►
From YouTube: 20200812 - 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
Okay,
hi.
Everyone
today
is
august
12th,
and
this
is
the
cluster
api
office
hours
cluster
api
is
this
pr
subproject
deficit
cluster
life
cycle?
Please
adhere
to
the
cncf
code
of
conduct
and
raise
or
use
the
race
hand
feature
on
zoom
if
you'd
like
to
speak.
A
A
So
I
don't
think
we
have
any
psas
today
the
for
those
who
weren't
here
last
week,
the
zero
three
eight
release
of
cluster
api
was
out
and
your
sorry
039
is
planned
for
end
of
august.
I
believe,
okay,
so
we'll
just
get
started
with
the
discussion
topics.
Unless
anyone
has
any
other
psas,
they
want
to
add
that
weren't
on
the
notes.
B
B
A
Thanks,
do
you
have
a
link
to
the
proposal
for
people
who
don't
have
it
already
if
she
could
add
it
to
the
doc
that'd
be
great.
B
A
Thanks
and
is
there
a
deadline
on
it
or
lazy
consensus
date
already,
or
is
it
just
open
for
now,
just
at
the
review
stage.
A
Okay,
thank
you
any
questions
for
near
about
the
proposal
for
external.
C
B
Why
do
you?
Why
do
you
think
it's
complex.
C
One
is
about
the
quorum
of
the
cluster,
that
is
the
nodes.
C
And
then
the
other
assuming
that
is
resolved,
then
the
question
is:
how
do
we
remediate
who
who
is
responsible
for
it?
So
how
do
we
really
externalize
it
in
the
sense?
Is
it
coming
from
the
bare
metal
side
of
it,
or
is
it
coming
from
the
ionic
side
of
it
or
whatever
the
mass
side
of
it,
and
how
do
we
ensure
that
the
node.
C
Is
which
is
not
now
in
the
food,
because
either
you
throw
it
out
of
the
pool
and
you're
not
then?
How
do
we
know
that
this
is
correct?
So
basically,
I
could
not
really
understand
that.
Will
we
be
able
to
successfully
do
it
without
understanding
the
implications
of
it,
because
we
have
to
deal
with
the
raft
we
have
to
deal
with.
What
is
the
concept
of
the
whether
it
is
in
the
pool
it
is
out
of
the
pool
if
it
is
out
of
the
pool?
C
Does
it
break
the
network
segments
and
how
do
we
really
deal
with
something
except
to
say
that,
okay,
we
can
probably
give
a
status
saying
that
hey
this
is
not
in
the
this
should
not
be
in
the
list.
This
needs
some
fixing,
but
is
it
going
to?
Are
we
going
to
fix
it
within
a
given
time
frame
or
somebody
else
is
going
to
fix
it,
and
is
it
just
a
alert
is
enough
or
if
you
try
to
remediate
by
alert
manually
that
is
different
or
what
is
doable
within
the
particular
watch.
C
If
we
watch
a
particular
node,
so
my
my
thinking
has
been
I'm
struggling
with
the
idea
and
I
really
after
going
through
the
document
which
is
there,
I
still
felt
that
I
wrote
something
and
then
withdrew
back.
So
my
thinking
is,
we
need
to
have
more
thinking
before
jumping
into
it
and
making
that
this
happens.
B
Yeah,
so,
first
of
all,
thanks
for
the
feedback,
I
think
we
might
be
mixing
here,
two
different
type
of
remediations,
so
all
the
other
quorum
stuff
you
mentioned
is
related.
I
guess
to
the
kcp
remediation,
which
is
something
else
and
in
any
case
the
user
can
configure
the
remediation
type
it
wants
to.
If,
if
you
don't
want
any
type
of
remediation,
you
can
just
not
use
it.
It's
an
opting
feature.
So
the
user
can
decide
whatever
it
wants
to
do
if
it
wants.
B
If
he
wants
to
remediate
bare
metal
hosts
for
a
masters
for
control
panel,
so
you
can
do
it.
If
you
don't
want
you
just
want
an
alert,
maybe
you
can
do
it
as
well.
So
this
is
an
opt-in
feature.
So
that's
that's
my
point
of
view
at.
A
Least,
okay,
thank
you,
prakash.
Does
that
help
a
little
bit
answer
your
question.
C
If
I
am
talking
on
the
wrong
topic,
maybe,
but
I'm
still
looking
at
it,
what
the
cluster
api
includes
the
health
check,
and
so
as
long
as
it's
only
limited
to
health
check
and
somebody
else
doing
the
remediation,
that's
different,
but
otherwise
specific
remediation
requirements
such
as
kcp.
If
you
are
excluding,
then
it's
fine
with
me,
then
I
think
it
should
be.
A
Maybe
it's
worth
clarifying
in
the
non
goals
of
the
proposal,
what's
in
scope
and
what's
out
of
scope,
if
that
wasn't
clear,
yeah
sure,
okay,
great
thanks
any
other
questions
about
external.
A
I
don't
see
any
hands
raised,
so
I'm
just
gonna
keep
going.
Okay,
mal!
You
have
a
question
about
kcp
upgrade
behavior.
D
Yeah
thanks
so
the
I
would
like
to
get
a
quick
confirmation
from
people
here
who
are
much
more
knowledgeable
about
this.
The
when,
whenever
we're
trying
to
do
an
upgrade
of
the
kcp
and
let's
say,
there's
a
misconfiguration
from
the
users
or
like
anything
that
prevents
the
upgrade
to
go
through,
and
the
user
like
changes.
