►
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
I'll
repeat
it,
which
is
we
put
up
a
doc
or
I,
wrote
up
a
doc
trying
to
summarize
the
blockers,
everyone's
blockers
and
wishlist
items.
It
is
knowingly
incomplete
in
at
least
two
ways.
If
you,
if
you're
a
blocker
or
wishlist
item,
is
not
even
represented,
please
go
and
add
it,
and
also
I've
tried
to
put
underneath
each
one.
You
know
a
list
of
the
tools
and
whether
it
is
for
them.
It
is
a
blocker
or
a
wish
list.
B
So
please
I
did
that
for
cops,
I
hope
accurately
and
for
Cooper
corn
and
the
CNC
a
project
I'm
sure
less
accurately.
I
did
not
do
it
for
anyone
else.
So
please
go
through
and
add
your
your
plus
ones,
your
votes,
so
you
know
we
can
get
idea
which
ones
are
actually
most
important
to
people
and
whether
they
are
actually
just
you
know,
just
on
the
cost.
If
it's
just
a
cop's
wishlist
item,
it's
very
not
a
priority
right,
but
if
it's
a
blocker
for
three
people,
that's
much
more
of
a
priority.
C
B
B
I
wish
this
is
I,
think
it
would
be
good,
but
if
we
can
adopt
who
they
excavate
am
without
it,
although
it
would,
we
think
it
would
be
a
good
thing
to
be
in
Canadian,
whether
it's
simply
because
it
would
make
us
more
comfortable
or
whether
we
think
it's,
because
it
would
be
additional
that
cube
ATM
should
do
rather
than
you
know,
cops
doing
and
googling
going
energy
cake.
Okay,.
B
B
I
think
yeah,
if
you
just
copy
and
paste
the
paragraph
and
edit
it
and
put
it
at
the
bottom,
I
guess
and
we'll
slide
it
back
into
place
based
on
the
votes.
And
if
you
do
at
the
bottom,
then
I
will
not
miss
it
as
I
read
through
them
now
so
like
the
resiliency
or
non
happy
path
into
ATMs.
So
obviously,
when
everything
works,
everything
works,
but
the
cases
when
it
doesn't
work
like
what
is
the
behavior,
the
insects
around
like
system
g-tube
that
hanging
and
has
to
be
named
recover
from
that
being
pulled
out.
B
B
B
It
came
up
that
cuvee
diem,
basically
bundles
a
set
of
versions
together
instead
of
components
instead
of
versions
and
that
for
some
people
is
actually
from
attic.
Particular
if
you
want
to
test
like
different
skewed
versions
of
the
project.
That
obviously
has
some
challenges,
because
you
know
we
don't
test
particular
combinations,
but
that
is
that
is
an
issue
support
for
cloud
providers
running
out
of
tree,
so
in
general,
we're
trying
to
pull
the
crab
spiders
out
of
coop
controller
manager.
B
B
Do
we
have
h
a
support
at
all
and
then
in
it
join
is
the
notion
of
like
the
model
for
initializing
it.
So
we
could
have
h
a
support.
This
still
requires
you
to
do
leader
election.
The
cops
so
kacica
probably
will
do
leader
election
itself
figure
out
how
to
do
in
exactly
one
of
those,
and
so
we
could
have
h
a
support
without
what
keeping
one.
D
B
B
A
That
well
must
be
spins
up
first
and
have
to
say:
does
the
third
generation
or
all
these
things
spins
up
the
first
API
server
uploads?
Maybe
something
I
don't
know,
set
up
orbital
and
then
most
a
and
C
comes
up
some
minutes
later
and
then
they
have
to
realize
that.
Well,
we
have
reader
and
then
do
cubed
M
join
if
master
or
something
like
to
hook
up
with
with
the
must
be
so
I.
Think
that's
the
thing,
but
I
think
also.
We
can
do
this
pretty
easily
by
requiring
external
a
TD
cluster
with
this,
but.
