►
From YouTube: 2020-08-24 - 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
A
We
abide
by
the
kubernetes
code
of
conduct,
which
means
please
be
kind
to
everyone
and,
among
other
things,
please
use
the
raise
hand
feature
of
zoom.
If
you
would
like
to
participate
in
the
discussion-
and
please
add
your
name
to
the
attending
list
as
well,
if
you
have
any
psas
or
group
topics,
please
feel
free
to
add
them
and
we'll
go
from
there.
Does
anybody
have
any
psas.
A
A
Did
you
all
want
to
talk
about
the
eks
work
that
is
going
on
check
in
on
a
progress
update?
Maybe.
B
C
We
could
probably
do
that.
I
know
so.
Basically,
the
there's
like
a
big
waiting
cascade
of
prs
that
will
all
basically
flood
in
in
sequence,
so
the
control,
the
bootstrap
provider
now
now
depends
on
the
control
plane
pr,
so
they
can't
necessarily
be
merged
independently
of
each
other.
C
So
what
I've
been
doing
is
just
kind
of
as
as
richard
who
is
here,
I
think
makes
changes.
We
just
keep
testing
and
finding
new
little
things
that
need
to
get
fixed
and
just
kind
of
work
through
those,
but
I
think
the
bootstrap
provider,
for
the
most
part
is
ready
to
go
in
just
waiting
for
importable
types.
A
Okay
cool,
so
I
saw
one
request
to
edit
the
doc
come
through.
Hopefully
having
this
in
bigger
text
will
make
that
clear.
Please
join
the
google
group
and
you'll
get
it
edit.
Access
to
the
document.
B
Yeah,
it's
really
just
to
add
to
what
andrew
said
is:
there's
a
few
outstanding
things
on
the
control
plane.
I've
got
an
issue
with
the
mutating
web
hooks
at
the
moment.
I
think
I've
just
realized
that
it
was
some
pep
cap
going
on
on
my
side,
so
hopefully
that'll
be
fixed
and
then,
as
far
as
I'm
aware,
it's
just
the
outstanding
issue
on
the
cluster
name,
stroke
kubernetes
cluster
name.
B
D
Yeah,
so
it's
I
think
it's
probably
worth
going
through
the
milestone
for
zero
six
zero.
In
relation
to
eks,
I
made
a
bunch
of
guesses
and
there's
quite
a
lot
of
issues
in
that
milestone
relating
to
egas,
but
you
should
decide
what
to
punt.
A
Okay,
do
you
do
y'all.
B
A
So
I
don't
know
how
this
milestone
view
sorts
things
unless
it's
like,
I
recently
updated,
because
it's
definitely
not
numerically,
so
this
may
not
be
the
best
view
to
use
or
we
can
just
go
through
it,
and
if
we
see
something
twice,
because
it's
an
issue
in
a
pr,
we
can
just
skip
over
it
the
second
time.
A
So
why
don't
we
try
this?
So
this
is
everything
in
the
060
milestone.
Now
I
missed
the
milestone
and
backlog,
grooming
and
planning
when
I
was
out
on
vacation
a
couple
weeks
ago.
So
if
y'all
can
catch
me
up
to
speed,
I
do
see
that
trying
to
target
a
week
from
today
for
that
milestone.
Does
that
still
sound
doable
or
like
do
either
well
number
one?
Does
that
sound,
doable.
A
C
So
probably
all
33
things,
perhaps
not,
I
think
the
the
thing
that
we
were
looking
to
get
out
of
the
end
of
the
month
was
just
basically
just
parody
functional
clusters
for
the
most
part,
so
that
would
be
the
bootstrapper
and
the
control
plane
like
as
far
as
I'm
concerned,
if,
if
those
two
things
go
in
and
there
are
no
like
earth-shattering
concerns
with
with
like
the
way
that
it
works
or
anything
like
that,
like
that's
a
success,
I
know
for
a
fact
that
some
of
these,
like
enhanced
ami
lookup
logic,
it's
just
going
to
take
some
time
on
its
own,
like
there's
a
whole
like
research
element
to
that
right.
A
So
why
don't
we?
Why
don't
we
look
for
things
that
are
not
critical
to
zero
six
zero
and
move
them
at
least
to
the
next
patch
sure,
okay,
so,
let's
just
go
from
top
to
bottom,
because
that's
easier
to
reason
through
so
default.
I
am
role
with
the
cluster
aws
adm.
B
Yep
tool
could
probably
go
to
the
next
one
because
short
term
people
can
enable
automatic.
I
am
roll
creation
by
a
feature
flag
if
they
want
to.
A
Okay,
I
don't
know
if
this
is
the
same
for
everybody
else,
but
your
mic
volume
is
pretty
low
for
me
richard.
If
maybe
you
can
up
that.
That
would
be
helpful.
I
think
you
said
move
this
to
zero.
Six
one
is
that
right:
okay,
next
up,
managed
control,
plane
controller
should
set
up
a
config
map
automatically
yeah.
We
can
push
that
forward
too.
C
C
That
pr
is
open
in
its
basic
form:
okay,
so
this
can
stay
for
now,
yes,
yeah
that'll
be
good
to
go.
The
enhanced
lookup
is
like
extension
of
of
that
capability.
Okay,.
A
So
you
mentioned,
bootstrap
provider
is
necessary,
so
I
assume
leaving
this
here-
makes
sense.
Self-Managed
node
group
support.
C
There
that
pr
is
that
pr
is
technically
independent
of
vks.
There
will
be
code
changes
to
those
logic
branches
like
when
eks
exists,
because
actually,
that
might
not
even
be
true,
so
aws
machine
pool
when
implemented,
will
support
both
eks
config
and
kubernetes
mq
and
cube
adm
config
together,
and
that
should
resolve
this
issue.
