►
From YouTube: 20200713 - Cluster API Provider AWS 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
hello
and
welcome
to
the
July
13th
edition
of
the
cluster
API
provider,
AWS
office
hours,
a
sub-project,
a
bolts,
a
cluster
lifecycle
and
cluster
api.
Please
remember
that
this
meeting
will
be
recorded
and
posted
up
to
YouTube
afterwards,
and
this
meeting
does
comply
with
the
kubernetes
code
of
conduct.
So
in
general,
please
be
excellent
to
one
another
to
get
started.
It
looks
like
we
have
at
least
one
PSA
I
think
that
was
you
that
added
it
nadir.
B
Yep,
so
we
just
didn't
after
release
last
week
or
zero
five
five,
so
that
includes
conditions
are
now
being
added
to
a
diverse
class
tonight
on
this
machine.
Debilitate
is
support
the
customization
of
ingress
rules
for
CN
eyes
and
some
changes
to
the
subnet
logic.
So
to
include
so,
we
can
now
deploy
multi
AZ
control
planes
and
also
the
rewrite
your
class
aid
on
the
sadm,
which
has
the
new
api
so
yeah.
So
quite
quite
a
lot
of
changes
so
go
check
it
out
we'll
be
doing
a
final
release
or
I'm,
not
sure.
B
C
C
A
D
Yeah
good
morning,
I'll
probably
add
this
topic
to
the
next
couple
of
these.
We
can't
have
a
mini
milestone
going
on
in
here,
but
yeah.
So
we
have
been
working
on
the
e
KS
bootstrap
provider.
I
think
we'll
be
in
a
place
where
we
can
open
a
whippy,
or
at
least
today,
so
people
can
kind
of
see
what's
going
on,
and
then
we
can
decide
like
what
to
merge.
When
you
know
we
can
do
like
half
implementations
or
whatever,
but
we
were
pretty
far
along
on
the
bootstrap
provider.
D
The
the
thing
that
kind
of
stops
us
from
being
able
to
test
is,
of
course,
a
matching
control
plane
provider
to
to
use.
There
are
some
hacks.
We
were
considering
with
just
like
making
a
joining
a
pretend
eks
cluster
or
something
like
that,
but
it
largely
works
and
we're
working
on
putting
together
the
tests
now
and
then
the
other
side.
So
half
of
our
team
is
working
on
the
a
Tobias
machine
pool
implementation
that
just
started
work
at
the
end
of
last
week.
So
we'll
see
what
happens
there?
D
C
I
think
somebody
from
your
team
Andrew
had
a
cap
B
issue
about
the
tilt
providers
file
and
wanting
to
support
multiple
providers.
I
added
a
comment
sometime
last
week:
I
think
about
trying
to
keep
it
backwards
and
forwards
compatible
to
support
either
one
or
multiple,
whether
it's
a
hash
or
an
array.
I,
don't
know
if
you
all
had
seen
that
yeah.
D
E
F
It's
sort
of
related
to
what
Andrew
just
mentioned,
so
I've
done
a
refactor
of
the
cluster
scope
and
related
services,
so
that
it
can
be
reused
between
the
current
AWS
cluster
stroke,
a
Demarest
machines
and
some
future
managed
versions
of
those
I've
tested
it
out
with
our
implementation
of
the
control
plane.
So
I
eks
control
plane
just
to
see
if
it
would
work,
and
then
it
does
so
when
this
does
merging
I
will
then
be
able
to
create
and
update
the
work-in-progress
PR
for
the
eCos
control
plane
with
the
new
AWS
managed
cluster.
F
Based
on
the
discussions
we
had
last
week
about
splitting
now
and
not
having
in
the
conditional
the
conditional
logic
within
the
current
AWS
cluster
controller,
really
just
followed
the
with
this
free
fact.
I
really
just
followed
the
pattern
that
the
caps
heads
have
introduced.