B
D
Together,
yes,
as
q,
medium
defines
what
you
know
these
interfaces.
It
begs
the
question:
what
is
cuvee
DM
and
was
could
again-
and
it
seems
like
it'd-
be
good
to
fill
that
if
it's
sort
of
expertise
around
what
the
correct,
if
the
I
service
lines
over
run
or
controller
manager
to
run
that's
really
different,
then
you
know
opinionated
command-line
tool
that
has
all
these
things
so
I
I
guess
it's
more.
It
brings
up
an
interesting
if
you
make
all
of
these
components
pluggable.
What
are
we
actually
agreed
on
I.
B
Don't
know
that
it's
pluggable,
or
is
it
pluggable
or
more
just
I?
Guess
it
consumable
through
code
rather
than
through
the
like,
exacting
it's
one
of
the
asked,
and
the
other
asked
is
the
ability
to
run
in
it
on
all
three
nodes
of
a
master
right,
I,
don't
think
those
change!
The
idea
that
you
radium
is
opinionated
I
do
think,
though,
in
general,
that
you
know
we
would
like
the
gold.
B
Successful
icicles
always
been
to
get
to
make
sure
when
it
is
easier
to
install
to
pull
things
into
the
core
and
QA
diem
is
maybe
a
stopping
ground
for
things
that
are
going
to
end
up
in
the
core.
A
component
comes
big.
All
the
work
that
we're
doing
for
component
configuration
will
hopefully
get
out
of
cops,
go
out
of
cube
ADM
and
go
into
component
company
in
the
core
and
I.
Don't
I
think
that's
been
the
plan
for
longer
than
I'm.
B
A
Yeah,
that's
kind
of
that
database,
but
yeah
the
cubed
M
package.
Now
it
is
that's
actually
not
depending
but
yeah,
it's
just
a
pre-flight
check,
but
in
a
way
that
the
main
value
of
cubed
M
is
is
like
cubed.
M
knows
what
the
best
practices
are
for
every
reversion
and
provide
a
consistent
interface
to
interact
with,
like
I
mean
you
run,
cubed
M
on
on
one
six
and
you're
going
to
get
some
functionality.
B
I
mean
I
think
it's
certainly
about
a
bad
point,
let's,
let's,
let's
finish
the
list
out
of
that,
and
then
we'll
come
back
to
with
more
opportunities
for
questions
about
how
about
that.
So
the
the
necklace
was
the
the
security
or
a
secure
alternative
to
bootstrap
tokens.
Tokens
are
indeed
insecure
or
cannot
be
made
secure.
Lucas
said
there
might
be
some
magic
fact
that
was
sort
of
a
cox,
a
problem
for
cops
where
you
know
mostly
API
servers
on
the
Internet.
B
D
A
And
if
kiddies
for
validation
and
it
can
be
used
for
authentication
and
it's
a
validation
part
you're
concerned
at
and
of
where
we
have
the
clustering
for
Big,
Mac,
public
and
yeah
I
mean
we
don't
have
rate-limiting
on
the
API.
So,
for
example,
so
yeah
I
understand
that
and
it
will
be
hopefully
already
in
1/8.
Maybe
in
one
line
I
mean.
B
B
B
B
C
I
XP,
but
not
really
quick,
I'm
happy
to
do
the
testing
outside
of
the
cube
ATM
official
repose
its
we.
If
I
you
just
be
given
an
avenue
to
like
report
back
data
on
specific
releases.
If
that
makes
it
any
here,
I
know
I'm
solutioning
already,
but
I
just
wanted
to
throw
that
out
there.
Yet.
A
E
B
I
might
be
you
guys.
Fine
next
up
is
a
cluster
API
which
actually
I
think
we've
discussed
a
little
bit
in
sick,
bastard
lifecycle,
but
the
idea
of
an
API
II.
