
post
post
conferences,
eating
avoidance
and
also
no
community
meeting
on
460
and
the
the
caveat
there
is
that
if
16
is
not
shipped,
then
we're
going
to
keep
putting
all
hands
on
deck
to
continue.
The
meeting
and
you'll
be
very
focused
around
getting
things
off
the
list
to
get
those
milestones
done
so
looks
like
Caleb
miles.
Asked
leadership
summit,
invite
in
detail.
I
can.
B
B
Hopefully,
we
are
pulling
together
a
list
of
invites
which
will
include
all
of
the
sig
leaders
and
then
core
company
contributors
will
be
giving
a
budget
of
people
to
will
be
making
recommendations
of
people
to
invite
from
those
companies
and
then
a
budget
of
people
for
the
tops
contributing
company,
so
that
I
think
I
have
a
list.
I
was
five
or
eight
top
contributing
companies
and
make
sure
we
pull
in
the
people
that
can
help
us
push
forward.