►
From YouTube: Kubernetes SIG Cluster Lifecycle 20170912
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
B
Thank
you
so
much
Lucas.
So,
just
briefly,
I
were
in
the
cleanup
stages
of
getting
the
branch
ready
to
go
so
we're
doing
daily.
Fast-Forwards
on
the
1.8
branch
and
sort
of
the
job
right
now
is
to
do
a
lot
of
the
the
pre-release
work
with
documents
and
release,
notes
and
all
that
stuff.
So
in
terms
of
deadlines
right
now,
documents
related
to
code
and
you
guys
have
a
lot
of
stuff
in
the
queue
I.
B
B
So
if
it's
just
some
generic
fix
or
something
like
that,
and
it
doesn't
have
an
issue-
that's
marked
as
critical
urgent,
where
those
are
gonna
have
to
be
skipped.
So
right
now
is
the
time
to.
If
you
have
any
of
those
specific
plot
requests
for
1.8.
We
should
look
at
those
now
before
we
encode
for
you.
Before
the
code
freezes
over.
B
There
was
a
bot
that
went
through
and
removed
things
out
of
a
milestone.
They
did
not
have
proper
three
criteria,
which
is
urgency,
a
cig
and
the
oh
gosh.
Let's
the
other
thing,
I
heard
to
see
the
milestone
and
the
the
documents
that
the
issue
is
either
true
or
false,
so
those
were
removed
from
the
milestone.
So
if
you
had
some
issues
that
we
need
to
go
back
in,
please
make
sure
there
you
few
any
issues
you
have
that
need
to
be
in
the
one
nine
eight
milestone
I
asked.
B
Yesterday
we
had
61
open
issues
and
all
those
are
considered
release
blocking
until
they're
resolved
or
taken
on
the
milestone.
And
if
you
go
to
bit
Dudley
slash
Cates
one
eight
release,
you
can
see
all
of
the
breakdown
of
all
these
things
that
I'm
talking
about
right
now.
So
that's
our
running
release,
notes
for
the
burned
out
meetings
for
specific
issues
for
Kuby
diem.
B
There
is
a
tracking
issue
that
Lucas
created
make
sure
that,
if
there's
anything
that
we
need
to
know
about
Canadian
related
that
it
goes
in
there
and
that
is
currently
in
the
1.8
milestone
and
milestone
approved,
so
we'll
be
tracking
that
till
the
very
end,
and
once
we
get
to
the
point,
we're
gonna
cut
the
release.
I
assume
that
any.
If
it
all,
the
issues
are
fine,
we'll
just
close
that
particular
milestone
and
then
we'll
be
open,
full
moon
at
9:00
and
just
do
the
same
tracking
for
it.
B
For
that
there
are
some
great
1.8
mile,
our
one
dot,
a
milestone
dashboards
as
far
as
testing
goes.
Those
are
also
linked
in
the
the
agenda
that
I
shared
with
you
moments
ago.
All
the
feature
exception,
requests
that
were
requested
have
been
granted
and
have
been
put
in,
so
we're
good
there
and
yeah.
A
B
D
Jess
I
pulled
up
the
links
for
released,
blocking
issues
and
filtered
by
our
cig,
there's
just
one
which
is
issue
Lucas,
created
about
the
release
tracking
issue
for
cube
admin,
which
is
basically,
let's
not
screw
it
up
like
we
did
before
I
think
that
means
we're
in
pretty
good
shape.
If
the
only
thing
we're
tracking
is,
let's
make
our
release
actually
work.
D
B
Fantastic
I
haven't
actually
looked
this
morning,
but
if
that
is
true
now
the
one
thing
I
would
say
Robert
is
that
we
do
need
to
make
sure
that
if
anything
fell
out
of
the
1.8
milestone
because
of
a
missing
label
or
something
that
it's
not
lost
in
the
in
the
wash.
So
just
maybe,
if
you
guys
could
look
at
all
the
all,
even
things
that
are
not
in
them.
One
item
I'll
stone.
There
they're
tagged
with
the
sake
just
to
make
sure
that
that
everything's
hunky-dory
there
yeah.
B
A
E
A
B
Yeah
just
stay
in
communication
with
me
about
whatever
you
need
to
emerge
in
and
we'll
we'll
handle
it
we'll
make
sure
you're
in
good.
It's
because
that's
critical
path,
so
yeah
right
there.
Any
other
questions
concerns
I.
F
Don't
question
or
concern,
but
regarding
the
majors
theme
section
and
the
release
notes,
this
is
Kayla
by
the
way
you
usually
I,
guess
historically,
SiC
p.m.
has
provided
these
major
themes
in
consultation
with
the
other
sig
leads,
often
in
the
setting
like
the
Leadership
Summit
we
had
last
year,
unfortunately,
for
1/8
there
doesn't
seem
to
be,
at
least
in
my
personal
view.
There
doesn't
seem
to
be
the
possibility
of
continuing
that,
at
least
for
one
eighth
and
looking
at
the
features
that
went
into
the
futures
repo.
F
It
says
it
would
be
hard
to
drive
an
overall
theme.
