►
From YouTube: 20180718 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
Hello
today
is
Wednesday
July,
18
2018,
and
this
is
the
ku
Beta
Gamma
Sigma
stroller
cycle
convenient
office
hours.
Currently,
our
agenda
is
pretty
light.
So,
if
you
have
things
to
discuss,
please
add
them
to
the
agenda
for
b2.
It
has
one
topic
that
I
know
that
Lucas
asked
us
to
talk
about
and
it's
basically
the
promotion
of
the
configuration
to
be
1.
Beta
1
I
have
initially
looked
through
it
a
long
time
ago.
A
C
Lucas
is
pushing
a
lot
on
this
worker
and
is
every
doing
a
good
job,
and
we
are,
we
already
had
in
place
at
least
once
one
big
step
of
the
of
the
API
change,
which
is
to
have
a
different
or
a
lazy
dependency
with
that
component,
configure
and
now
component
config
are
not
any
more
part
of
the
public
FBI
of
cupid
mean,
but
they
are.
They
should
be
passed.
C
The
spirit
document-
and
this
is
already
merged
at
as
as
PR
now
Luke
is
working
in
the
second
big
step,
which
is
to
split
the
cluster
configuration
from
the
inter
configuration.
So
we
will
have
a
cleaner
cupid
mean
config
map
that
doesn't
have
tokens
or
everything.
It
is
a
temporary
or
being
today
today
join
action,
and
then
there
is
the
the
third
part
of
the
weft
API
change.
C
Some
action
are
already
well
defined
at
the,
for
instance,
having
the
sub
structure
for
each
control
playing
component
found.
Some
other
detail
should
be
better
divided,
for
instance,
ow
or,
if
and
how
we
are,
we
changing
how
we
manage
images
now
there
is
a
an
approach
which
is
kind
of
articulated.
Where
were
we,
we
have
to
set
up
the
image,
repository
or
unified
contour
painting
page
and
this
for
the
contraband
component
Y
for
the
etcd.
C
You
can
find
a
rabbit,
a
brief
discussion
on
all
the
topics
and
in
the
final
part
of
the
cap,
there
is
the
last
proper
version
of
the
new
api
that
lucas
and
that
that
lucas
and
me
in
some,
and
also
some
other
folks
in
in
in
in
some
session
of
pepe
program
and
prepared
that,
but,
from
my
point
of
view,
it
is
necessary.
Some
are
still
RFI
elements,
so
any
contribution
will
be
more
than
el.
D
C
E
A
A
C
C
A
C
Yes,
the
critical
point
is
that
the
PR
Lucas
is
working
on
is
kind
of
blocking
because
he's
splitting
the
unit
configuration
from
the
cluster
configuration
and
until
we
have
this
in
place,
it
is
also
difficult
to
work
in
parallel
and
and
and
to
move
to
move
on
and
to
move
on
quickly.
So
I'm
helping
look
after
Lucas
from
remote,
but
as
you
know,
it
is
not
always
a.
A
C
A
C
A
Are
there
other
topics
to
discuss
today?
I
know
that
Lee
had
mentioned
he
wanted
to
join,
but
I
thought.
Yesterday's
conversations
regards
to
testing
was
pretty
closed
for
the
time
being,
I
did
follow
up,
I
am
gonna,
follow
up
with
sig
testings,
especially
what
Ben
is
trying
to
do
with
regards
to
didn't
testing,
because
I
think
that
would
be
highly
beneficial
to
everyone.
If
we
can
get
like
a
PR
blocking
den
tester
for
Covidien,
that
would
be
super-dee-duper.
Do.
A
Because
that
way
we
get
PR
blocking
jobs
on
kopitiam
deploys
of
some
sort
right
so
that
that
alesis
gets
us
first
pass
signal.
I
was
going
to
sign
up
the
sig
mailing
list
today
for
the
spam
bot
failure.
So
if
it
becomes
too
much
noise,
please
let
me
know
and
I'll
try
to
turn
down
the
gain
on
that
one,
because
there
are
parameters
to
tweak
that,
but
that's
the
only
other
updates
I
can
think
of
at
this
point
live
with
me
or
other
folks.
Do
you
have
other
updates
join
us
guys,
I
wanted.
B
A
A
Know
if
there
is
a
hard
date
deadline,
so
I
think
if
we
get
something
up
their
initial
PR
within
the
next
week
or
so
a
steering
committee
is
backlogged.
The
problem
is
from
the
steering
committee
currently
is
that
we
have
a
lot
of
the
people
who
started
the
project,
who
also
have
a
million
other
responsibilities
so
getting
getting
the
time
to
X
review.
All
the
charters
is
a
problem
yeah.
B
G
B
A
Would
take
that
I
would
take
that
to
sig
release
Channel
and
they
want
to
I
know
that
Tim
pepper
is
trying
to
make
changes
to
allow
this
to
be
better,
so
definitely
file
that
with
them,
because
they're
pretty
much
the
owners
of
the
relationship
auditory,
but
only
a
Googler
can
push
to
the
GCS
buckets.
