►
From YouTube: Kubernetes Community Meeting 20160929
Description
We have PUBLIC and RECORDED weekly video meetings every Thursday at 10am US Pacific Time.
https://docs.google.com/document/d/1VQDIAB0OqiSjIHI8AWMvSdceWhnz56jNpZrLs6o7NJY
Demo: Digital Rebar Kubernetes Underlay; 1.4 has been released; 1.5 Planning and Features from SIG-Node, SIG-Cluster-lifecycle; SIG-Service-Catalog, SIG-autoscaling/SIG-instrumentation; Introducing SIG-CLI
A
All
right,
hello:
this
is
the
Cooper
Nettie's
community
meeting.
It
is
Thursday
September
29th
and
I
am
recording
this
since
Sarah's,
not
here
I'm
playing
the
Sarah
for
tonight's
performance
and
are
posted
publicly
after
after
the
meeting
on
a
YouTube
channel.
So
don't
say
anything
that
you
don't
want
the
world
to
see
and
and
and
Rob
don't
share
anything
on
the
screen
that
you
don't
want
the
world
to
see.
I
will.
C
A
A
A
A
A
Jason,
if
you
could
somehow
so
there,
we
go
you're
you're
way
ahead
of
me.
A
All
right
well
great!
So
it's
three
minutes
past
we
have
41
participants
and
let's
go
ahead
and
get
started
so
then
we're
going
to
start
with
a
demo
from
Rob
and
Greg
Greg
online
to
Robbie
no.
B
A
From
digital
art,
okay,
Rob
from
rack,
n
and
you're
going
to
be
demoing
digital
rebar,
which
is
there
open
source
sort
of
compliment
here,
and
what
we
have
here
is
visual
multi
infrastructure,
deploy
with
automated
pki
right.
So
take
it
away,
we
rob!
Can
you
key
to
your
screen
share
and
they'll
make
sure
we
see
that
awesome.
B
There
you
go,
you
should
have
it
excellent,
so
digital
digital
rebar
is
an
open
source
project
that
boy
we've
been
working
on
bits
and
pieces
up
for
quite
a
while,
and
its
goal
is
really
just
to
do.
Composable
operations,
some
break
big
operations
to
a
very
small
parts
and
what
what
we're
showing
for
this
is
the
cooper
Nettie's
work.
B
We
did
and
we've
shown
some
cooper
Nettie's
in
the
past,
where
we
just
took
an
ansible
playbook,
the
cargo
stuff
specifically
and
just
ran
that
in
this
case
we
looked
at
kelseys
cooper,
Nettie's
the
hard
way-
and
we
said
wow-
that's
really
cool.
Let's
do
that,
and
so
what
we
did
was
we
broke.
We
call
it
composite
came
on
POS
8,
and
so
we
broke
those
steps
into
small
units,
and
then
we
automate
the
small
units
and
so
I'm
I'm
going
to
walk
you
through.
B
This
is
the
rack
and
UX
in
front
of
digital
rebar
I'm
going
to
actually
do
a
cooper,
Nettie's
wizard
to
install
on
Google
I
already
have
one
running
in
amazon,
so
we'll
jump
to
that
in
a
second.
But
it's
going
to
go
through
I'll
pick.
Google
is
my
provider,
so
I
could
pick
different
different
providers
like
OpenStack
or
packet
or
Amazon
or
metal
from
a
wizard
perspective.
These
are
just
some
of
the
things
that
we're
exposing
like
what
version
Cooper
Nettie's
you
want
to
pull
down.
B
If
you
want
to
actually
check
the
the
checksum,
so
you
could
use
an
experimental
version,
all
sorts
of
stuff
I'll
show
you
there's
actually
a
lot
more
settings
available
and
then
we're
building
a
simple
cluster.
So
in
this
case
one
control,
three
workers
and
then
you
could
actually
switch
different
networking
models
so
we're
working
on
plugging
in
right
now
it's
flannel
and
no
sdn,
but
you
could
plug
in
calico
or
contrail
or
something
else
the
way
the
orchestration
works.
B
It's
perfectly
acceptable
to
substitute
blocks
of
function,
but
I'm
just
going
to
take
the
defaults
and
move
through
and
then
the
wizard
doesn't
do
anything
except
build.
This
JSON
file,
the
JSON
file
is
designed
for
you
to
create
a
repeatable
infrastructure
or
CSV
infrastructure,
Cooper
Nettie's
deployment.
We
can
deploy
other
other
things
too.
I
won't
name
the
thumb
because
you
would
consider
them
by
market
competitive.
B
B
So
I
have
the
AWS
when
I
pre-baked,
here's
the
google
one
I'm
it's
requesting
for
nodes
somewhere
in
here,
I
actually
have
my
compute
engine
running
and
it'll
start
spinning
those
nodes
out
in
just
a
min
or
two
whoops
and
all
right,
and
so
what's.
This
is
the
visual
part
of
the
Cooper
Nettie's
deployment.
So
remember
we
took
these
uber
Nettie's
the
hard
way,
so
here's
our
four
nodes
and
then
these
are
all
the
steps
that
are
necessary
to
build
the
Cooper
Nettie's
cluster.
B
B
We
just
felt
from
an
operational
perspective
that
this
was
simple,
repeatable
multi-os,
very
portable,
because
in
our
experience
when
somebody
looks
at
a
deployment
like
this,
they
want
to
be
able
to
point
their
finger
to
exactly
what
happened
where
and
what
you'll
see
going
on.