This
I
would
not
be
able
to
drive
an
overall
theme
for
the
police
and
so
wondering
if
it
makes
more
sense
in
the
major
theme
section
well,
one
either
to
drop
the
major
theme
section,
or
rather
to
instead
break
up
the
major
theme
section
by
sig
and
the
city
put
in
their
goals
their
what
their
dominant
goals
were
for
1/8.
Since
she
you
all
would
have
a
better
idea,
then
we
wouldn't
say
p.m.
F
B
Think
that's
great
because
it's
really
starting
to
align
with
the
overall
governance
model
I
mean
sick
p.m.
was
really
critical
in
earlier
releases,
because
the
themes
helped
people
who
are
consumers
weren't
familiar
with
the
underlying
structure,
to
understand
where
we're
going,
but
now
to
say
that
we
have
release
level
themes
feels
a
bit
anachronistic
I'm.
A
huge
fan
of
that
idea.
So
so
caleb
is
the
idea,
then
that
we
would
have
radical
and
Jennifer
help
curate.
The
that
section
by
sake
would
that
you
know
yeah.
B
That's
how
I
would
see
that
really
amped?
Okay,
so
you
so
just
to
be
clear
to
everybody
and
when
the
team
either
radical
Jen
will
will
contact
you
to
help
get
that
information.
So
if
you
could
spend
a
little
time
in
this
meeting
to
even
just
decide
what
are
the
you
know,
one
or
two
things
that
you
guys
intended
to
to
accomplish
and
and
were
able
to
roll
out
this
this
release,
but
I.
F
Guess
yeah
I,
don't
wanna
have
like
a
unfunded.
If,
if
that
doesn't
sound
like
it
makes
sense
teal,
let
me
know
because
I
don't
want
to
I
mean
I,
can't
obviously
compel
you
to
do
any
work,
but
it,
but
if
that
makes
sense
and
you
don't
believe
that
the
major
themes
are
collectible
at
the
project
level,
at
least
for
this
current
release,
that's
I
think
then
yeah.
This
might
be
one
to
approach
to
solve
that
problem.
I.
A
Think
that
makes
a
lot
of
sense
as
well
to
break
break
it
up.
I
think
I
also
have
comments
on
writing
a
blog
post
on,
like
sig
cluster
lifecycle,
update
what
are
the
new
stuff
and
so
forth.
Probably
it's
gonna
be
part
of
like
five
days
of
kubernetes,
one
eight.
If
that's
the
thing
I
don't
know
yet
anyway,
I
yeah,
just
just
bring
me
one
of
those
through
I'm
gonna,
collect
them
and
all
basically
it's
like
upgrades
and
extensibility,
or
the
two
major
for
cupid
image.
Please.
B
B
G
G
So
I'm
from
one
of
the
main
taters
of
Cordia
nests
for
another
ciencia
project
and
one
of
the
big
things
we're
doing
is:
we've
built
an
integration
with
kubernetes
to
be
the
in
cluster
DNS,
essentially
a
replacement
for
kube
dns,
and
we,
you
know
we're
a
single
process
where
it
means
ripping
go.
So
it's
a
bit
of
simpler
than
having
different
containers
at
QB.
Ns
runs
today
and
also
it's
you
know
we're
a
general-purpose,
DNS
server
as
well,
and
so
there's
a
lot
more
functionality
and
features
and
use
cases
that
you
can.
G
You
can
accomplish
with
it
than
you
can
with
cheap
DNS
and
we've
talked
to
Bowie
and
Tim.
Are
the
maintainer
zuv
keep
DNS
and
they're
open
to
the
idea
of
using
coordinates
there?
Although
we're
you
know,
there's
no
conclusion
yet,
but
as
part
of
that
process,
I
wanted
to
see
if
we
can
integrate
or
DNS
with
the
Cuban
moon
and
NIT
process,
or
some
other
other
way
that
you
could
deploy
optionally
to
play.
Accordion
ask
with
Cuban
men,
so
people
could
start
trying
it
out
more
easily
than
they
can
today.
G
G
We
have
looked
at
doing
a
new
flag,
2
cubed
moon,
admit
somebody
mentioned
feature
dates
when
he
goes
mad
and
then
I
think
Lucas
said.
Maybe
it
should
just
be
a
issue
in
the
communities
featuring
posit
or
easily
other
employment
tools
could
pick
it
up
and
can
do
the
same
thing,
but
I
wanted
to
get
the
thoughts
of
the
team
here.
G
D
So
they
starts
to
shred
into
a
little
bit
of
controversial
territory
in
terms
of
DNS
is
an
add-on
for
the
cluster
and
right
now,
Qube
admin
installs
it
because
it's
since
all
DNS,
because
it's
required
for
the
cluster
to
function
but
doesn't
actually
sort
of
take
ownership
of
add-ons
as
a
whole.
Right.
There
are
a
lot
of
things
we
consider
cluster
add-ons
that
cube
admin,
just
sort
of
ignores
and
allows
people
to
install
after
the
fact,
but
it
doesn't
solve
DNS
because
that's
actually
critical
for
this
system
to
function
so
I
think
you
know.
