►
From YouTube: sig cluster lifecycle kubeadm office hours
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
A
So
I'll
start
with
the
current
state
of
ducks
what
outstanding,
Docs
issues?
Why
don't
even
share
my
screen
here
for
a
second,
we
have
a
bunch
of
Docs
issues
and
it's
it's
almost
unintelligible
at
this
point
now
of
where
what
things
actually
need
to
get
done
and
if
we
should
just
close
out
what
we
currently
have.
One
second
I'll
show
your
screen
in
just
a
second
here.
A
Hopefully,
folks
can
see
my
screen
yep
right,
so
there's
a
bunch
of
copy
edits
that
are
currently
in
waiting.
I
know
that
we
we
have
do
talk
this
morning.
We
know
that
the
folks
are
working
on
updating
the
h8
documentation
they
work
and
now
I
have
already
have
PRS
for
the
the
standard
documentation
we
should
hopefully
be
merged
by
this
afternoon.
C
So
I
need
to
look
at
this.
What
we
had
for
this
page
but
I
think
it's
pretty
much
done.
I
submitted
a
PR
for
this
and
I
I
think
that
people
from
sick
networking
already
like
submitted
their
changes,
and
you
know,
places
open
yeah
the
reference
PR
it
Smurfs
already
yeah,
so
so
I
think
we
are
pretty
much
done
with
this
one.
Okay,
just
close
it
up.
Oh
I'm,
pretty
sure
it's
done
and
I
mean
it
should
be
up
to
Signet
working
to
save
they
want
something
else.
A
A
A
C
And
I
wonder
if
she,
if
she
it
should
overlap
with
a
great
and
a
cluster
and
I'm,
not
sure,
because
when
you
install
cook
medium,
is
it
only
about
the
installation
from
the
package
managers
yeah?
And
where
should
we
put
this
document
because
I
I
think
it
correctly
does
not
have
the
right
location?
It's
in
independent.
C
A
C
A
B
C
C
A
B
A
C
A
A
A
B
C
Sure
we
can
close
it
and
in
yeah
every
should
be
much
in
there,
there's
one
that
it's
not
much
waiting.
So
the
check
box
is
in
the
first
first
message
of
the
these.
A
D
E
B
B
C
B
I
what
I
wanted
when
I
like
filed
this
issue
is
I,
want
us
to
be
able
to
automatically
from
from
our
code
from
our
Cobra
and
go
duck
stuff
generate
markdown
files.
That
is,
that
is
what
I
want
to
do,
and
then
we
don't
have
to
and
on
every
new
release
we
don't
have
to
ourselves
go
and
copy
over
manual
manually
like
copy
over
the
current
current
reference.
B
C
So
so
I
think
we
should
we
can
do
it,
but
this
has
to
be
discussed
like
the
website
repository
how
how
we're
gonna
integrate
it
with
them,
because
they're
currently
having
some
issues
themselves
like
with
yourself
and
so
I
broke
this
PR
into
a
couple
of
different,
because
this
luck
escalated
from
the
discussion
here.
C
So
this
is
about
how
we
document
the
latest
config
and
the
cube
a
ad
I
mean
it,
but
it
has
to
move
from
there.
I
think
and
the
other
one
is
the
card.
This
is
what
this
is
an
easy
one,
the
second
one
how
we
generate
commands
I'm
gonna,
explain
that
in
contributing
indeed.
So
a
question
now
is:
should
we
like
close
this
one
and
use
eight
nine
to
four
like
how
we
generate
configuration
files,
basically
how
we
integrate
them
automatically
into
the
website?
E
F
A
So
I
think
there
was
a
punt
on
some
of
that
until
the
next
cycle,
like
we're
not
going
to
be
able
to
get
to
the
ideal
state
in
this
time
frame
like
the
the
docs
people
I
have
too
much
on
their
tasks
to
try
and
get
it
done,
and
this
is
too
close
to
the
end
of
cycle.
If
we
want
to
have
an
ideal
world
in
use
case,
we
should
start
planning
for
it
for
1:12
and
just
deal
with
that
in
a
retrospective.
So
I
definitely
won't
have
a
retrospective
for
our
cig
for
this
cycle.
B
B
B
A
G
A
A
E
A
Going
back
to
the
issues
list,
both
Liz
and
I
are
currently
working
on
trying
to
distill
down
the
data
and
get
like
at
least
an
initial
cut
of
something
in
place
that
doesn't
that
actually
works
for
both
outlining,
at
least
for
AWS,
outlining
the
entry
cloud
provider
details
and
just
links
out
to
the
appropriate
section.
So
people
can
pass
in
the
right
arms
because,
right
now,
it's
totally
unintelligible
from
the
external
observer
and
I
do
think
that
this
is
blocking
or
blocking
worthy.