Is
the
system
is
behind
the
scenes
literally
doing
the
work,
and
we
can
see
it
point
by
point
conducting
those
operations.
B
So
if
I
come
back
in
here,
there's
my
AWS
one
is
already
finished.
If
I
look
at
that,
this
is
what
a
fully
complete
matrix
would
look
like.
This
node
is
my
control
node,
and
I
can
take
any
one
of
these
steps
I
could
look
at
this
is
where
we
installed
the
scheduler.
It's
just
ansible
I'll
show
you
the
ansible
in
just
a
second
and
then
I
can
take
this
from
this
node.
I
can
go
in.
B
It
actually
is
running
here
and
then
I
can
do
something
like
uploading,
the
guests,
a
guest
book,
app
where's,
my
guest
book,
and
I'm
not
going
to
have
quite
enough
time
in
the
demo
to
have
everything
coming
in
the
end,
but
you
could
actually
see
it
running.
It's
a
live
amazon
in
this
case
amazon.
The
deployment
that
I'm
doing
now
is
going
to
produce
a
comparable
experience
out
of
out
of
a
google,
and
I
could
then
do
the
same
thing
with
metal
or
up
and
stacked
or
other
other
platforms.
B
The
composition
that
we
did
is
ansible,
so
it's
still
an
ansible
set
of
playbooks
the
differences
we're
using
digital
rebar
to
orchestrate
them.
So
if
you
were
to
look
at
the
API
server,
install
right
still
follows
all
ansible
stuff,
but
it's
very
short
and
simple,
because
it
only
installs
the
API
server,
the
things
that
do
CNI
configurations
or
storage
configurations
or
certificate.
Those
are
all
in
their
own,
isolated
pieces.
B
From
that.
From
that
perspective,
so
you
can
go
in.
This
is
all
open
source
stuff
Patchi
to
license.
You
can
modify
these
pieces,
you
can
change.
You
can
add
extra
extra.
You
know
you
could
actually,
instead
of
adding
making
the
API
server
more
complex,
we
have
a
demo
trying
to
be
respectful
for
time,
but
you
could
go
into
a
working
deployment.
B
Like
my
amazon,
one
I
could
switch
it
into
editable
mode.
I
can
add
in
something
like
a
dais
capability
in
here
and
then
I
can
then
have
my
worker,
my
pic
Deus
here,
sort
of
fun
I
can
come
in.
I
can
say
I
want
to
add
deus
to
the
controller
I
can
tell
it
to
go
and
the
system
will
come
back
and
add
deus
into
the
infrastructure.
So
it's
going
to
go,
find
new
work
and
add
that
capability
into
the
infrastructure
jual
that
building
and
work.
A
A
Okay,
people,
you
meet
me
repeat,
if
you
don't
have
a
question,
otherwise
we're
just
going
to
assume
that
you're
about
to
talk
well,
I'll
start
so
so
so
this
is
interesting.
Drops
because
I
mean
one
of
the
things
that
we've
been
actively
talking
about
in
a
lifecycle,
clustered
life
cycles,
moving
more
towards
self-hosting
and
boot,
cube
and
essentially
use
kuber
Nettie's
to
manage
cooper
Nettie's,
and
it
seems
like
your
explicitly
going.
The
other
direction,
oh,
which
I
think
you
know,
is
not
wrong.
B
We
have
a
lot
of
experience
from
the
OpenStack
on
OpenStack,
a
mess
put
it
politely,
I
we're
really
excited
about
Cooper
Nettie's
and
its
ability
to
manage
them
sustain
applications.
From
our
perspective,
we
have
the
tools
to
just
build
exactly
what
Kelsey
says.
You
need
to
build
in
a
secure
way,
signing
this
very
operationally
transparent,
even
with
the
self-hosted
stuff,
we're
still
talking
about
isolated
control,
planes
and
all
those
pieces,
and
so
when
I
look
at
the
best
way
to
accelerate
Cooper
Nettie's
adoption.
B
It's
simple
operational
controls,
easy
networking,
easy
storage
right,
remove
all
the
complexity
from
the
install
and
so
yeah
our
operational
experience.
Building
platforms
is
under
platform.
Underlays.
Is
that
the
simpler
you
can
make
it
the
faster
it
is
to
adopt,
and
so
yeah
you're
right,
Joe
I?
Am
you
know
down
a
different
path
from
the
way
life
cycle
is
gone
and
I
gave
up
on
trying
to
influence
that
okay.
A
Well,
I
mean
that's
interesting
and
I
think
you
know
one
of
the
notes
that
that
Clayton
put
in
the
chat
is
that
I
think
you
know
there
is.
There
are
a
lot
of
pragmatic
concerns
there
and
it's
a
journey
and
oh
yeah,
I,
think
the
proof
is
in
the
pudding,
so
we'll
have
to
sort
of
see
how
this
stuff
plays
out.
B
And
there's
a
lot
of
things
in
life
cycle
that
I
think
are
very
beneficial
to
what
we're
doing
and
we're
excited
to
see,
because
anything
that
makes
Cooper
Nettie's
itself
easier
to
install
is
great.
Our
experience
has
been
that
right,
especially
because
part
of
what
we
do
is
metal
right
out
of
the
box
right.