Well,
there's
the
qadian
portion,
there's
the
non
forbidden
portion,
but
we
certainly
discussed
the
idea
of
an
API
that
controls
your
infrastructure
or
defines
your
infrastructure
and
could
be
used
to
control
it
and
I
guess
this
is
visual.
So
the
notion
that
do
we
also
want
an
API
for
Kubb
ATM
itself.
I
guess
is.
D
A
question
if
I
wrote
this
on
the
dot.
This
video
is
currently
not
concerned
with
provisioning
and
I
like
that
layering
of
not
having
it,
you
know
something
else
should
provision
the
machines
and
cuvee
diem
should
just
so
I'm
a
little
confused
about
how
this
like
what
what
scope
this
is.
Adding
to
video
I
think.
C
C
C
A
This
is
more
like
documenting,
documenting
the
like
integration
points
between
these
two
projects
with
the
clustering
cluster,
API
spec
or
whatever.
We
call
it
and
cubed
n.
It's
not
going
to
add
anything
really
like
on
a
code
basis.
It's
not
going
to
add
anything
to
cube
it
in
it's
more
like
we
recognize
that.
Well,
it's
like
the
layer,
one
like
cubic
corn
or
whatever
that
implements
the
cluster
API
might
just
use
cube
am,
and
we
should
just
make
sure
that
it
plays
well
together
or
something
like.
D
C
B
Thank
you
for
that
and
yes,
sir,
then
that's
yeah
and
then
the
next
one
was
the
ability
to
easily
run
Federation
on
top
of
tube
ADM
Custer's
again
I
think
this
is,
you
know
we're
not
going
to
like
subsume
coop
said,
but
whatever
integration
we
can
do
to
make
that
easier.
B
E
B
There's
certainly
a
there's,
certainly
something
where
the
the
specific
topic
we
just
changed.
Where
we
had
the
clusters.
Api
right,
that's
gets
so
cubes
or
Federation
has
its
own
notion
of
clusters
and
other
notion
of
clusters,
and
so
I
I,
don't
know
honest
I,
don't
I
didn't
I,
don't
know
what
the
someone.
If
someone
has
more
details,
Lucas.
B
A
Haven't
checked,
I'm,
not
100%
sure,
but
I
I
talked
briefly
to
module
and
he
thought
it
would
be
pleased
that
it
would
be
just
an
addition.
On
top,
like
Frankie
wait,
a
minute
in
three
different
zones
on
Google
cloud
and
then
add
some
notes
them
and
then
just
run
cube
head
with
Google
cloud,
DNS
integration
and
point
to
these
three
cubed
and
minuted
clusters.
A
That's
what
we
think
is
the
case
at
least
but
I
mean
it
totally
makes
sense
to
have
this
like
item
here
and
just
recognize
that
this
is
something
we
really
should
document
and
when
talking
to
matter.
This
is
also
something
we
should
like
actually
test
and
have
mean
we
should
set
up
GC
test
for
this
and,
as
part
of
moving
from
like
cube
up
yeah.
E
I
guess
because
there's
a
news
email
thread
about
the
fact
that
the
Federation
tests
reuse
some
of
the
cube
up
code.
I,
don't
want
that
to
be
completed
with
this
right.
So
the
fact
that
they're
reusing
some
the
code
doesn't
really
matter
it's
our.
Is
there
something
the
way
we
configured
g5
in
clusters
that
prevents
them
from
federated
them?
And
if
so,
we
should
fix
that
and
if
not,
then
the
Federation
should
be
purely
layered
on
top
yeah.
A
B
I
guess
that
leads
to
I
guess
our
next
one,
which
is
more
documentation
in
general
and
I
guess
this
would
be
one
where
it'd
be
great,
like
someone
could
try
writing
up
the
docs
for
like
I,
try,
running
curating,
went
through
clusters,
Erin
coop
side
and
roll
work
story
ously
or
who
them
something
failed
horribly
right,
but
in
general,
like
just
more
documentation,
is
a
we've
mentioned.