C
A
B
So
so,
from
from,
at
least
for
my
discussion,
we
said
with
Andrew,
like
the
cloud
provider
is
gonna
like
do
that
stuff
with
with
the
generate
Docs,
and
this
is
like
how
the
AWS
cloud
provider
works,
etcetera,
etcetera
here
is
how
you
run
the
out
of
three
cloud
provider.
Hereis
I
run
the
entry
on
a
general
basis
and
what
we
want
to
improve
there,
for
example,
is
to
put
them
under
the
same
folder
and
make
it
more
discoverable,
because
now
they're,
scattered
along
around
in
Doc's
repo,
you
know
places.
C
B
R-Cubed
inbox,
which
I
want
to
fix
so
so
they're
gonna,
do
that
eventually,
probably
not
for
111's
it's
kind
of
late,
but
then
we
now
we
from
our
cube
ATM
and
cloud
provider
like
cross-cutting
doc.
We
should
duck
or
we
should
link
to
all
the
places,
but
we
hope
to
have
those
places
more
in
a
uniform
directory
or
something
like
this
is
the
place
you
go
to
every
time
you
want
to
deal
with
a
craft
order,
so.
B
A
I'm
not
gonna,
go
that
far
I'm
gonna
basically
say
like
go
to
these
other
locations
to
understand.
The
arguments
that
you
need
to
pass
in
in
the
derivation
is
left
of
the
reader,
because
I
say
cloud
providers
fractal.
If
we
start
doing
it
for
one,
then
we're
gonna
have
to
do
it
for
all
the
other
stuff.
No.
B
So
without
being
like
without
saying
explicitly
for
one
but
instead
like
this,
is
how
you
generally
configure
your
cloud
provider
you
need
to.
If
you
want
to
use
the
entry
cloud
provider,
you
need
to
pass
the
extracts
for
the
API
server
controller
and
cubed.
If
you
want
to
do
the
outer
tree,
go
ahead
and
do
cubelet
extracts
cloud
provider
external
to
everyone
keeps
available.
A
A
H
So
the
two
things
that
we
need
to
mention
here
on
the
cuban
vm
side
is
how
did
do
the
cloud
provider
and
the
cloud
config?
That's
it
and
then
everything
else
is
punted
over
to
the
other
side
and
yeah.
I
can
I
kick
the
tires
yesterday.
After
for
the
OpenStack
one
at
least
so
I
know
what
we
need
to
do
there.
So
I
can
help
the
tattoo.
I
A
A
C
B
C
Or
so
my
question
here
is:
is
this
overlapping
with
the
I'm
a
good
kind
of
confused,
because
we
we
have
so
we
have
a
plan
for
the
transition
from
version
1
alpha
alpha,
1,
2
or
2,
and
do
we
have
to
like
combine
these
with
the
version
file
because
I
don't
adore?
We
have
anything
before
it
yet.
Yesterday,
I,
don't
really
understand
what
to
write
in
there.
That's
because
I
don't
know
the
like
the
destruction,
yeah.
F
A
C
E
B
First,
one
that
I
have
there
that
is
packaging,
cubed,
M
and
running
an
unseen
system.
These
systems,
it's
basically
gonna,
be
the
same
dock
as
configuring,
the
cubelets
in
your
cluster
on
using
another
z
area
right.
It's
like
it
all
ties
into
each
other.
So
it's
like
the
same
dock,
but
securing
your
Cuban
Plaza
photo
is
gonna,
be
kicked
out
to
112.
If
you
like,
could
edit
the
issue
or
something
we
could.
B
B
A
A
E
B
A
B
B
C
A
B
A
A
I
wanted
to
give
a
shout
out
quick
to
Jason
for
fixing
their
period
test.
That
was
great.
It
looks
like
things
are
greenish
for
the
most
part,
there's
some
other
issues,
but
I
don't
think
they're
related.
It
could
be
any
him
at
all
and
that's
the
good
thing.
The
bad
thing
is
I
walk
through
the
upgrade
test,
and,
oh,
my
god,
the
horror,
the
kubernetes
anywhere,
how
it
does
the
upgrade
test.
That
is
Wow.
B
B
So
I
on
Jason's
PR
check
questions.
Q
Falls
ice
under
from
the
release
team
asked
why
the
tree
tensile,
SAP,
stateful,
set
upgrade
HPA
upgrade
and
service
upgrade
are
failing,
and
this
is
a
question.
We've
got
every
single
release
assets.
None
of
this
and
someday.
We
should
fix
it.
If
you
open
this
s
grid,
we
can
like
look
at
its
illustrate.