D
You
know
QB
NS
versus
core
DNS,
that
is
seems
unlikely
to
land
in
the
one
nine
timeframe,
which
is
why
I
said
you
might
not
be
happy
with
that
solution,
and
so
I
think
you
know
in
the
in
the
short
term,
if
this
is
something
that
people
want
to
play
with
like
making
a
flag
or
a
feature,
gate
seems
reasonable,
but
I
do
think
that
in
a
long
term
like
we
should
be
solving
this
of
the
core
proper
way
using
the
add-on
mechanism
that
we
haven't
yet
built.
What.
C
C
I
think
we
can
probably
decide
on
like
a
add-ons
flag
that
takes
a
set
of
strings
with
the
default
being
cue
DNS,
and
then
we
have
like
a
switch
statement
somewhere
that
says:
okay,
we
can
do
core
DNS,
and
this
is
all
done
in
code
versus
a
generic
system
right
and
then
over
time,
I
actually
sort
of
back,
and
we
can
mark
that
flag
as
alpha,
and
you
know
because
we
might
change
it.
But
that's
enough
to
is
hint
towards
starting
to
build
towards
a
more
generic
add-on
system
without
having
to
sort
of
swallow.
H
Like
that,
a
lot
I
also
I
wanted
to
like
kind
of
lean
against
the
idea
of
feature
gates,
I
think
they've
been
over
utilized
currently
and
the
way
the
binaries
are
have
been
smashed
together.
The
feature
gates
will
propagate
to
other
components
that
have
nothing
to
do
with
this,
so
it
becomes
confusing
the
folks
who
are
trying
to
rationalize
what
is
the
dependency
graph
in
this
component.
You
know
if
they're
not
familiar
with
kubernetes.
What
are
the
dependency
graph
of
this
feature
on
this
component
right,
Lucas.
A
Regarding
that
I
mean
I,
I,
don't
understand
why
feature
gates
were
implemented.
The
way
they
were.
Basically,
you
just
put
like
well
we're
out
of
that
process,
or
that
package
cube
atom
has
its
own
decoupled,
but
anyway
like,
if
you
add
the
city
gate
the
coop
rocks,
it
will
show
up
in
the
API
server
like
I
said
that
doesn't
make
sense
at
all
anyway.
A
H
Depends
upon
testing
a
Needleman
really
right
if
there's,
if
there's
enough
thumbs-up
and
people
feel
comfortable
and
it's
been
thoroughly
vetted,
then
I
think
the
answer
might
be,
maybe
as
long
as
the
key
stakeholders,
namely
Tim
I,
think
in
this
front,
we're
okay
with
it.
So
I
think,
there's
an
aspect
that
we're
kind
of
missing
here
to
you
is
that
we
need
to
enable
testing
of
this
yeah.
D
There
because
I
would
say
it
sounds
like
that:
John
has
gone
and
talked
to
the
Signet
working
folks
and
those
are
the
people
that
would
decide
if
this
should
be
the
default
or
not
right.
So
I
think
the
question
here
is:
should
we
allow
it
to
be,
can
turned
on
optionally
so
that
we
can
start
testing
and
he'll
think
give
them
data
about
whether
it
should
be
the
default
and
it
promotes
onset.
It
sounded
like
they
were
open
to
that.
D
A
G
G
A
G
A
Yeah
that
works
well
I
do
think
that
we
should
have
a
features
issue
for
this
anyway,
like
if
sig
network.
Besides
that,
this
is
the
way
we
want
to
go.
We
want
to
deprecated
what
we
have
in
Combinator,
slash,
DNS
and
use
core
DNS.
Is
that,
then
we
should
like
advertise
that
the
general
community
and
say
let
like
this
is
the
long-term
plan.
Here
you
can.
You
can
deploy
that
on.
However,
you
want,
but
cube.
Dns
is
going
to
be
the
like
core
DNS.
It's
gonna,
be
the
the
recommended
Adamo
to
use
weights,
kubernetes
I.
I
Think
so
I'm
writing
well
communicate
on
behalf
of
the
keys
may
seem
like
not.
Everyone
is
a
course
on
this
call.
But
can
you
hear
me?
Okay,
yeah,
okay,
I,
just
don't
see
like
some
wiggling
bar,
like
I,
do
in
hangouts
so
keeps
trying
has
been
around
for
a
while.
It's
almost
as
though
those
cops
not
as
cool,
of
course,
because
they
get
way
more
press.
But
it's
been
around
it's
kind
of
popular.
It
doesn't
use
cube
idiom
yet,
but
I'm
about
to
change
that.
I
I
have
some
code
and
review
it's
almost
merged,
but
we
have
to
get
some
things
shift
ourselves
around.
We
prefer
deploy
hypercube
because
it's
less
bits
to
download
and
so
on.
So
there's
a
bit
of
a
different
strategy
to
to
key
phrase,
approach
from
Cuba
DMS,
so
I'm,
trying
to
synergize
and
use
some
more
buzzwords
to
say
that
we
want
to
you
know,
do
things
the
cube.