D
The
kcp
back
to
something
that
is
supposed
to
be
working
is
like
is
this
kind
of
a
supported
scenario
where
the
the
node
would
actually
be
deployed
with
a
new
configuration
or
not,
because
basically,
what
we've
seen
in
our
scenario?
So
we
might,
we
might
have
a
problem
somewhere,
but
whatever
is
that
the
the
cubed
m
config
that
is
generated
from
the
for
the
first
upgrade
is
actually
not
modify
afterwards.
F
Yeah,
if
you
want,
you
can
try
to
answer
a
little
question.
So
as
far
as
as
I
know,
kcp-
and
this
better
behavior
for
me-
is
that
you
could
you
changed
sap
you,
you
schedule
an
upgrade
kcp
starting
or
only
logging
out
the
new
nodes
and
in
if,
in
the
middle
of
the
the
operation
you
basically
define
a
new
desired
state.
Tcp
will
try
to
align
to
the
new
desired
state.
So
I
do
expect
that
what
we
are
talking
should
work.
D
Okay,
great
thanks
just
wanted
to
confirm
that
that's
how
it's
supposed
to
work
cool.
A
Thanks
for
watching
cool
all
right,
so
warren
metadata,
emol.
G
G
However,
we
so
the
case
I'm
making
is,
can
we
or
the
case
I'm
advocating
for
is:
can
we
have
metadata
yaml
as
part
of
the
provider
releases
like
provider
release
artifacts?
That
way
certain
issues
like
that
specific
one
like
three
four
one,
eight?
Would
we
wouldn't
run
into
this
more
often
because
right
now,
what's
happening?
Is
people
have
to
constantly
up
update
cluster
cuddle
cli
to
you
know,
making
sure
that
like
if
something
goes
wrong,
you
know
the
default
would
be.
G
A
Thanks,
I
can
speak
from
azure
perspective
from
the
azure
provider
perspective.
I
don't
think
there's
any
reason
why
we're
not
shipping
it
it
just
I.
I
wasn't
aware
that
that
was
something
we'd
want
to
do.
So
I
don't
see
any
blogger,
but
does
anyone
else
have
any
thing
that
comes
to
mind?
Why
we
would
not
want
to
do
that.
A
G
Yeah,
that
makes
sense,
I
believe
it
is.
I
linked
the
in
the
provider
contract
for
cluster
cuddle.
There
is
a
it's
it's
there
in
the
cluster
cuddle
provider
contract,
but
if
there's
any
other
places
that
would
like
to
see,
then
just
let
me
know-
and
I
can
remedy
that.
A
I
think
it'd
be
great
if
we
could
open
issues
in
like
providers
to
have
like
chuck
making
that
change,
adding
it
to
the
released
artifacts
that'd
be
a
good
step.
Awesome.
Yeah
thanks.
I
feel
good
too.
F
I
only
would
like
only
to
add
as
a
small
clarification,
so
basically,
the
the
metadata
yammer
is
something
that
links
are
released
to
a
version
of
the
cluster
api
contract.
For
instance,
it
tells
that
cup
v
0
5
series,
that
is,
series
is
linked
to
v
1,
alpha
3,
okay
and
up
up
to
now
everything
worked
because
most
of
the
provider
basically
are
keeping
the
the
same
releases
seriously
so
zero
to
find
this
case
without
changing.
F
But
the
problem
is
when
you
basically
change
already
series.
So
when
you
create
a
new
0.6
and
encase,
a
cattle
cannot
link
these
release
to
a
cluster
api
contract.
F
So
maybe
we
can
embed
in
custer
cattle,
some
sort
of
defaulting
rules,
but
I
see
risk
on
this,
so
I
will
take
charge
of
opening
an
issue
in
copy
to
discuss
how
a
possible
solution
will
be
to
the
problem
and
how
we
can
make
it
the
life
of
the
provider
easier,
and
if
there
is
no
good
solution,
maybe
it
is
better
to
be
explicit
and
ask
the
provider
to
explicitly
declare
for
each
version
which
cluster
api.
There
is
a
contract
that
they
are
using.
A
All
right,
so
that's
the
end
of
our
discussion
topics
unless
anyone
has
anything
to
add
last
minute.
But
otherwise
I
do
see
some
new
faces
and
I
skipped
the
introduction
part
at
the
beginning
of
the
meeting.
But
since
we
have
some
extra
time,
if
anyone
would
like
to
go
ahead
and
introduce
themselves
if
you're
new
to
this
community
and
this
meeting-
and
you
know
just
say,
hi
and
tell
us
why
you're
here
and
what
you're
interested
in
I'll
give
some
space
for
that.
E
Hey
everyone,
I'm
xander
nice
to
see
you
cecile,
it's
been
a
while.
I
work
for
apple
and
here
hanging
out
we're
looking
at
possibly
setting
up
virtual
cluster
for
conformance
with
cluster
api,
and
so
I'm
just
kind
of
here
to
listen
and
learn
for
now
and.
F
H
Anyone
else
hi
everybody
chris
samp,
I'm
with
vmware
just
trying
to
get
a
pulse
on
what's
up
with
cluster
api
as
we
work
with
it
more
so
kind
of
lurking
and
following
along.
I
A
Cool,
well,
it's
great
to
see
all
the
new
people
here.
If
anyone
has
any
questions,
feel
free
to
reach
out
on
slack
if
you're
not
comfortable,
like
asking
in
this
forum.
I
know
it's
a
lot
of
people
but
great
to
have
you
all
right.
So
I
think
this
is
all
for
today.
Unless
anyone
has
any
questions
for
the
group
or
any
comments,
if
not
we'll
end
it
here,
I
don't
see
any
hands
raising
all
right.