I
said
no
one
actually
put
it
as
a
wish
list
or
a
blocker,
but
it
was
just
a
general
like
the
malaise
it
opens
doors,
every
McCoy,
documentation,
yeah.
B
F
Clients
get
get
frustrated
with
running
stale
software
and
all
these
things,
so
you
have
to
somehow
work
on
actually
upgrading
their
infrastructure,
while
keeping
all
the
pots
and
containers
like
online.
So
this
one
thing
that
it's
a
painful
for
us
and
and
yeah
coming
here
that
it
will
be
addressed
at
some
point.
Thank
you.
Yeah.
E
F
And
yeah,
as
I
said
like
they,
we
have
great
process
is
not
like
upgrading
said.
This
is
like
I,
have
something
critical
running
on
the
pods
and
I
want
them
to
be
online
because
users
are
using
it
hitting
it
and,
like
my
client,
editing
and
all
the
things
and
somehow
I
won't
be
all
the
underneath
layers
to
be
to
be
upgraded,
kubernetes
in
principle
and
does
handle
this
case
because,
like
you
know,
if
you
kill
some,
if
you
kill
a
worker,
the
post
will
get
my
creative
and
you
have
replication
set
and
all
these
things.
F
So
if
you
have
set
up
your
your
infrastructure
correctly,
you
should
be
able
to
handle
a
bigger
place,
but
on
the
coop,
because
they
could
bad
mean
basically
somehow
hides
the
complexity
of
setting
up
infrastructure
and
it
has
to
document
a
it
has
at
least
document
what
am
I
going
to
do.
If
I
want
to
go
from
one
seven
to
one
eight
and.
E
And
that
the
fact
that
it's
out
of
scope
is
exactly
why
I
think
the
cluster
API
effort
is
important
is
because
this
is
a
problem
that
we
sort
of
want
to
be
able
to
solve
once
and
let
everyone
reuse,
like
we've,
done
a
lot
of
work
on
solving
exactly
that
problem
on
gke.
But
it's
like
proprietary
code
that
can't
be
reused,
which
is
really
a
shame
like
it's.
E
B
G
Sorry
I
wasn't
here
last
week:
it's
just
there
just
a
couple
missing
holes
in
the
phases
for
a
form
like
say:
writing
the
static
pod
manifest,
but
the
reason
we
want
to
do
this
is
we
don't
like
the
idea
of
say
the
master
being
able
to
sign
cert.
So
we
want
to
be
able
to
do
the
cert
distribution
ourselves.
Maybe
we
could
move
later
on
to
a
delegating
that
to
assert
approver
or
cert
signer
through
the
API
server,
but
right
now
we
don't
have
that.
So
we
want
to
be
able
to
take
that.
G
Other
things
like
mark
the
master
will
just
fail
because
we
don't
register
our
master
node
with
the
cluster
at
all
and
we'd
like
to
be
able
to
basically
do
the
entire
process
but
skip
that,
and
we
also
do
something
special
with
that
on
some
like
to
handle
that
ourselves
or
move
to
an
add-on
manager
project.
That's
I
think
still
in
the
incubation
days,
basically
more
fine-grained
control
over
the
steps,
so
yeah.
A
A
So
you
can
generate
everything
like
from
your
special
CA
somewhere
else
or
somewhere
else,
and
then
run
your
custom,
cert
approver
and
an
signer
most
of
them,
maybe
just
the
the
signer,
because
it's
requires
to
see
a
key.
It
should
be
run
somewhere
else.
A
So
I
think
that
you
skate
well,
that's
absolutely
something
that
should
be
documented,
even
better
I
think
we
have
some
comments
in
a
doc,
maybe,
but
not
like
this
full
guide
or
blog
post,
where
we
can
say
that
this
is
exactly
how
you
should
do.
You
should
run
these
OpenSSL
commands
pick
up
like
really
private,
CI
key,
then
the
CCP
over
these
files
into
these
directories.
Thank
you
better
minutes,
and
it
will
just
work.