So
you
have
to
deal
with
networking
interfaces
and
which
Nick
you
have,
and
building
storage
and
prepping
a
system
and
making
sure
that
you
can
patch
it
individually
and
self
hosted.
You
know
it's
going
to
have
to
you.
A
Yeah
and
I
think
that's
one
of
the
things
that,
with
life
cycles,
that
Louie
at
least
want
to
do
start
at
a
certain
level
and
there's
a
whole
bunch
of
work
under
that
I'm.
So
so,
first
of
all,
so
very
cool
I'm.
One
last
thing
is
that
how
do
you
handle?
Do
you
handle
management
or
deployment
of
things
on
top
of
the
gerber
Nettie's
cluster
once
the
Guru
Nettie's
cluster
is
up
because
that
doesn't
necessarily
map
to
a
node
or
an
ansible
playbook,
because
in
those
things
are
a
node
century.
Yes,.
B
I
didn't
show
this
in
the
demo,
so
this
is
actually
all
the
configuration
items
that
we're
exposing.
Fundamentally
the
way
we
look
at
the
system
is
it's
a
it's
a
cluster
and
so
the
row
of
icons
that
I
just
I
just
typed
on
for
this,
our
cluster
level
services
and
the
cluster
itself
is
a
cluster
level
service.
B
So
it
is
entirely
possible
to
then
add
other
things
into
this
into
a
deployment,
especially
from
you
know,
it's
all
restful
api
is
it's
just
an
angular
app
so
with
a
restful
api,
go
in
and
say,
tell
me
the
cluster
information
about
this
deployment
and
then
take
actions
against
it,
and
you
can
do
that
like
the
dashboard.
Here
is
something
inside
that
deployment
or
you
could
reach
in
and
get
the
capabilities
of
the
deployment.
If
you
wanted,
okay,.
A
B
If
I
had
add
it,
if
I
had
a
couple
of
control
nodes,
I
could
do
that
or,
if
I
need
to
add
more
nodes,
yeah,
it's
it
knows
it's
a
cluster.
The
thing
we
haven't
done
yet
is:
do
all
the
DNS
integrations
cloud,
specific,
DNS
integration
so
like
we
don't
yet
we're
aware
of
which
cloud
were
deploying
into,
but
we
haven't
done
the
work.
This
is
where
this
is
community
stuff.
So
we
would
love
to
see
people
who
want
to
collaborate
on
getting
those.
You
know
DNS
integrations
done,
taking
advantage
of
wait
a
second.
B
These
guys
already
did
all
of
the
base.
I
want
to
add.
My
I
want
to
plug
in
functionality
in
the
middle
of
the
Cooper
Nettie's
deployment
yeah,
and
that's
a
that's
exactly
why
our
design
is
for
its
to
say,
look
if
you
need
to
take
action
in
between
the
API
server
and
the
controller
coming
up
or
I
want
to
substitute.
One
condone
a
box
is
doing
this
like
a
different
controller.
The
way
it's
designed
is,
you
could
actually
say
this.
B
This
controller
is
equivalent
to
the
default
one,
and
you
can
substitute
in
a
controller
and
run
the
deployments
just
just
fine.
So
the
logic
assumes
that
you
have
alternatives
in
this
workflow,
because
in
physical
infrastructure
right
you
swap
one
brand
of
gear
for
another
brand
of
gear.
You
actually
have
to
take
different
steps,
and
so
that
philosophy
carries
all
the
way
through
this
right.
A
Role
for
another,
so
I
hate
tenor
too,
but
I.
We
got
to
do
a
time
check
here.
All
the
words
out.
There's
a
couple
of
questions
in
the
chat
that
we're
not
going
to
have
time
to
get
to
rob.
Can
you
go
contact
info
the
best
way
to
get
ahold
of
you
in
the
notes
and
folks
can
always
check
that
and
in
follow-up
offline
with
you
will
do
thing,
offer
them
to
appreciate
it.
Thank
you
for
the
demo,
sir.
Alright.
A
So,
let's
see
the
next
item
on
the
agenda
here
is
to
really
first
of
all,
1.4
is
out
the
door.
Thank
you
everybody.
It
was
an
amazing
effort
and
it's
it's.
It's
awesome
to
see
so
much
come
together
from
so
many
different
people
across
different
time
zones
and
companies-
and
you
know
really
is
what
makes
the
the
community
that
is
very,
very
exciting.
To
see
that
happen.
There
is
a.
A
I,
like
the
hump,
the
noise
on
the
thing
here,
yeah
alright
think
we're
good
yeah,
so
the
the
I'm
sorry
so
so
Jason
will
be
will
be
facilitating
there.
Thank
you,
Jason
and,
and
then
also
anybody
who
you
think
went
above
and
beyond
to
help
get
one
point
for
out
the
door.
There's
the
the
Cooper
Nettie's
community
award,
please
send
nominations
or
any
feedback
on
that
too
tessera
and
shall
shall
collect
and
fried
that
process,
which
is
super
exciting.
A
So
so
please
send
that
out.
It's
it's.
You
know
big
part
about
building
the
community
is
actually
recognizing
folks
who
are
really
going
above
gun,
and
this
isn't
just
code.
This
is
you
know,
all
sorts
of
different
things
and
I'm
not
going
to
name
names
because
I
don't
want
to
pry
I,
don't
want
to
put
my
thumb
on
the
scale
at
all.
Okay,
so
next
topic
here
any
comments
on
1.4
before
we
move
on.