C
I'll
have
to
double
check
the
comments
currently
on
that
pr,
I
know
me
too
was
working
that,
but
the
pr
is
open
and
largely
complete
by
my
understanding.
There
are
a
couple
follow-up
items
that
I
think
have
already
been
broke
out
and
broken
out
into
separate
issues
that
may
not
even
be
in
this
milestone.
A
Okay,
well,
I'm
going
to
leave
it
alone
for
right
now
and
we'll
we
may
come
back
to
it
if
we
need
to
all
right,
so
we
have
a
control
plane
issue
which
we
need.
This
I'm
assuming
is
like
the
overall
sort
of
epic
or
umbrella
for
eks.
C
Yeah,
if
we're
actually
going
to
cut
a
release
for
o60
we,
I
can
probably
spend
a
couple
minutes
figuring
this
out.
It's
really
just
looking
at
the
way
cappy
breaks
down,
it's
yaml
files.
Do
we
want
the
eks
things
to
all
show
up
in
separate
yaml
files?
I
assume
so-
and
you
know,
like
any
make
targets
required
to
make
that
work
or
changes
to
existing
make
targets.
Okay,.
A
A
A
Okay,
this
is
the
performance
one
which
is
not
specifically
eks
related
nadir.
Are
you
actively
working
on
this
one
still.
D
Yeah
yeah,
so
I
started
off
adding
scale
tests
stuff,
which
I've
got
in
a
branch.
I
need
to
push
up
yeah.
I
am
working
on
that
right
now
and
there's
a
bunch
of
related
issues,
you'll
see
which
are
sort
of
kind
of.
A
A
All
right.
We
have
the
ede
pr
for
installing
cni
with
cluster
resource
set,
but
I
guess
so
death
we're
waiting
for
a
cappy
release.
A
Yeah
we
can
leave
this
for
zero
six
zero
for
now,
but
if
the
timing
doesn't
work
out,
we
can
always
do
it
as
soon
as
we
get
zero.
Three
nine
vendor.
Again,
it's
awesome.
D
A
Okay,
yeah
feel
free
to
tag
me
and
I
can
look
at
stuff
when
you've
got
it
ready.
A
Okay,
documentation,
for
I
am
roles,
help
wanted.
A
I
mean
we,
we
do
have
the
ability
to
generate
the
policy
files
from
the
command
line.
Is
this
like?
Do
you
all
think
that
this
should
go
into?
D
Yeah,
I'm
actually
thinking,
maybe
so
I
think
this
is
stuff
that
was
in
the
05x
milestone
left
over,
which
I
just
punted.
D
A
A
A
Okay,
next
up
is
the
pr
to
look
up
the
eks
ami.
So
I'm
assuming
you
all
need
this
for
the
full
eks
supporting
zero
six
zero
right.
A
What
is
this,
I
vaguely
remember
this
forward
port,
something
the
root
volume
should
inherit
additional
tags
from
the
instance.
Oh,
that
is,
this,
isn't
done,
doesn't
look
like
it.
D
C
I
I
can
check
the
code
on
this
one.
The
reason
why
I
think
it
might
be
done
is
there
was
like
a
big
pr
that
changed
the
structure
of
the.
B
C
D
A
A
Custom
fashion
am
I
I
know
we
have
a.
This
is
a
pr
I
think
like
if
it
makes
the
cut
great
if
it
misses
zero.
Six,
because
I
know
we're
waiting
on
some
code
updates
from
comments.
Then
they
can
roll
to
zero
six
one.
F
A
Okay,
update
documentation
for
aws
manage
cluster
cube
configs
richard
since
I
see
andrew
had
to
drop.
Is
this
one,
something
that
we
need
for
zero?
Six
zero.
B
Potentially,
just
because
there
is
two
cube
configs
now
one
is
just
purely
for
cappy
and
then
one
is
for
the
user.
Okay,.
F
A
Sent
us
seven
fails,
I
know
there.
There
was
a
full
need
for
the
image
builder
pr
to
pull
in
centos
and
rel
support
did.
D
It's
fixed
we're
waiting,
waiting
for
a
new
tag
of
image
builder
and
then
generate
amis
based
of
that
tag,
rather
than
building
straight
off
master.
So
travis
said
it
was
going
to
happen
this
week.
F
D
E
Right
so
in
case
of,
we
cannot
find
the
instance
that
we
are
looking
for.
I'm
not
sure
what
should
be
the
way
to
go
other
than
retrying
re-entering
yeah
material
is
mentioning.
There
may
be
a
vmts
problem,
but
if
instance
is
not
there.
This
may
be
caused
a
lot
of
reasons
so.
A
Okay,
how
about,
if
you
and
I
look
at
this
sometime
this
week
and
see
if
we
can
come
up
with
a
plan
for.
A
A
F
A
A
B
F
A
Okay
and
then
we
have
ci
pre-submit,
should
make
sure
that
the
e
to
e
tests
compile,
I
thought
somebody
was
yeah.
I'm
gonna
zero,
six
x
this
and
check
in.
B
A
And
I'm
gonna
put
this
into
the
x's
okay,
so
I
think
now
that
we've
looked
at
everything
the
bulk
of
it
tends
to
be
eks
with
a
couple
of
potential
bugs
or
scaling
improvements,
and
I
think
we
can
just
touch
base
on
slack
over
the
next
few
days
to
see
if
it
looks
like
we
need
to
adjust
the
date
or
move
anything
in
or
out
that
sound
good
to
everybody.
A
Okay,
any
other
topics
questions
anything
worth
chatting
about.