A
So
if
you
want
to
run,
if
you
just
want
to
do
some
certificates
generate
certificate,
you
run
cubed
and
face
certificates
all
or
you
run
the
certificates
just
like
API
server,
signing
search
so
having
search
or
something
Oh
or
if
you
just
want
to
generate
a
cube,
config
file
or
if
you
just
want
to
do,
manifests
or
whatever
it
will
be
possible.
We
are
this,
but
it
will
I,
don't
think
it's
worthwhile
to
add
the
cube
am
in
it
to
skip
our
things
as
well.
G
Api
would
work
fine
for
us,
it's
just
not
fully
and
when
fully
implemented
from
what
I
could
tell,
and
so
the
blogger
is
for
that
to
be
finished
like
I
did
not
see
a
face
up
command
to
write
the
static
pod
manifest,
which
is
currently
our
biggest
pain
point
on
GG
aid,
and
we
want
to
get
rid
of
the
templates
that
we
have
and
actually
just
use
cube
ATM
to
generate
them.
Yeah.
A
G
D
Just
know
that
we
basically
discussed
making
all
the
phases
skippable
also
maybe
not
having
a
public
config
map.
What
I
think
this
is
a
good
opportunity
to
sort
of
talk
about.
You
know.
To
what
extent
does
queue
radium
become
so
pluggable
that
it
meets
everyone's
needs,
but
also
isn't
released?
Doing
like
everything
pretty
much,
we've
discussed
a
lot
about
where
we
need
all
of
these
before
the
goal
and
I
think
it'd
be
very
helpful.
A
There
might
be
well
basically
like
the
end
goal
is
have
one
face:
do
pretty
much
one
thing
like
atomically
and
hopefully
also
like
a
documented
route,
that
this
is
what
cube
am
actually
does
and
expects
like
in
the
next
next
place.
I,
don't
think
Cubana
should
be
broken
down
into
smaller
parts
of
the
faces.
A
D
Give
an
example:
so
blue
kuba,
while
back
made
this
choice,
that
they
we
weren't
going
to
do
in
a
templating
or
rendering
like
we
have
a
render
phase,
but
people
don't
use
it.
Tektronix
doesn't
not
use
the
render
page
to
generate
manifest.
It
provides
its
own
so
Google
long
time
ago
decided
it's
not
going
to
be
intimidated
about
any
of
that
or
how
you
grew
up.
Yeah,
basically,
most
of
the
worker
configuration
of
the
master.
It's
just
going
to
do
self
hosted.
D
So
if
you
use
bukuvu,
you
know
you're
getting
self
hosted
and
it's
your
job
spring.
Everything
else
at
the
point
where
your
ATM
is
being
so
pluggable
that
it's
allowing
you
to
you
know,
skip
phases
and
do
all
that
kind
of
stuff.
I.
Think
it's
important
to
highlight
like
what
are
the
phases
that
work?
The
Canadian
is
not
going
to
let
anyone
skip
like
what
is
the
essence
of
refugee,
a
set
you
can't
skin
ever
because
that's
intrinsic
to
the
value
proposition
of
the
tool
so.
A
I
don't
think
like
with
cubed
M
in
it
you're
not
going
to
be
able
to
skip
anything
at
any
time.
That's
just
going
to
do
this,
like
full,
full
thing
full
meal
for
you,
but
with
the
faces
like
well.
If
it
basically
becomes
your
responsibility,
I
guess
I
mean
if
you,
if
we
have
six
or
seven
different
places
and
you
choose
to
not,
then
you
have
to
write
like
have
seven
CLI
commands
invoking
the
face,
and
if
you
skip
the
fifth
one
you
have
to
do
something.
Thank.
B
I,
let
me
cause
have
something
here
which
annoys
things
up:
that
that
try
to
use
cops
but
I,
think
is
actually
helpful.
So
with
cops.
You
have
this
notion
that
you
can,
you
can
put
faces
in
to
validate
mode
but
or
put
tasks
in
to
validate
mode.