Oh.
A
Alright
great
the
the
next
topic
here
is
1.5
planning,
and
we
have
one
two
three
four
different
sort
of
mini
presentations
to
go
over
here,
so
we're
going
to
have
to
be
a
little
bit
fast
going
through
this.
The
first
one
here
is:
we
have
sick
note
from
derek
and
let's
try
and
keep
these
things
to,
like
you
know
five
ish
minutes
each
and
see
how
that
works.
Derek
are
you
around
yeah.
E
Yes,
I,
don't
I'm
have
like
a
set
of
slides.
Anything
talk
through,
but
I'll
give
seems
at
signode
has
been
discussed,
saying,
give
little
details
so
for
signal,
we're
kind
of
breaking
things
out
by
certain
high-level
themes
the
first
year
that
we're
tacking
is
around
improving
our
runtime
portability
and
firming
up
the
container
runtime
interface
and
getting
basically
all
the
necessary
eyes
dotted
and
t's
crossed
to
support
multiple
runtimes
going
through
there.
E
We
did
a
lot
in
one
that
for
we're
looking
to
do
it
more
on
125
from
a
node
upgrade
standpoint,
there's
some
people
that
are
looking
to
tackle
how
to
do
in
place
updates
of
the
key
blitz
and,
in
addition,
let
make
it
easier
to
set
up
your
node
using
some
dynamic
config,
the
workaround.
A
farmer
support
enhancements
is
still
moving
forward
and,
let's
see
when
it
comes
to
resource
provisioning,
we're
trying
to
get
the
support
for
per
pod
level
c
groups
in
place.
E
E
There's
an
effort
going
on
across
signode
mcig
scheduling
to
improve
your
ability
to
support,
ok
character
resources.
I
would
sigh
describe
this
like
making
sure
that
node
doesn't
overwrite
what
it
reports
us
capacity,
forgery
resources,
but
we're
not
doing
anything
in
isolation
around
those
things
and
wind
up
HUD
and
then
the
last
two
areas
would
be
around.
E
A
E
A
E
I'm,
assuming
that
memes
can
the
CRI
be
back
by
things
like
hyper,
potentially
I
believe
the
answer
there
is
yes,
but
we're
still
working
through
some
of
the
edge
details
around
that
for
things
like
the
monitoring
metrics
pipeline
of
the
associated
that
okay
and
I
think
some
of
these
items
are
what
falls
under
the.
What
makes
this
an
alpha
level
API
or
not
in
general,
we're
just
trying
to
push
the
ball
forward
here
and
there
may
be
edges
yeah.
A
And
one
of
the
one
of
the
questions
there
was
like
qm,
UK
BM,
but
then
also
a
question
of
binaries
and
I.
Think
if
I'm
not
mistaken,
the
big
sticking
point
there
is
is
networking
right.
Goober
Nettie's
has
chosen
not
to
do
sort
of
port
assignment
and
enforcement,
and
so
that
that
becomes
one
of
the
sticking
points
there.
Yeah.
E
Yeah
I
would
say
that
the
Sierra
has
been
a
really
good
exercise
and
trying
to
understand
what
decisions
that
were
made
by
default,
that
we
should
revisit
now
and
I
would
just
say
that
it's
been
a
rather
extensive
process
to
understand
everything
that
we've
already
settled
on
in
that
process.
So
it's
an
exercise
so
all
right
take
your
equal
all.
A
F
G
F
Yeah,
so
the
three
slides,
including
the
life,
including
the
the
front,
slide
slew.
We
released
cube
admin
with
one
doc,
for
which
we
believe
is
a
like
a
big
step
forwards
in
terms
of
making
cuva
Nettie's
easy
to
install
there's
I
won't
go
over
everything's
in
the
blog
and
and
the
getting
started
guide,
but
please,
please
give
it
a
go
and
give
us
feedback
and
tell
your
friends
and
yeah.
We
really
want
to
iterate
on
on
cube
admin
quickly
to
to
make
it
even
better
so
and
additionally,
it's
designed
to
be
easy
to
automate.
F
So
if
you,
if
you
have
your
own
automation
for
setting
up
Cuba
Nessie's
already-
and
you
can
try
integrating
cube
admin
into
that,
it
should
simplify
the
amount
of
work
that
you
need
to
do
so
Rob
the
it
will
be
interesting
to
see
how
many
fewer
steps
your
your
setup
system
needs
to
take
if
it
were
to
use
cube
admin,
for
example.
So
just
a
interesting
thought
experiment
there.
So
Joe
asked
me
to
speak
a
little
bit
about
what's
next,
so
this
isn't
a
complete
list
of
everything
that
we're
going
to
do.
F
Next,
we
haven't
had
another
six
since
the
release
and
we
were
all
pretty
much
hands
to
the
pumps
for
for
the
in
the
last
meeting
so,
but
we
have
a
few
ideas
on
the
table.
So
the
first
thing
we
want
to
do,
of
course,
is
react
to
user
feedback
and
improve
on
what
we've
already
got
so,
for
example,
we'd
like
to
support
more
distributions.
We
currently
support
only
a
bunches
annual
and
central
seven
we've
already
had
people
show
up
in
slack
and
say:
hey
would
be
really
cool.
If
this
worked
on
trustee
and
I
said.