So
it's
still
in
their
boat.
We
can't
as
I
say
we
don't
control
that
apparatus.
A
A
The
major
releases
are
usually
cut
with
a
lot
of
more
buy-in
from
the
SIG's
to
make
sure
that
everything
is
done.
The
minor
releases
are
kind
of
done,
asynchronously
from
the
SIG's,
so
you
know
it's
not.
There
isn't
person,
that's
double-checking
everything
before
it
all
goes
out
the
door,
so
I'm
less
concerned
about
the
major
releases
and
as
I
am
but
the
minor
release
updates
so
I
think
talking
with
Fox
ish
I'm
sig
release
them
as
well
as
Tim
peppers.
It
worked
yeah.
B
F
H
A
There
is
a
possibility:
well,
there
is
a
possibility
for
a
breakage
if
there's
a
cuvette,
DM
bug
and
that
Covidien
bug
isn't
fixed,
but
part
of
the
upgrade
instructions
are
step.
One
update
update
your
packages
step.
Two
then
pull
the
images,
so
I
mean
like
it's
documented,
to
do
the
right
thing.
Whether
or
not
users
do
the
right
thing.
Is
you
know
another
story
so
in
that.
B
B
J
Yeah
so
I
just
wanted
to
in
this
whole
conversation
and
the
conversation
from
yesterday
I
had
another
thought
which
is
do
we
want
to,
since
we
know
that
we
want
to
change
the
CI
jobs
and
things
like
that.
Add
a
DND
job,
and
things
like
that.
So
is
this
the
right
time
to
actually
consider
moving
our
code
into
a
separate
repository
as
well.
This
Jason
and
I
were
talking
on
another
one,
and
we
were
talking
about
something
else.
J
A
The
problem
is
that
slightly
orthogonal
to
the
actual
problem
statement.
Even
if
we
were
to
a
secular
elites,
we
depend
upon
packages
such
as
couplet
right
so
like
we
need
the
Kula
package
along
with
ku
medium.
They
need
to
be
in
lockstep
synchronization,
so
I
don't
see
it
as
being
beneficial
at
this
stage
and
again,
there's
there's
a
bunch.
We
can't
really
a
synchronously
release,
cube
a
DM
just
because
of
how
the
vetting
process
works.
J
A
Won't
see
it
as
a
priority
right
now.
It's
that
it's
kind
of
inter
it's
conflating
the
problem
set
like
I
I,
don't
see
it
as
being
a
problem
per
se
right
now.
The
biggest
problem
is
that
the
build
artifacts
are
in
one
repo
and
the
actual
testing
for
most
of
the
other
stuff
is
in
another
repo.
If
we
actually
just
built
and
tested
for
one
single
repository
and
we
use
Basile
for
everything
that
would
solve
ninety
percent
of
these
problems
and
I,
would
you
know
I've
been
wanting
that
for
what's
the
date?
J
Would
be
a
clear
signal
to
everybody
that
okay,
we
are
starting
from
scratch.
You
know
this
is
the
new
repository.
This
is
how
we
are
going
to
do
stuff,
and
you
know
that.
That's
basically
why
the
thought
came
about
another
example.
What
Jason
and
I
were
talking
from
yesterday
was
in
the
class
in
the
cluster
API
stuff.
We
needed
things
that
are
incubating.
J
For
example,
we
were
talking
about
the
configuration
validating
the
configurations
before
we
pass
things
on
to
QB
diem,
for
example
right
and
then
we
were
talking
about.
Are
there
multi
going
to
be
multiple
versions
of
these
things,
and
where
do
we
render
it
from
so
that's
how
we
got
into
that
rabbit,
hole.
A
I
Not
really
but
I
think
that
it
well
in
my
mind,
I
think:
that's
we.
We
don't
even
want
to
try
to
address
breaking
it
out
a
repo
at
this
point,
especially
with
the
amount
of
work
that
we
have
ongoing
with
the
current
cycle,
with
trying
to
stabilize
the
config
I
think,
once
things
settle
down
a
bit,
then
we'll
have
a
bit
more
kind
of
cycles
to
be
able
to
try
to
take
on
something
like
that.
I
agree,
10,000
percent,
so.
J
A
A
Another
conundrum
that
bothers
me
that
it
relates
to
test
automation
which
I
need
to
talk
with
sig
note
about
too,
as
well,
is
that
there
is
no
CRI
test
automation
for
these
other
things,
and
we
are
like
literally
the
focal
point
of
the
universe
like
a
laser
beam
for
all
the
CRI
deployment
configurations.
So,
like
they've
kind
of
said,
yeah
cierra,
it's
great,
you
know,
use
container
D
in
cryo,
but
no
one
actually
did
PRS
on
our
side
to
validate
and
set
up
test
automation.
A
J
B
H
B
H
K
Twitch
its
for
Brazil
on
yes,
he
is
well
I,
just
probably
gonna
send
in
a
quick
patch
to
the
Canadian
development
repo
to
update
the
requirements
list
for
the
playground,
because,
like
older
versions
of
ansible,
are
basically
incompatible.