►
From YouTube: 20190828 - Cluster API 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
and
welcome
to
the
Wednesday
August
20th
edition
of
the
cluster
epi
office
hours,
a
sub-project
of
C
cluster
lifecycle.
If
you
have
any
items
you
want
to
discuss,
please
go
ahead
and
add
them
to
to
the
agenda
I'm,
linking
that
in
the
chat
now
to
start
off
with
I
wanted
to
give
everybody
another
update
on
the
face
to
face.
If
you
are
planning
on
attending,
please
do
RSVP
with
the
link
in
the
agenda
doc.
If
you
do,
we
may
not
be
able
to
fit
you
in
the
room
that
day
we
are
limited
on
seating.
A
A
B
A
Okay,
all
right,
it
looks
like
Moshe
is
finishing
up
work
on
pull
request,
number
153,
I,
think
that's
related
to
having
each
of
the
CA
certificates
in
an
individual
secret
versus
a
single
secret.
Chuck
is
working
on
the
request
number
177,
which
is
C,
calculating
the
cert
hashes
and
verifying.
So
we
can
verify
the
CA
certificates
instead
of
you
know,
skipping
validation.
A
A
C
And
of
this,
we
fixed
an
issue
where
the
exponential
back-off
the
last
ten
billion
years,
if
it
will
continue
to
return
there,
and
this
actually
needs
to
be
backward,
it
release
zero.
Three
I'll
try
the
backward
the
change
today
before
we
release
this
week
or
next
week
and
on
master
beso
on
the
bender
fall
that
have
been
removed
and
for
cap
as
well.
We're
targeting
release
this
Friday
for
b1.
A
D
I
consistently
saw
we
merged
just
be
one
of
two
controllers
and
you're.
Also
using
the
human
input
store
right
now.
Those
two
are
merged
and
your
example.
Examples
mostly
work,
but
I
didn't
get
to
it
yesterday,
after
classical
most
much
until
it
could
be
working.
I
think
they're
awesome
not
really
published
rush,
published
images
of
the
classic
controller
and
everything
else
right
now,
yet
so
think,
I
wait
for
those
and
I'm
like.
A
A
E
A
A
All
right,
thank
you
for
the
cluster
API
upgrade
tool.
We
haven't
made
any
progress
there.
Yet
we've
been
mostly
focused
on
the
v1
alpha
to
work
for
core
cluster
API,
the
bootstrap
provider
on
the
ws
provider.
Once
we
get
the
release
out,
we
should
be
able
to
start
better
and
numerating
what
the
work
is
to
do
to
get
that
updated
next
week.
Hopefully
and
Vince,
can
you
give
an
update
on
cluster
cuddle,
yeah
I'm,
the.
C
Vr
I
support
for
like
the
pivot
face
and
the
create
workflow
has
been
merged
under
the
lead
worker.
Who
has
me
in
test
yet
so
expect,
like
other
peers
to
come
after
we
release,
but
the
pivot
has
been
dust
from
the
new
relic
folks,
we
did
with
it
some
like
tests
on
zoom
like
last
week,
so
that
has
been
working.
Fine,
so
close.
Ricardo
is
also
like
agnostic
right
now,
so
you
can
have
one
build
and
point
it
to
any
provider,
see
release
and
it
will
just
just
work.
A
Okay,
let's
move
on
to
the
the
core
cluster
API
PSAs.
We
are
targeting
to
kind
of
release
this
Friday,
so,
barring
any
major
issues,
we
should
be
able
to
get
the
cluster
cuddle
binary
released
and
get
the
initial
images
up
for
that
dazzling
vendor
have
been
removed,
so
hopefully
nobody
has
any
major
issues
related
to
that.
If
you
do,
please
raise
them.
A
So
hopefully
this
will
help
out
with
iterative
development
after
the
release,
so
that
you
know
people
can
start
consuming
the
latest
tag
based
on
from
the
staging
bucket
directly,
if
you're
trying
to
test
newer
changes-
and
it
looks
like
somebody
added
in
their
congratulations
to
VMware
on.
Let
me
see
if
I
can
pronounce
this
right,
Tong's
ooh
I
was
me
Congrats
guys
thanks
Jessa.
C
Yeah
I
added
it
but
I
saw
Moshe
was
online
and
I
got
all
I
wanted
to
like
try
to
see.