F
Okay,
awesome,
do
you
want
to
work
on
that
and
they
said
yes,
so
that's
really
nice
so
yeah
get
involved
and
we
can
try
and
point
you
in
the
right
direction,
then
sort
of
the
three
big
areas
that
we
want
to
work
on
in
the
future
and
I
don't
know
if
these
are
all
going
to
make
it
by
15.
But
these
are
the
sort
of
the
areas
that
we're
looking
at
the
first
one
is
high
availability,
so
being
able
to
deploy.
F
Multiple
masters
have
multi
no
debt,
CD
for
extra
resiliency,
etc,
and
this
is
actually
kind
of
hairy
as
a
problem,
but
it'll
be
interesting
to
get
our
teeth
into
that
and
and
try
and
tackle
that
the
other
aspect
is
upgradability.
We
want
to
make
it
possible
for
people
who
have
installed
clusters
with
cube
admin
to
easily
upgrade
to
the
next
release
of
Cuba
Nettie's,
and
there
are
lots
of
different
ideas
for
ways
to
solve.
F
F
And
it's
plausible
that
you
to
imagine
that
you
might
want
to
control
pretty
much
every
aspect
of
every
component
in
coover
Nettie's
that
can
be
configured
via
these
flags
and
we
don't
want
to
end
up
with
just
with
it
like
a
massive
list
of
five
pages
of
flags,
because
that
would
be
a
bad
user
experience.
So
we're
also
looking
at
ways
that
we
can
sort
of
manage,
manage
that
complexity
and
one
example
would
be
to
give
cube
admin
a
config
file.
So
that's
my
very
short
update.
So
please
get
involved.
A
A
You
know,
technology
stacks,
so
that's
kind
of
it's
kind
of
what
we're
looking
for
there.
Hopefully,
hopefully,
we
can
actually
make
some
significant
progress
in
1.5
all
right
great!
Thank
you.
Look!
That's
awesome!
The
next
one
we
have
is
any
questions
on
it,
for
we
all.
F
A
All
right
and
so
all
right
and
then
some
questions
on
finding
the
docs
and
making
sure
they're
up
to
date,
so
yeah
so
so
docs
are
a
big
big,
big,
big,
big,
big
big
part
of
this,
and
so
I
very
much
like
as
you
go
through
this.
This
is
going
to
be
some
of
the
stuff
that
you
know.
Users
are
going
to
hit
early
on.
If
you
see
typos,
if
you
see
like
it's
hard
to
find
stuff,
if
you
think
that
there's
information,
that's
missing,
please
you
know
file
an
issue.
Let
us
know
submitted
PR.
D
One
of
the
question
which
is:
is
it
cross-platform
right
now
in
the
sense
of
like?
If
I
take
this
joy,
Raspberry
Pi
cluster
is
they're
going
to
work
well,
I
think
I.
F
So
Lucas
in
the
sig
has
worked
tirelessly
to
make
sure
that
it
works
on
his
Raspberry
Pi
cluster,
so
I
think
we
did
manage
the
land
arm
and
arm.
64
builds
all
right
I'll,
give
it
a
minute,
then
so
yeah,
please,
please
give
it
a
try,
I'm,
not
sure
if
we've
met
supports
arm
yet
so
you
might
need
to
use
something
else
for
networking.
Why.
D
A
What
interesting
thing,
though,
that
I
did
notice
I,
don't
know
people
are
talking
about
this
as
we're
doing
cube
admin
is
that
it
will
I'm
trying
to
remember
it.
Will
it
will
launch
a
a
daemon
set
for
the
proxy
on
a
for
one
for
each
of
the
architectures
that
it
might
want
to
support?
So
there's
this
question
of
like
do.
We
want
to
support
heterogeneous
architectures
in
the
same
cluster,
which
is
missing,
I?
Don't
think
anybody's
really
tested
or
thought
about
that
too
deeply,
yet
or
maybe
I.
D
Me
just
as
a
concrete
example:
the
windows
work
that
aprenda
is
doing
is
running.
The
H
is
running
the
proxy
with
a
different
config
and
it
will
run
with
the
proxy
with
a
different
config
and
there's
very
clearly
good
use
cases
for
having
a
hybrid.
You
know,
Windows,
Linux,
cluster
and
so
I
think
we're
going
to
want
to
absolutely
I
gather.
A
D
F
Is
an
undocumented
cloud
provider
flag
on
cube
admin
in
it?
The
reason
it
isn't
documented
yet
is
that
I
didn't
know
enough
about
how
to
make
various
cloud
providers
work
to
document
it
well,
so
we
could
really
do
with
some
help
with
that,
and
so
it
does.
I
think
in
figures.
The
controller
manager
with
the
cloud
provider
argument
may
be
wrong
about
that,
but
it
passes
it
through
to
to
something.
Certainly
so
it's
there,
it's
not
documented
because
we're
not
sure
about
how
to
make
sure
that
people
have
a
good
experience
with
it.
A
There's
an
open
question
around
sort
of
auto-detection
there
right
like
how
fancy
do
we
want
to
get
about
doing
heuristic
e
stuff,
we're
going
to
just
have
to
feel
our
way
through
that
yeah,
okay,
I,
don't
want
I,
don't
want
to
take
too
much
time.
Thank
you,
Luke!
That's
awesome!
The
next
is
Paul
with
sleep
service
catalog
and
we
are
just
for
time
check
we're
about
10
33,
so
we're
moving
along
okay.