I
Atm
way
still
have
some
blockers
related
to
DNS
and
basic
users
file,
but
basically
we're
gonna,
get
it
up
and
out
the
door
probably
next
week,
which
means
we're
more
or
less
satisfying
the
requirements
for
graduating
incubation
that
were
set
out.
Last
year
we
have
until
October
26
to
reach
that
path,
so
I'm
reaching
out
kind
of
in
advance
so
that
we
can
find
what
gaps
there
are
what
the
community
prefers.
A
D
C
C
C
C
I
mean
it's
okay,
so
I'll
tell
you
where
my
stand
on.
It
is
and
I'm
one
of
many,
and
some
of
this
is
going
to
change
as
we
get
this
during
community
committee.
You
know
in
place
I'm,
not
a
huge
fan
of
the
incubation
project
in
general,
I'd
like
to
see
kubernetes
the
organization
be
smaller
and
more
targeted,
I,
don't
think
we
necessarily
have
to
bless
these
things
as
much
as
poss
I.
C
Think
that
there's
plenty
of
room
for
folks
to
sort
of
form
their
own
project
projects,
their
own
associations,
I,
don't
think
we're
necessarily
going
to
have
a
sort
of
soup
to
nuts
official,
install
or
I.
Just
think
that
there's
too
many
variations
and
opinion
even
Qube
admin
is
on
the
bubble
and
there
are
people
who
are
not
a
fan
of
keybed
and
being
an
official
part
of
the
project.
C
C
Okay,
we
really
want
to
have
a
tool
box
that
enables
folks
to
do
a
lot
of
stuff,
and
so
because
we're
building
to
that
it
ends
up
being
sort
of
a
platform
level
thing
versus
an
end
user
facing
thing
so
I
think
I'm
really
uncomfortable
with
anybody
in
this
group
or
in
the
steering
committee,
actually
playing
favorites
and
making
you
know,
and
and
in
nineteen
winners
and
with
respect
to
how
to
actually
install
and
manage
kubernetes,
especially
since
there's
so
many
commercial
interests
across
everything.
That's
that's
happened
in
kubernetes.
C
That
being
said,
there
are
checks
like
cops.
There
are
already
in
that
situation,
and
so
we
do
have
this
sort
of
you
know,
we've
already
sort
of
started
biting
that
off.
There's
a
question
for
how
do
we
want
to
handle
this
stuff
moving
forward?
How
do
we
want
to
actually
establish
what
are
the
requirements
for
incubation
graduating
from
incubation?
How
do
we
list
and
recommend
these
things
and
we're
going
to
have
to
deal
with
the
fact
that
we
already
have
projects
like
cops
that
are
in
that
situation?
C
Now
that
are
my
opinion.
Just
to
be
clear,
there
are
other
other
folk
that
are
you
know,
leaders
in
the
community
that
have
different
opinions,
the
current
equation
project,
as
it's
documented
states
that
graduate
you
need
to
have
you
know
a
track
record
and
sponsor
from
the
top
level
owners
updating
that
taking
into
account
the
steering
committee
is
probably
something
that's
going
to
be
on
the
backlog
of
stuff
for
the
steering
committee
to
take
care
of
after
it's
actually
seated
after
the
elections
that
are
happening
right
now,
Joe.
I
Can
I
get
it?
We're
done.
So
sorry,
that's
that's
my
feel.
Okay,
we
shouldn't
pick
favorites,
although
cops
went
through
incubation
and
graduated
and
is
official
and
bypassed
incubator
projects
in
some
degree,
but
that's
okay
and
so
nothing's
been
announced.
The
incubation
has
changed,
there
are
a
list
of
criteria
and
the
state
should
approve
it.
The
state
can
still
set
requirements.
What
would
you
like
to
see?
I
C
Part
of
the
problem
here
is
there
is
thing
that
says:
SIG's
should
approve.
How
do
we
approve
it
right?
We
don't
have
enough
governance
and
sinks
to
actually
be
able
to
say
this.
Sig
approves,
or
what
have
you
we
have
sig
leads,
but
sig
leads
are
really
more
of
a
advisory
sort
of
facilitator
type
of
role
right.
We
don't
have
an
official
number,
a
set
of
members
that
can
vote
in
terms
of
actually
what
happens
inside
the
SIG's
and
so
a
lot
of
that.
C
I,
don't
know
what
to
tell
you
right
now.
Right
now
can't
tell
you.
Is
that,
like
over
time,
I
really
want
to
win
Clary
here
so
that
you
don't
feel,
like
you
know,
you're
you're,
working
with
the
process?
That's
not
clearly
documented
because
you're,
not
the
only
one
feeling
that
that's
the
one
thing,
though
the
other
thing
is
like
okay.
What
do
we
do
now?
My
Lu
again,
my
personal
opinion
is
I.
C
C
That's
already
official
I'm
not
going
to
stand
in
the
way
of
other
people
want
to
help
assist
to
sort
of
make
this
stuff
happen,
because
I
don't
feel
like
it's
my
place
before
the
steering
committee
is
in
place
to
really
sort
of
you
know,
try
and
change
things
I'm,
just
not
going
to
enable
it
personally
myself
to
be
able
to
do
that.