If,
like
we
can
like
get
the
synchronous
discussion
that
have
been
a
back-and-forth
on
the
PR
straighten
out
during
this
meeting,
so
that
we
can
move
it
along
I
get
all
my
screen
or
somebody
else
can
share
their
screen.
C
So
there
has
been
some
con
for
like
for
this,
like
who
don't
know
like
what
this
world
does
it.
It's
just
like
sets
like
a
document
for
kind
of
like
having
like
standards
like
around,
like
how,
like
we
manage
certificates
of
where
to
expect
the
certificates
to
be
found.
Most
has
been
like
mostly
like
pushing
this
forward,
and
thank
you
very
much
like
for
doing
this
in
so
little
time.
C
A
Most
of
that
was
just
some
myths
around
the
specific
verbage
for
making
sure
that
we're
clear
on
that.
It's
not
anything
that
I
would
consider
blocking
the
PR,
maybe
something
that
we
can
easily
fix
up
and
post
I
did
call
out.
The
the
si
key
pair
is
different
from
the
other
one.
So
we
may
want
to
call
that
out
separately,
I'm,
not
familiar
enough
with
short
manager
myself
to
know
how
that
would
be
handled
differently
if
creating
those
would
start
manager
for
just
a
normal
public/private
key
pair
versus
an
actual
self
signed
certificate
authority.
A
F
I
don't
know
if
certain
manager
can
actually
create
those,
but
I
also
don't
think
it
would
be
necessary
for
so
that
manager
to
create
those,
because
there
wouldn't
really
be
it's
not
training,
there's
no
value
in
creating
those
externally,
so
yeah,
but
I
think
maybe
we'll
just
call
that
out
and
make
it
explicit
in
the
documentation
so
that
we
can
definitely
do
I.
Think
the
the
label
requirement
is
probably
something
we
should
put
in
bed,
though
one
way
or
the
other.
F
C
F
F
F
C
F
F
F
So
if
you
go
I'm
looking
at
this
from
like
security
auditing
perspective
and
if
I
were
auditing
views
on
the
secrets,
I
wouldn't
want
cluster
aprt
view
all
the
secrets
all
of
the
time.
What
would
want
to
be
a
little
bit
more
restrictive
as
to
what
it
views
and
maybe
that's
possible
to
do
it
not
actually
to
get
the
list
of
names
and
not
the
values
but
I
think
that
potentially
over
complicate
something
that
was
quite
simple
with
just
the
label.
F
F
A
It
would
be
potentially
having
IIIi,
don't
necessarily
know
that
we
would
be
worrying
all
of
them
yeah
it's
I
mean
this
is
what
the
realistic
look
like
whirring,
the
the
only
thing
that
I
worry
about
is
is
whether
or
not
because
we
don't
require
labels
on
a
lot
of
the
other
components.
If
we're
changing
up
user
expectations
when
they're,
if
they're
creating
these
themselves.
C
C
Can
see
it's
one
way
or
the
other
right
like
you
would
need
the
cluster
name
to
have
a
cluster
name
in
here.
I
guess
so.
You
need
a
cluster
to
use
two
certificates
anyway,
but
the
other
way
around,
like
I,
also
see
like
that
label.
It's
like
not
required
elsewhere
and
there's
like
an
opening
shear
to
fix
that.
Maybe
in
the
future,
your
iteration,
but
like
I,
don't
discover
for
this
discussion.
A
A
A
A
A
All
right
so
I
don't
think
we
scoped
whether
we're
gonna
try
to
get
that
in
place
as
part
of
a
follow
on
for
the
0
to
X
release
or
if
we
wanted
to
do
it
too.
Next
I'll
just
move
it
to
0
to
X
all
right
make
sure
documentation
is
clear
that
machine
deployment
is
the
right
model
for
disruption,
style,
machine
updates.
A
A
A
A
A
I'll
bump
it
to
zero
to
X
for
now
sounds
great,
don't
abort,
cluster
creation
when
applying
the
add-on
field.
C
A
A
C
C
A
C
A
A
A
A
Okay,
all
right
in
that
case,
I'm
gonna
go
ahead
and
go
ahead
and
bump
this
one
to
next.
All
right,
let's
see
reconcile
should
set
CA
cert,
hashes
and
settlements.
They
skip
CA
verification
and
there's
pure
and
light
for
this
already
use
controller
runtime
everywhere
for
consistency.
I
think
at
this
point,
that's
a
nice
to
have.
A
Alright
and
add:
spec
Doc's,
not
provider,
to
cube
ATM
config
I.