I
Well,
I'll
make
this
click
quick,
because
I
don't
have
a
voluminous
spreadsheet
or
a
fancy
presentation,
unlike
some
other
sig,
leads
that
I
know
so
I
for
those
of
you
that
aren't
familiar
with
six
Service
Catalog
six
service
catalog
is
about
making
the
cloud
foundry
service
broker,
API
usable
in
Coober,
Nettie's
I,
and
what
that
API
is
is
an
API
that
a
right
now,
it's
it's
mostly
SAS
providers
that
implement
it
and
in
what
it.
What
is
the
value
added
this
API
is?
I
So
right
now,
what
we're
kind
of
talking
about
in
the
sig
is
determining
the
scope
of
what
we
want
to
do
for
our
initial
milestone,
and
we
paired
that
down
to
we're
going
to
get
probably
a
big
black
cat
walking
into
frame
here
in
a
second,
so
just
be
ready.
What
we
paid
that
down
to
is
sorry
he's
he's.
A
monster
is
basically
just
the
use
cases
of
consuming
a
service
broker
and
running
your
own.
I
I
We
all
agree
that
this
is
an
instance
of
something
that
we'd
like
to
see
be
a
separate
API
server
be
federated
with
the
main
API
and
and
perhaps
even
have
a
binary
extension
to
cube
CTL,
which
are
all
things
that
we
wanted
you
to
make
it
easy
for
folks
to
build.
Thank
you.
Thank
you
Joe!
That's
you
know.
My
applause
meter
has
been
pretty
low
recently.
So
that's
good.
I
These
are
all
things
that
we
want
to
do
to
make
it
easy
for
the
community
to
build
out
things
like
this
so
right
now,
then
the
main
tension
in
the
group
is
between
a
expediency
and
making
a
longer-term
investment.
So
if
you
have
an
opinion
on
this
subject,
I
encourage
you
to
join
us
at
four
pm
eastern
1
p.m.
pacific
time
on
mondays
and
I'll
post,
a
link
to
the
community
page
for
this,
which
has
all
the
other
links
that
you
can
follow.
So
I
won't.
I
I
won't
speculate
on
15
items
that
will
actually
get
into
main
crew
benetti's
in
15
because
we
haven't
scoped
yet,
but
that's
where
we're
at
and
that's
where
we're
going
thanks
a
bunch.
A
Awesome,
thank
you
Paul
and
I.
Just
you
know,
and
the
applause
there
I
want
to
just
be
clear,
I'm
I'm
applauding
for
the
idea
of
us.
You
know
making
sure
we
prove
out
the
extension
mechanism
so
that
we
can
build
really
interesting
things
on
top
of
Cooper
Nettie's
without
having
to
the
start
working
at
the
you
know
at
the
core
right
and
this
definitely
imports
that
whole
sort
of
what
score.
What's
not
core
discussion,
I'm
not
going
to
let's
not
rat
hole
now,
but
but
that's
a
that
was
my
my
response.
There
are
there.
A
No
all
right
well,
thank
you,
Paul
awesome,
so
II
or
a
feature
status
just
across
1.5
plant,
or
we
don't
know
with
it.
Let
me
make
sure
I'm
not
wonder
twin
yeah,
okay,
so
II
or
duh.
You
want
to
go
over
real,
quick
where,
where
you're
seeing
things
well,.
H
So
it'll
be
like
an
on
the
brief
update,
because
we
don't
have
any
like
real
big
it
shall
status
or
diffuse
because
we
don't
have
on
the
strategy
wanted
my
filtration.
So
these
this
limb
provides
you
with
a
table
where
I
have
collected
all
the
actual
leaves
for
of
one
of
her
release
from
our
feature
story
book
so
I
see,
remember,
have
prepared
a
similar
spreadsheet,
wonderful
release
where
we
have
collected
all
the
actual
featured
it
at
being
included
and
very
included
into
wonderful,
is
over
different
statuses
and,
like
these
disparate
collects.
H
Also,
the
new
features
that
mark
this
milestone,
one
at
five,
it's
like
a
word
or
the
draft,
but
I'm
expecting
from
the
feature
on
us
and
on
other
interest
people
to
feel
it
and
provide
with
the
actual
status.
Also,
if
someone,
if
someone
s
any
new
features,
ideas,
please
go
ahead
and
preparing
you
assure
like
a
new
feature
issue
in
the
future
straight
for
and
start
working
on
it.
So
that's
all
from
me.
A
H
A
Yeah,
no
thank
you
for
doing
doing
the
good
work
of
organizing
this
stuff
I.
Think
it's
going
to
give
us
a
lot
more
clarity
around
the
themes
of
15
as
it
develops
and
where
things
are
at
and
where
the
efforts
and
are
going
and
where
all
the
exciting
stuff
is.
So.
Thank
you
so
much
any
questions
on
that.
G
Right
awesome,
thank
you
so
much
since
somebody,
there
was
a
question
about
Ian
features,
repo
versus
spreadsheet
and
since
I
was
the
last
give
this
who
went
through
this
exercise.
I
will
say
again:
I
really
appreciate
the
features,
repo
and
I
do
think.
That's
where
all
of
our
automated
tools
should
be
pointed
out.
What
I
ran
into
was
we,
as
humans
did
not
use
the
feature
repo
effectively.
It
was
a
very
noisy
mess
and
the
only
way
I
had
anything
clean
to
look
at
was
thanks
to
e,
wonderful,
ongoing
long
manual
efforts.