So.
D
I
have
two
questions:
Joe
one.
If
people
want
to
have
code,
that's
you
know
affiliated
with
kubernetes.
That's
you
know
own.
You
know
by
the
CNCs,
which
is
kind
of
what
you
get
by
being
in
the
official
repo.
Do
we
have
another
place
where
they
can
go
right?
You
said
people
can
set
up
their
own
products,
their
own
communities,
but
likely.
Where
is
that
set
so.
C
That
the
hold
of
all
this
ncf
thing
is
is
a
question
because
that's
a
two-way
relationship
and
I,
don't
think
we've
been
super
clear
in
terms
of
of
what
are
the
rights
and
responsibilities
for
having
like
who
has
access
to
the
cops
Twitter
account?
Is
it
the
CN
CF
Martin
I
know
like
Chris,
does
right
who
controls
usage
of
the
helm
logo
right?
We
don't
have
clarity
around
this
stuff
and
when
stuff
is
part
of
the
CMC
f,
it's
not
just
ownership
of
code.
It's
a
two-way
relationship
and
I.
C
Don't
think,
we've
been
super
clear
and
crystal
clear
about
that
stuff,
and
so
I
don't
want
to
have
you
know
kubernetes
through
the
CN
CF
sort
of
like
you
know,
have
a
bunch
of
sort
of
you
know
varying
degrees
of,
like
you
know,
special
agreements
between
different
products
right
and
I
think
we
I
think
we
have
a
mess
right
now
that
we
have
to
clean
up
I,
don't
know
I
mean
like
please.
My
gut
here
is:
let's
not
make
the
problem
worse.
C
Until
we
can
actually
really
understand,
you
know,
really
define
sort
of
a
a
you
know
workable.
You
know
system
moving
forward
that
I've
been
involved
in
like
like
caisson
it
as
a
community
project,
but
we
didn't
go
through
the
incubator
project
process
because
we
feel
like
we
can.
We
can
be
you
know
just
as
effective
without
having
to
actually
go
through
and
do
that.
We
have
people
submitting
code.
It's
an
Apache
2
license.
It's
DCO,
the
the
you
know
over
time,
like
that.
C
The
ownership
of
that
comes
so
commingled
that
it
might
as
well
be
like
nobody
can
do
license
it
or
anything
like
that.
So
it's
messy
I
just
don't
want
to
make
the
problem
worse
until
we
can
figure
it
stuff
out.
That's
my
personal
opinion,
though.
So,
like
you
know,
I'm
not
gonna
stand
in
the
way
of
it.
Okay,.
I
I,
don't
want
to
waste
a
more
time.
It's
a
mistake.
I
have
two
quick
questions
for
you:
Joe
one.
What
happens
on
October,
26
we're
in
action,
needs
to
this
project,
failing
provisions
and
are
you
the
source
of
authority
on
all
things,
akyuu
Batian?
Or
is
this
just
a
personal
opinion
or
secondhand
information?
This.
C
Is
my
opinion
I've
seen
other
projects
you
know
looking
to
graduate
incubation,
what
I
would
do
if
I
were
you
is
I
would
send
mail
to
the
like,
so
number
one
is,
if
you
can
find
somebody
else
in
the
top-level
owners
to
actually
help.
You
know
sponsor
this
thing
through
have
a
sponsor
just
a
vacation,
so.
C
So,
if
Brian's
willing
to
Robertson
what
we're
standouts
are
Brian
Brian,
alright
I
mean
like
I'm,
not
gonna,
stand
in
the
way
and
try
and
walk
Brian
from
doing
this
stuff.
Fine
doesn't
have
the
same
concerns
that
I
do
around
this
stuff,
but
I'm,
not
gonna.
You
know
I'm,
not
gonna,
like
the
other
thing.
That
I
think
is
totally
reasonable.
For
you
to
do
is
to
say,
hey
look.
We
have
this
clock,
that's
ticking
with
respect
incubation.
There's
a
lot
of
stuff
in
flight
with
the
steering
committee
can
I
at
least
get
some
approval.
C
That
hey.
Can
we
put
some
of
that
clock
on
pause,
and
so
we
can
get
some
more
clarity
from
the
steering
committee.
That's
something
that
I
would
be
totally
willing
to
support.
Hey
yeah
I
mean
I.
Just
you
know,
I,
don't
think
we
have
clear
criteria
in
terms
of
what
it
means
to
actually
graduate
in
terms
of
responsibilities
and
rights
and
I.
Don't
think
that,
like
we've,
pretty
much
just
rubber-stamp
everything
and
I,
don't
like
where
that's
taking
us
at
the
project,
but
again,
that's
me
personally,
right
I'm,
not
I'm,
not
I'm,
not
the
one!
D
C
So
yeah,
if
you
want
to
Santa,
sent
you
no
mail
to
the
steering
committee,
saying
hey
I'm
in
the
incubation
project,
there's
a
lot
of
stuff
going
on.