So
it's
nothing
new
I!
Guess
it's
really
yet,
because
you
get
that
in
at
some
point,
if
possible.
Just
so,
we
can
carry
on
with
the
control
plane,
stuff.
C
A
D
C
C
So
maybe
we
do
something
similar
like
you
could
split
out
the
I
guess
you
could
have
one
yeah
mole
for
eks
and
another
yeah
mole
for
what
we
currently
have
and
then
an
aggregated
one.
That's
got
them
all
together.
If
we
really
wanted
to,
although
I
don't
know
would
would
somebody
want
to
be
doing
both
in
a
single
management
cluster.
D
Yeah
I
mean
so
that's
an
interesting
question.
I
definitely
think
there
will
be
a
couple
of
clusters
of
CTL,
like
maybe
tweaks,
only
because
like
the
well,
maybe
not
clusters
ETL,
but
you're.
Certainly
not
gonna,
install
well
I,
don't
know
how
I
feel
about
to
control
plane
providers
like
how
that's
how
that's
gonna
go
like
how
well
they
would
work
together
position
and.
D
From
this,
but
from
the
self-managed,
where
it's
very
hilarious
but
yeah
I
mean
I,
think
they
could
play
well
together.
I
would
also
love
to
try
and
join
eks
nodes
and
qadian
nodes
to
like
the
same
control
point
it
just
it
just
wouldn't
work.
D
C
Yeah
with
regards
to
multiple
control,
plane,
controllers-
I-
don't
know
if
there's
anything
in
cluster
cuddles,
specifically
that
would
need
to
change.
But
the
api's
are
designed
to
support
multiple
in
the
same
management
cluster,
because
we
have
the
control
plane
breath
from
the
cluster
to
the
actual
instance
of
qubit
and
control,
plane
or
dks
control,
plane
and
presumably
the
individual
control
plane
providers
won't
be
bothering
to
deal
with
clusters
that
don't
link
to
them
for
the
type
of
control
plane.
It
is.
C
D
A
F
A
A
A
B
A
D
A
G
C
B
So
as
soon
as
you
run,
Dinko
with
more
than
one
node
is
gonna
have
problems
if
you've
done
with
the
default
account
limits.
So
we
decided
to
pivot
from
doing
not
an
existing
e
to
e
testing
to
doing
it.
If
we're
gonna,
do
it
on
the
new
cluster
cuttable
framework
based
testing
and
D
dupe
some
of
these
tests.
So
we
got
less
individual
clusters
being
built,
and
maybe
some
external
lock
to
be
able
to
serialize
Dinko
a
bit
which
you
can't
do
with
engine
code
by
default
and
and.
A
B
A
Yeah,
so
I
do
think
that's
you're
realizing
some
of
these
out
is
a
good,
relatively
short-term
solution.
I
think
we
probably
also
want
to
look
at
potentially
seeing
what
we
can
do
about
increasing
the
limits
in
the
for
the
boscoe's
accounts,
but
we
would
need
to
work
with
Justin
Santa
Barbara
to
sort
some
of
that
out.
A
A
So
I'm
went
ahead
and
I
put
priority
important
soon
on
here,
because
I
do
think
we
want
to
have
testing
around
the
defaults
I.
Don't
necessarily
know
that
this
needs
to
be
a
blocker
four
zero.
Five
five,
though
so
I'm
tempted
to
make
it
zero
five
x.
Does
anybody
else
have
thoughts
for
the
milestone
here.
A
A
G
A
Yeah
all
right
so
I
will
go
ahead
and
I
think
for
it.
Yeah.
Let's
get
this
one:
zero,
five
X
and
then
AWS,
multi-tenancy
I
know
this
is
going
to
be
somewhat
of
a
larger
change.
Is
this
something
that
would
likely
target
zero
five
X?
Or
is
this
something
that
we
should
probably
look
towards
zero
at
0.64
I?
Think
zi6
is.