G
I
think
we
now
have
a
bit
of
time
ahead
of
us
to
discuss
during
the
retrospective
coming
up
about
how
we
could
more
effectively
use
the
release,
features
repo
and
how
we
might
think
about
timeline
wise
sort
of
summarizing.
What's
going
to
go
into
the
release
a
little
bit
earlier
to
help
give
people
a
heads
up
of
like
what
they
should
anticipate
for
beta
testing
and
what
there.
A
H
It
would
be
great
and
I
am
second
then
eran,
eran,
sus
forget
in
debts
or
the
car
home.
We
were
able
to
click
it
in
a
manual
way,
but
it
will
be
great
little
automated
in
somewhere
because
feature
evaporates
like
the
source
for
all
your
feature
initiatives,
but
we
have
to
collect
it
in
some
specific
way.
One
single
document
11-1
salute
bios,
like
it.
Ok.
A
And
then
short
run
what
about
just
linking
the
spreadsheet
publicly?
You
know
it's
in
the
notes
here
Josh.
What
did
you
have
in
mind
in
terms
of
making
it
more
better
I
I
can.
H
A
I
think
you
know
there's
there's
definitely
room
for
somebody
to
build
a
tool
that
sort
of
scrape
some
stuff
from
the
features,
repo
and
populate
to
spreadsheet,
and
at
least
that's
at
least
a
good
starting
point
for
some
for
automating
some
of
this.
But
I
think
let's
leave
that
for
the
for
the
retrospective,
perhaps
alright
very
cool
okay,
let's
see
and
then
the
next
one
is
Sally
Ross
talking
about
siga,
auto
scaling
and
cig
instrumentation
and
the
future
of
monitoring.
So
getting
a
quick
update
there.
Alright,
so.
J
Sadly,
I
do
not
even
have
a
cat
to
spice
up
my
presentation
but
I.
We
we
have
started
in
cigarette
scaling
and
cig
instrumentation
kind
of
discussing
a
vision
for
the
overall
future
of
the
uber
nettings
monitoring
architecture
and
how
we
want
how
we
want
that
to
look
and
what
components
we
want
to
be
replaceable
common
and
how
we
want
the
api's
and
the
interaction
to
look
there.
J
We
haven't,
we
haven't
quite
finalized
anything
yet,
but
we're
hoping
to
get
some
of
the
initial
work
started
in
15
I
will
post
a
link
to
the
kind
of
work,
the
vision
document
as
well
we're
discussing
on
it.
We
have
some
questions
that
we
need
to
discuss
with
around
significant
with
sig
a
peon
machinery,
but
if
you're
interested
attend
the
next
sugano
scaling
and
sig
instrumentation
meetings.
A
All
right
awesome,
thank
you.
Any
questions
or
comments
there,
real,
quick,
all
right,
Thank,
You,
Sally,
ok,
and
then
we
have
a
cig
update
just
in
general
fabiano,
it's
gonna
be
talking
about
six
CLI
and
sort
of
an
announcement
that
hey.
This
is
a
thing
yeah.
C
You
know
between
all
the
tools
we
have
on
under
cube
CTL
today,
and
not
only
that
things
in
terms
of
the
architecture
of
command
line
tools,
but
also
how
we
keep
our,
for
example,
POSIX
compliance
and
how
we
look
good
in
terms
of
user
experience,
usability
for
our
end-users
developers
and
develops
those
kind
of
stuff,
so
we're
still
discussing
the
exact
boundaries
of
this
new
sig.
So
it's
pretty
clear
that
in
terms
of
code,
it's
what's
under
the
cube
CTL
package.
C
But
what
exactly
are
our
boundaries
and
the
scope
of
our
work
is
still
under
this
discussion.
So
as
soon
as
we
have
something
we
are
going
to
update
the
the
community
ripple
with
more
information,
so
things
like
how
we,
how
what's
our
relationship
with
other
tools,
tools
other
than
cube
CTL,
specifically,
so
the
other
command
line
tools
we
have
so
that
kind
of
stuff,
so
basically
we're
right
now,
finding
a
time
for
our
meetings,
it's
going
to
happen
every
two
weeks,
so
probably
not
starting
the
next
week
on
Wednesday.
C
It's
not
already
defined
the
exact
time,
but
we
are
discussing
that
in
the
mailing
list.
So
by
the
way
there
is
now
a
QQ
Burnett
is
sick,
CLI
mailing
list
that
we
might
may
be
interested
in
joining
and
then
you'll
seek
CLI
selection
channel.
So
that's
it.
So
if
you're
interested
in
command
line
tools,
so
please
get
involved.
We're
available
for
questions.
Ideas
concerns
I'm
not
sure.
If
Phillips
or
tony
are
here
now,
but
if
you
or
do
you
have
anything
else
to
add
other.
A
Well,
I,
for
one
can
say,
I'm
really
excited
about
this.
I
think
you
know
usability
and
sort
of
really
looking
at
the
the
you
know,
continuing
to
evolve
and
improve
and
extend
cube
control
is
very
exciting,
so
I'm
going
to
be
I'm
going
to
be
tuning
in
personally.
Any
anybody
else
have
have
questions.
A
No
all
right.
Thank
you
very
much
hobby.
No,
thank
you
you're
in
a
bunch
of
plus
ones
in
the
chat.
So
that's
good
good
news.