So
we
verify
the
output
that
we
don't
do
it
for
you,
and
we
have
a
discussion,
a
discussion
about
like
what
happens
if
the
validation
fails,
but
the
idea
being
like
to
medium
to
find
the
requirements
for
what
is
installed
right.
So
you
have
to
have
a
manifest.
B
You
have
some
keys.
You
have
to
do.
Some
are
back
stuff
right.
All
those
are
done
automatically
when
you
do
creating
in
it.
If
you
are
an
advanced
installer
that
has
special
requirements,
then
you
can
opt
out
of
some
of
those
phases,
but
you
still
have
to
satisfy
the
requirements
of
those
phases.
In
my
opinion-
and
they
says
the
cops
approach
cops
actually
actively
checks
that
you
validated
that
you
like
control
the
contract,
but
it
may
be
the
cube
a
team
decides
not
to
it
may
be
that
some
of
those
contracts
are
too
flexible.
B
The
way
we
the
the
philosophy
encompasses,
if
you
have,
if
TK
said
we're
going
to
not
going
to
use,
you
know,
RSA
crypto,
we're
gonna,
use
some
magic
Google
thing
right,
then
we
would
actually
add
support
for
magic,
google
crypto
as
well
as
RSA
crypto.
We
wouldn't
say
we're
not
going
to
we're
just
going
to
skip
over
the
flag
to
not
validate
them.
So
the
approach
that
cuffs
takes
is
the
requirements
of
requirements.
A
B
The
notion
in
cops
is
the
validate,
doesn't
validate
the
the
model
that
reality
matches
the
model,
and
if
you
want
to
do
something
crazy,
like
google,
crazy,
crypto,
Google
secret
crypto
right,
you
have
to
implement
Google
secret
crypto
in
the
model
and
like
plumb
it
all
the
way
through
through
into
the
spec
end,
and
the
validation
set.
Is
the
configuration
in
reality
equal
to
the
configuration
in
the
model.
So
there's
no
like,
oh,
you
can
cheat,
it's
you
have
to
like
add
support
for
it.
If
you
want
to
support
it,
yeah.
A
B
E
Does
that
mean
abductor
Eric's
point
to
be
considered
a
cube
event
cluster?
That
needs
to
be
like
a
final
page.
That
is
verify
that
you
completed
the
second-to-last
phase
or
that
we
require
people
to
run
the
last
days,
because
that
sort
of
verifies
all
the
inputs
up
to
the
end.
And
then
we
can
say,
assuming
you
pass
that
last
phase,
we
know
that
you're
a
specific
baseline
because
I
think
part
of
what
he's
getting
at
is.
E
If
everybody
says
we
use
cube
admin,
but
everybody
skips
all
the
phases,
but
one
and
it's
like
technically
using
cube
admin
they'll
be
end
up
with
every
cluster
looks
vastly
different
from
every
other
cluster
and
we
haven't
really
accomplished
scible
right.
The
goal
is
to
get
clusters
to
sort
of
a
standard
baseline
where
you
can
build
software
on
top
of
them.
That
assumes
that
baseline,
that
is
compatible
across
installations.
I.
B
Mean
this
is
the
this
is
where
I
say
like
whether
or
not
we
actually
implement
it.
We
should
have
the
notion
of
validate,
so
it
should
pass
validation
on
every
stage
implementing.
It
is
a
different
matter,
but
also
it's
not
lit
yeah.
It's
not
linear,
but
there's
no
last
phase
in
leasing
cops
but
yeah.
A
So
isn't
that
kind
of
a
conformance
test,
or
do
you
mean
we
should
have
something
even
I
mean
that's
what
we're
doing
now
in
the
cube
a
times
always
does
Duke
you
better
minutes
and
join
five
nodes
or
something
and
then
run
the
conformance
test,
and
also
now
the
bootstrap
token
test
and
the
node
authorizer
test
are
coming
soon.