H
H
B
If
you
do
yeah
yeah,
but
if
you
do
upgrade
110
111
that
I'm
looking
at
pretty
so
this
is
the
current
state
under
steam
said
like
there
are
wrongly
marked
the
overalls
cutest
apprentice.
So,
but
these
three
states
will
set
HPA
a
service
upgrade
are
expected
to
fail.
They
fail
for
every
single
qubit
an
upgrade
job,
because
stateful
said
upgrade
requires
persistence
for
automatic
persistent
fully
in
creation.
B
B
Problem
is
that
it
doesn't
work,
there's
something
there's
something
really
weird
like
if
you,
if
you
select
the
master,
upgrade,
it's
gonna,
execute
those
unconditionally
and
not
respect
the
flags
or
stuff.
So
I,
remember
being
like
grumpy
about
that,
like
you
half
a
year
ago,
but
but
yeah
it
due
to
how
the
current
code
is
done,
which
will
god
stop
great
tests,
we
can't
skip
this
and
HPA
upgrade
requires
hipster
and
service,
upgrade
request
service
type
load,
balancer,
so
yeah
I.
H
That
is
like
that
is
really
good.
So
I
should
just
show
you
what
I
was
talking
about.
Can
you
click
on
the
first
green
box
on
the
top
left
corner,
yeah,
green
yeah,
and
so
here
you
can
see
the
test.
Three
failed
and
202
succeeded
right.
So
that's
what
we
need
to
look
for.
You
know
weather
to
actually
see
whether
the
job
is
basel
thing.
Yeah.
A
One
now
that
the
tests
are
green,
but
we
don't
have
testing
coverage
on
a
lot
of
the
new
additions
that
we
have
other
than
sort
of
the
the
unit
test.
Validation.
We
don't
have
like
you
know,
sort
of
partial
integration
or
end-to-end
test
for
a
bunch
of
the
new
things.
Are
there?
Are
there
a
bunch
of
new
features
and
see
lies?
Do
are
there
things
that
folks
would
like
anyone
to
focus
on
while
we're
out
here
in
this
meeting?
They
do
that
last
mile
push
to
get
eyes
on
particular
focus
areas,
so.
B
What
would
be
super
powerful
is
install
cubed
and
110
using
everything
110
like
do
that,
get
install
install
normal
cubed
in
clusters.
You
would
do
when
you
have
all
data
installed.
You
go
ahead
and
fetch
the
new
cubed
M.
You
basically
follows
off
grid.
Dr.
fetch
attack
the
cubed
and
binary
from
head.
B
B
A
C
C
B
B
E
B
K
A
B
B
L
L
B
K
B
I
B
B
This
might
might
be
sometimes
a
little
bit
of
problems
with
like
converting
the
G
Ducks
markdown,
but
it
does
work
kind
of
well
and
it
has
been
like
at
least
manageable
to
edit.
The
release
notes
from
like
the
maintain
is
perspective
in
Google
Docs
versus
checking
everything
into
github,
because
the
the
dark
I
mean
it
is
like
several
several
ten
pages
long
and.
C
How
about
using
Google
Docs,
but
instead
of
using
the
formatting
of
the
fonts
and
stuff
inside
the
Google,
Doc
editor?
How
about
we
like
in
wine,
the
markdown
inside
the
Google
Docs
document
like
when
a
copy?
Please
yeah
you
when
copy
pasta
to
get
help,
is
gonna,
be
converted
immediately
to
the
correct
model.
Yeah.
B
B
H
B
Cuz
case
it's
like
we
want
to
do
that.
Do
that
as
soon
as
possible.
Do
this
I
spent
like
countless
several
hours
just
moving
stuff
around
and
there's
a
lot
of
manual
editing
needed
to
make
make
the
release
notes
sane,
basically
from
the
auto-generated
material?
There's
there's
a
lot
of
good
stuff
in
there
by
default
from
from
the
auto-generated
stuff,
but
a
lot
of
polishing
is
needed
and
we
have
a
lot
of
features
for
111,
so
I
really
want
to
see
us
get
us
to
get
the
release.
Notes
right.
B
So,
when
you're
reading
the
simplest
life
cycle
chapter
along
with
the
major
teams,
there's
like
first
the
major
teams
come
then
all
the
action
required
for
cyclist
life
cycle
stuff
and
then
later
down.
It's
like
here
are
the
major
changes
between
110
and
111
nerdling
to
pyaare
some
general
issues
and
like
everything
so
so
yeah.
That
is.
That
is
one
thing
we
we
should
collaborate
on
as
a
secret.
Well,
just
just
like
a
heads
up
there
as
soon
as
we
get
as
soon
as
that
is
possible.
That
is
okay,.