There's
this
there's
this
arbitrary
clock.
Can
we
put
that
on
hold
while
we
figure
stuff
out?
You
know
this
Mike
with
Robert,
saying
that
I
sound
terrible
this
Mike
with
zoom,
it's
just
with
zoom,
because
I
do
mind
my
TGI
K
with
this
thing
and
it
works
fine,
so
I'm,
gonna
switch.
Sorry
I
would
totally
support
going
ahead
and
doing
doing
that.
D
C
E
C
A
E
Cool
drain,
you
talk
louder,
no,
okay.
So
after
a
faith,
you
know
a
lot
of
the
same
frustration
that
we
were
just
talking
about
in
cops,
I
decided
to
write,
Cuba
corn
to
get
me
out
of
all
this
stuff,
so
Cuba
corn
is
not,
and
more
than
likely
will
not
ever
be
official
kubernetes.
Anything
it's
mine,
I
make
all
the
rules.
If
you
don't
like
it.
E
Sorry,
not
sorry,
so
that's
kind
of
what
I'm
dealing
with
now
and
I'm
giving
an
update,
because
we
are
getting
a
lot
of
interest
in
it
and
a
lot
of
people
are
moving
over
to
the
project.
And
you
know
people
are
even
it's
a
good
idea.
I
guess
so,
where
we're
at
is
I
have
intentionally
said.
We
are
not
going
to
make
a
release
until
we
have
a
core
set
of
features
done
and
testable.
E
Most
of
those
are
done,
but
there's
a
handful
of
outliers
I'm,
waiting
to
get
sort
of
coded
and
kind
of
hardened
a
little
bit
from
people
just
banging
on
the
codebase,
and
those
big
features
are
I
want
to
have
an
azure
and
terraform
implementation
that
actually
uses
the
terraform
provider
to
deploy
at
kubernetes
on
how
you're,
using
the
same
API
that
were
using
for
everything
else.
Right
now.
We're
gonna
talk
more
about
the
API
later
just
go
with
me
for
now.
E
I
want
to
have
an
intimate
testing
suite
for
all
four
of
the
major
cloud
providers
that
we
support:
digitalocean,
Amazon,
Rodger
and
Google,
and
when
I
say
end-to-end,
test
I
mean
spin
up
a
cluster
hit,
the
API
validate
nodes
and
all
the
infrastructure
resources
are
there
as
expected
and
then
rip
it
down
and
verify
that
they
all
got
deleted
as
expected,
and
probably
do
that
a
couple
of
times
we're
also
implementing
a
new
feature
called
sets.
That
is
identical
to
exactly
how
helm
does
set
that
basically
says
at
runtime.
E
E
So
those
are
the
big
ones
I'll
step
over
implementing
when
FTO
czar
so
again
like
this
is
my
idea,
and
it's
crazy
and
most
people
disagree
with
me
and
that's
why
it's
my
project
and
not
in
kubernetes
incubator,
but
we're
going
to
attempt
to
allow
a
user
to
grab
a
copy
of
the
application
stuff
using
FD
a--'s
arc
and
also
grab
a
copy
of
the
infrastructure
stuff
using
cubic
or
and
stick
all
that
into
one
monolithic,
tar,
ball
and
say:
here's
your
kubernetes
cluster.
E
You
know
top
middle
and
high
layers
and
or
top
middle
and
bottom
layers,
and
you
can
go
and
pass
that
off
to
another
user.
They
will
be
able
to
recreate
your
infrastructure
and
your
apps,
and
if
you
wanted
to
tweak
some
sort
of
like,
maybe
you
want
to
call
it
a
different
name
or
you
want
to
run
it
in
a
different
cloud
or
you
want
to
adjust
your
note
pools
or
whatever
you
could
do
that
as
well.
So
those
are
kind
of
the
big
blockers
for
version
1.0
again,
we
purposely
are
not
doing
it.
E
Release
right
now,
because
I
kind
of
went
down
the
story
with
cops
would
be
released,
version
and
I.
Think
5.4
I
want
to
say
and
all
of
a
sudden
our
feature
cadence
came
to
a
halt
because
we
were
so
busy
supporting
bugs
that
we
even
actually
get
to
code
need
more
features
in
so
trying
not
to
make
that
same
mistake
again.
E
Lucas
had
mentioned
something
about
in
doing
testing
in
Amazon
I
have
some
ideas
and
opinions
around
that,
and
we
certainly
can
look
at
what
it
would
take
to
bootstrap
a
cluster
with
cube,
ATM
and
Amazon,
and
we
have
some
pretty
large
amounts
of
flexibility
as
far
as
what
we
want
to
test
so
any
way
we
can
help
out.
I'm
happy
to
do
that
as
well.
Yeah.
A
Basically,
I'm
I've
been
looking
into
how
federated
testing
works
like
getting
my
head
around
like
what
JSON
files
needs
to
be
where
and
what
artifacts
upload?