So
I
mean
well
just
like
we're
saying
that
these
cute
tattoos
should
pop
and.
B
Have
a
test
of
you
could
implement
rieflin
kubernetes
in
java
right,
you
can
pass
the
conformance
test
and
there's
a
whole
different
debate,
which
I
don't
want
to
get
into
it
to
whether
or
not
that
is
communities.
But
I'm
saying
that
like
there's
a
coup
by
diem
test
will
be
like
you
need
keys,
and
you
have
a
manifest
which
looks
like
this
and
it
uses
keys
and
it
uses
the
node
authorizer
and
ideally
we'd
say
you
know
like
a
strict
string
match
right,
but
maybe
someone
will
come
with
a
valid
reason.
B
A
D
Think
it
is
an
issue
that
we
that
Cuba
DM
needs
to
think
about
because,
like
we
turn
off
all
of
the
verifications,
whenever
we
use
cout
Vidya,
we
have
to
because
we
run
in
the
container
right.
So
the
if
I
can't
even
trust
rubidium
is
current.
You
know
pre-flight
checks
because
they
make
assumptions
that
are
valid
for
my
environment,
then
I'm
concerned
about
what
validations
would
look
like
further
down
the
line,
because
right
now
my
only
evidence
is
that
I
they
don't
help
me
in
place.
I.
D
B
A
A
Mostly
this
happy
path
where,
where
we
have
you
download
from
App,
get
cubed,
M
or
yum
install,
and
then
we
have
some
assumptions
like
you're
running
system
D,
for
example,
and
checking
if
the,
if
ducky
running
and
these
things
and
just
to
make
giggity,
because
initially
in
one,
for
we
had
a
lot
of
problems
with
users,
not
having
this
valid
configuration
and
then
it's
cubed
M,
just
hang
with,
which
is
this
vessel
insisting.
This
was
the
first
first
item
and
yes,
therefore,
we
made
the
proof
of
checks
and
they're
pretty
much
bundled,
unfortunately,
like.
D
Different
emergencies,
that
replace
don't
happen
or
inherently
that's
fine.
For
me,
it's
more
like
at
the
point
where
who
ADM
is
going
to
define
some
validation,
conformance
I,
think
that's
a
priming
issue,
because
those
assumptions
so
far
have
not
been
kind
of
like
broken
mind,
use
of
it
right.
So,
if
now
my
my
path
to
adopt
cubed
M
is
like
I
have
to
hassle
professor,
then
that
pitiful
yeah.
It
is
interesting
what.
B
D
A
And
it's:
this
is
really
much
more
more
like
do
exiting
than
a
technical
thing,
it's
more
like
warning:
the
user.
If
he
hasn't
been
able
I
mean
systemctl,
stop
docker
and
then
tries
to
run
cube
any
minute
and
that's
why
we
we
have
such
cases
and
and
that's
why
we
added
like
such
a
verification.
I
mean.
A
It's
hard
I
mean
it's
a
fine
line
between
having
coding
into
cubed
in
all
possible
configurations,
we'd
like
to
run
what
is
it's
called
like.
Well,
every
other
system
be
clone
or
or
init
system,
but
we
can
add
I
mean
some
really
useful
like
running
in
a
container
absolutely,
but
it's
really
hard
I
mean
also
if
you
are
using
cRIO
this
breaks.
We
have
an
issue
for
that.
I.
B
H
A
I
mean
Brian
grant
has
had
as
written
some
a
couple
of
these
documents
earlier
and
I
think
it's
one
of
the
nation's
for
the
Gothic
architecture
to
actually
like
check
these
into
kubernetes,
repo
or
whatever,
and
finalize
them
get
consensus
across
the
ecosystem
and
well
then
we
also
have
CNCs
conformance
and
all
these
other
tests
that
validate
that
you're,
something
that
can
have
the
kubernetes
trademark
on
it,
but
TLDR
I
guess
we
don't
know,
really
excited
yeah.