Alright,
so
we
have
a
little
over
ten
minutes
left.
There's
a
bunch
of
notice
is
in
the
in
the
notes
that
I'm
going
to
actually
read
through
first
off.
There
is
the
first
Cooper
Nettie's
meet
up
in
austin
or
Austin's.
First
Cooper
Nettie's
meet
up,
which
is
exciting.
A
So
if
you
are
within
driving
distance
of
Austin,
it
would
be
great
if
you
could
go
help
out
there
and
show
your
face
and
help
build
that
community
q
con
the
the
schedule
has
been
posted.
Things
are
looking
very
exciting
I.
You
know
just
I
want
to
shout
out
to
the
folks
who've
been
putting
that
together.
That's
a
lot
of
time
to
actually
pull
together
those
CFPs
and
go
through
them
and
and
let
everybody
know
a
lot
of
work.
A
Okay,
so
question
is
okay:
who's
running
the
austin
cooper,
Nettie's
users,
group
I,
don't
know,
there's
a
link
from
the
notes
and
so
I'm.
Assuming
that
there
is
some
an
answer
behind
that
link
Josh
the
next
questions.
Any
word
on
q,
con
lightning
talk,
CFPs
I
saw
something
I'm,
not
super
involved
in
that.
But
let
me
your
quick
I
can.
H
Riff
a
little
bit
so
let
enlighten
talk
so
Steve
be,
is
already
closed
but
see
and
see.
If
and
some
volunteers
from
from
the
cubicle
on
committees
are
working
on
that
and
we
are
going
to
Wyatt
some
results
racer.
So
we
don't
have
any
final
working
goodbye
estate
here,
please,
okay,
so
that
is
happening.
Oh
skip.
A
Either
closed
all
right
and
then,
let's
see
so
Diane
here,
has
a
in
the
chat.
The
open
shift
community
is
hosting
a
prequel
to
q
con
on
november
seventh
in
seattle
and
with
a
link
there,
let's
see
so
yeah,
so
we'll
make
sure
that
gets
copied
into
the
notes:
Thank
You,
Diane
and
then
yes,
lightning.
Talks
are
closed.
Okay,
so,
let's
see
so
also
around
q
con
there's
the
Cooper
Nettie's
developer
summit.
A
There
is
a
lottery
to
attend
and
to
submit
targets
and
sort
of
more
information
about
sort
of
how
that's
going
to
be
structured.
All
that
information
is
in
the
notes,
and
so,
if
you're
going
to
be
in
town,
please
please
take
a
look
at
that
as
appropriate.
A
Anyone
interested
in
being
a
mentor
for
outreach,
e
or
google
Summer
of
Code
I'm,
not
familiar
with
outreach
II
myself,
but
if
you're
interested
go
ahead
and
contact
sarah
and
somebody
got
that
oh
okay
yeah
go
contact,
sarah
&
and
let
her
know,
I
think
it's
a
great
opportunity
to
help
grow
our
community-
and
you
know,
start
meeting
some
new
folks
and
mentoring
them
and
really
really
cool
stuff
there,
and
then
also
there
will
be
a
new
community
calendar
invite
coming
so
stay
tuned.
They're
managing
google
calendar
with
a
group.
A
A
A
Ok
and
then
so
josh
is
asking:
how
do
we
get?
The
invite
list
is.
How
do
we
get
on
the
invite
list?
Is
this
for
the
the
the
developer
summit?
I
assume
that
that
josh
is
asking
about.
There
is
a
a
lottery
link
to
try
and
keep
it
somewhat
limited.
There's
there's
a
link
where
you
can
submit
your
name
and
and
OH
for
this
meeting.
How
do
we
get
for
this
meeting?
I?
Don't
know
it
was
a
manual
process
for
Sarah.
Hopefully,
she'll
find
a
way
to
automate
that
it.
G
A
There
was
something
like
where
it
topped
out
after
a
certain
number
of
people
like
calendar
just
gave
up,
you
know
just
threw
up
his
hands
and
said:
I
can't
deal
with
this,
so
maybe
that's
one
of
the
things
that
that
Sarah's
helping
to
fix
so
yes
but
I,
hear
you
Josh.
That's
it's
useful
to
have
this
automatically
updated
on
your
calendar.
A
A
Emails
around
response
looks
overwhelmingly
positive,
so
this
is
this
is
very
exciting,
moving
towards
a
more
automated
sort
of
who's
responsible
for
what
thing
and
expanding
the
set
of
reviewers
and
and
building
out
some
official
expectations
around
what
it
takes
to
actually
be
be
on
that
list.
So
that's
very,
very
exciting.
It's
going
to
be
one
of
the
ways
that
we're
going
to
scale
the
community
and
make
sure
that
we
have
people
ways
to
actually
get
involved
early
and
then
grow
their
responsibilities
over
time.
A
So
that's
very,
very
exciting
around
that
and
thank
you
for
helping
their
heel
and
yeah,
so
any
other
announcements
or
notes
that
folks
would
like
to
share.
If
not,
I
will
say
that
we
we
can
end
a
few
minutes
early
going
once
going
twice.
Oh
wait!
The
seattle
looks
like
the
extra
line.
Okay,
so
there's
some
question
around
the
cooper
Nettie's
meet
up
in
the
notes.
I'm
not
quite
sure
some
talks
might
have
been
swapped
out
further
speakers.
So,
if
X
in
chat
line
there,
someone.