What
metadata
to
use
all
that
it's
currently
like
it's
not
documented
anywhere
as
far
as
I
know
so
I
just
started
looking
at
the
like
source
code
that
scrapes
it
and
then
also
like
the
existing
data
that
that
we
have
I
think
I
have
now
a
pretty
good
idea.
What
it
looks
like
and
I'm
currently
working
with
suggesting
to
like,
probably
eventually
get
some
real
docs
in
place.
A
Maybe
even
a
criminalist
at
I/o
like
here
is
what
you
need
to
do
to
upload
the
things
it's
it's
testing
in
cluster,
so
I
mean
it's
just
a
regular
image
with
cubes
ETL
and
ETA
dot
test.
So
I
you
just
submit
the
job
to
the
cluster
and
it
will
work
so
yeah
I
mean
it
would
be
cool
to
have
some
kind
of
more
IDI
testing
for
cubed
cubed
I
mean
that's
nothing
core
right,
so
I
mean
it
could
show
show
up
in
a
test
grid
or
if
we
ever
find
something
like
community
test
grid.
I.
A
E
Yeah
I
think,
though,
I
think
that
would
be
great
I
mean
if
we
just
want
to
demonstrate
that,
like
we
have
a
handful
of
instances,
they
can
route
to
each
other
and
we,
you
know,
execute
on
the
command
line.
Err
in
the
back
script,
cube
ADM
your
flags
here
and
then
you
know
next
thing
you
know
we
can
hit
the
EPI
yeah.
That's
a
really
really
valuable
test
to
just
run
and
I.
Think
it's
relatively
cheap
as
well.
A
Yeah
I
mean
I'm
I'm,
working
on
IDI
testing
for
cube
admins,
specifically
as
well,
and,
of
course
it's
like
obviously
under
tested,
like
I
mean
we
have
it's
getting
better,
but
I
mean
still
any
anything
like
that.
The
slitting
the
cubed
M
does
its.
We
need
like
a
hundred
test
suits
and
it
would
be
too
little
anyways.
So
if
we
could
somehow
like
that's,
why
federated
testing
is
so.
A
E
That's
the
point.
So
anyway,
there
is
a
pull
request
in
the
cubicle
RN
pull
request
log
that
I
think
Jacob
put
together.
I've
got
some
comments
in
there
from
me,
and
Zack
and
Robert
I
have
not
had
a
chance
to
really
give
it
as
much
time
as
I
would
like
it.
Every
time
I
start
to
sit
down
I'd
like
a
few
minutes
like
no
I,
need
to
put
more
time
aside
to
really
think
about
it.
E
But
the
good
news
is:
is
that
there's
a
really
awesome
like
if
you
space
in
there,
where
it
says
the
cube?
Adium
API
goes
here
which
is
effectively
a
representation
of
the
control
plane,
and
then
we
have
this
other
concept
of
like
I
said.
If
you
don't
know,
pools
anyway
feel
free
to
go.
Look
at
the
PR
commented.
E
We
would
love
to
hear
your
opinions
and
I
think
this
is,
for
all
intents
and
purposes,
are
working
like
demo
of
the
direction,
we're
kind
of
all
hoping
the
api
goes
and
then,
after
we
sort
of
hammered
that
out,
do
a
proof
of
concept
start
creating
infrastructure
and
all
of
the
major
clouds.
Then
we
can
say:
hey
we've
done
this,
we've
thought
about
it.
We
have
API
overall
happy
with
it.
Now,
let's
go
and
propose
this,
get
it
through.
Whatever
incubation
or
not
computation
process,
there
may
or
may
not
be
at
that
time.
I.
D
Just
to
jump
on
there
also
like
I,
know
Lucas
you've
been
really
pushing
to
try
and
get
like
some
sort
of
cube
admin
API
out
of
alpha
and
I
think
some
of
the
reason
that
we've
been
hesitant
in
terms
of
jumping
behind.
That
is
because
we
really
want
to
think
about
it
more
holistically
in
terms
of
the
overall
cluster
API,
because
it
is
a
critical
component
of
the
cluster
that
we
can
describe
the
control
plane.
D
But
I
would
also
love
for
that
description
of
the
control
plane
to
be
consistent
between
people
who
are
using
a
cluster
API
to
describe
their
entire
infrastructure
and
people.
Who
are
you
just
using
do
Badman
directly
so
that
it's
portable
back
and
forth
I?
Think
having
that
that
description
be
the
same,
is
gonna
be
really
powerful
and
we
just
need
to
make
sure
we
get
it
right
so
that
it
satisfies
all
of
those
use
cases
and
so
I.
D
Think
people's
haven't
had
the
time
to
like
look
at
the
cube
API
that
you
started
working
on
and
try
to
figure
out
what
it
would
look
like
to
make
it
fit
into
the
overall
cluster
API.
And
what
it
would
look
like
standalone
for
cube
admin,
but
I
think
we
need
to
make
sure
we
satisfy
both
use
cases
and
so
I
think,
as
we
keep
pushing
on
the
cluster
API
that'll.
Also
the
momentum
behind
the
queue
when
API
and
hopefully
we
can
get
it
further
along
in
the
next
release
cycle.