B
I
think
you're
I
think
I
think
certainly
a
great
score,
especially
like
the
absolute
core
functionality.
You
really
cannot
be
proven
as
I
understand
it.
Your
economics
image
without
it
and
there's
also
like
conformance,
which
is
like
a
slightly
broader
like
this-
is
real
kubernetes
but
I
think
I,
think
Rigas
I
think
if
Cuba
name
was
installing
a
lot
of
things,
we're
not
in
core
or
requiring
lesbians
or
not
in
court,
then
we
then
we
have
a
flag
of
like.
B
A
A
It
comes
up
again,
that's
more
like
testing
the
controller
manager,
but
we
probably
need
something
that
tests
more
the
deployment
and
also
this
is
this
turns
into
soon
like
after
API
that,
where,
if
we
specify
something
in
that
after
API
and
want
to
upgrade
node,
then
then
we'll
they
will
get
all
upgrade
or
whatever
I
mean
it.
It's
definitely
a
hard
problem,
and
but
but
we
should
definitely
I
mean
this
group
should
solve
with
like
documenting
these
things.
What
what
do
we
expect
of
our
conformance
like
cluster
on
the
deployment
level,
like
the
setup
level.
B
B
D
Yeah
I
mean
I,
don't
think
I
think
this
isn't.
This
was
really
an
issue
for
asylum
per
se,
but
it's
like
if
we
use.
If
we
were
going
to
make
everything
in
Covidien
pluggable,
then
I
think
it
it
it
quick
helpful
to
defined.
You
know
things
like
ops,
where
you
do
validation
or
that
kind.
This
is
more
a
general
concern
than
me,
representing
my
you
know,
for
less
or
or
whatever
right.
B
Know
you'd
be
like
we
can
if
we
make
atom
separately.
Deanna
manager
is
a
separate
issue,
but
if
we
make
that
a
little
bit,
the
word
doesn't
have
any
output
on
disk
right.
It
can
be
easier
to
replace
or
do
whatever
we
want
to
do
with.
It.
I
think
that
the
the
idea
that
we're
not
if
we
skip
a
phase,
we're
actually
replacing
the
phase
we
are
we
implementing
the
face
is
the
way
to
put
it.
Maybe,
like
you,
don't
skip
a
phase
because
you
don't
want
to
do
it.
B
G
B
But
good
I
don't
know
what
what
the
next
step
should
be
in
terms
of
company
I'd
like
to
see
some
more
I
think
we
don't
have
a
lot
of
blockers
on
this
list
and
I
feel
like
we
could
knock
off
the
blockers
and
we
would
not
yet
necessarily
see
immediate
adoption
by
everyone.
So
I
certainly
will
be
adding
trying
to
dig
dig
more
and
add
some
more
cups
ones,
I,
don't
know
most
can
add
at
least
their
Plus
Ones
or
list
of
blockers,
if
they're
not
yet
on
here.
B
G
A
Yeah,
that's
indeed
the
more
complex
pot
and
just
just
like
plug
for
contributing
cubed
em
I
mean
all
these
suggestions
found
Fung
fine
to
me,
but
still
someone's
got
to
do
them.
So
if
you,
if
you're
wanting
to
like
actually
dig
into
something
like
ding
nails
like
or
whatever
and
I,
can
guide
you
to
some
first,
maybe
like
starter
things
and
then
maybe
some
some
more
heavy
ones,
because
I
mean
that's
indeed
a
lot
a
lot
to
do
here.
So
I'm.
B
A
B
B
B
E
B
B
Well,
that's
it,
then
everyone
has
a
great
rest
in
there.
Tuesday's
I
think
we're
doing
want
to
go
every
two
weeks
from
now
on
is
that
right
that
works
for
people?
As
we
talked
about
that
time,
I
know,
there's
still
some
people
that
are
NAT
with
the
timeslot
I
encourage
everyone
that
is
unhappy
with
the
timeslot
and
is
watching
this
on
YouTube,
presumably
to
to
go
in.