A
D
Think
Chris
had
talked
about
setting
up
a
working
group.
I,
don't
think
we
ever
set
up
an
email,
less
lists.
We
don't
have
official
like
weekly
meetings
that
are
recorded
and
posted
to
the
sig
channel
or
anything
we're
mostly
just
sort
of
coordinating
on
slack
and
sort
of
working
through
the
cubicle
and
repository
as
a
place
to
test
out
what
like
whether
we
think
this
API
is
sort
of
the
right
shape
and
again,
like
she
said,
there's
sort
of
a
big
hole
in
Jacob's
PR.
D
A
I'm,
basically,
like
you
know
that
I
I
mean
I,
would
love
to
move
forward
on,
as
you
said,
like
a
cube,
a
cube,
medium
API,
which
is
basically
a
representation
of
the
control
plane.
I
should
be,
and
yeah
I
mean
just
how
can
folks
get
involved
in
this
process
like
in
order
to
make
that
happen,
probably
like
make
alpha
2
of
the
cubed
M
API
in
one
line
or
whatever,
to
get
a
proof
of
concept,
then
whatever
hopefully
pushed
be
dine-in
110
or
whenever
we
feel
ready
right?
D
D
C
Ideally,
working
groups,
sort
of
cut
across
different
SIG's,
so
I
think
there's
there's
a
little
bit
of
a
question
around
that
I
mean
this
does
cut
across,
maybe
some
of
those
from
a
cloud
provider
or
specific
SIG's,
as
we
start
looking
at
the
cloud
specific
stuff.
So
maybe
that's
the
motivation
for
for
pulling
it
out.
Yeah.
D
And
I
think
that
the
main
thing
is
like
you
know,
Chris
has
mentioned
when
we
started
this
process
that
you
know
a
lot
of
people.
It's
a
gay
WS
were
already
talking
about
this
and
when
sick
guys
were
started,
they
were
already
talking
about
this
and
I.
Think
the
same
thing
is
gonna
happen
when
sig
GCP
starts,
and
so
it
really
is
sort
of
like.
Can
we
pull
some
of
those
people
who
may
not
come
to
this
meeting
into
that
working
group.
A
Also,
we
have
a
cloud
provider
refactoring
Burcu
starting
tomorrow.
That's
indeed
cross-cutting,
but
anyway,
I
mean
it's.
It's
not
exactly
the
same
as,
like
your
work
groups
goals
right.
We
would
like
to
design
a
cluster
API
versus,
what's
starting
tomorrow,
10:00
a.m.
Pacific
Time,
whose
goal
is
like
three
power
of
the
cloud
providers
out
of
core
and
try
to
do
something
better
I
mean
and
these
all
little
girls,
as
what
Joe
mentioned
earlier,
like
what
is
core.
A
What
is
not
and
there's
a
lot
of
people
thinking
about
it
and
also
like
what
should
we
do
with
cloud
providers?
The
work
groups
goals
right
now.
It's
like
make
it
possible
to
run
your
club
forward
out
or
go
fast.
Then
we
can
have
the
discussion
like
what
should
happen
with
the
existing
ones,
etc.
So,.
A
A
So
right
now
we
have
on
the
documentation
of
Cuba
in
the
Cuban,
my
commentation,
that
we
list
our
implementation
alpha
as
we
at
the
one.
Seven
time
knew
that
bootstrap
tokens
would
change.
We
knew
that,
like
some
resource
names,
we're
in
the
right
ones,
we
didn't
like
use.
The
right
are
back
names,
for
example,
had
some
mistakes
there,
it's
Ezra
should
we
declare
with
bead
and
now,
when
we
have
B
the
bootstrap
tokens
and
all
the
like
reefs
or
names
resource
names
are,
as
they
should
and
have
upgrades
in
place.
A
A
A
A
Basically,
there's
a
new
when
in
the
host
path,
struct,
that's
a
new
field
called
type,
and
that
has
a
bug
which
makes
cubelets
panic
at
head
and
fortunately,
the
queue
bed
and
Elise
that
I
set
up
last
week
caught
this
and
haven't
been
green.
So
far
due
to
this
bug,
we,
the
author
of
the
feature,
is
were
in
it
and
I.
Think
it's
it's
gonna
be
fine
anyway,
that's
mostly
six
storage,
but
I
had
some.
A
K
D
K
A
E
I'm
sorry
I,
don't
know
what
my
microphone
is
being
weird
today,
so
yeah
I
was
just
wondering
if
we
had
talked
about
the
future
of
cube
ADM
in
regards
to
high
available
kubernetes
clusters.
Recently,
we.
H
Have
beaten
this
horse
sufficiently
to
death?
Okay,
there's
a
there's,
a
tentative
plan
that
I
think
Lucas
could
probably
elaborate
on
that
he's
been
we've
been
talking
about
on.
We
have
a
call
tomorrow
that
we
do
just
a
chase
off
hosting
and
upgrades
which
we
could
probably
go
into
more
detail
there
and
then,
because
we're
actually
out
of
time,
I
think
right
now,
yeah.