►
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
A
B
Awesome
before
I
ask
my
question
just
a
little
bit
of
context,
so
I'm
coming
from
I
want
to
use
class
API
right
and
ideally
I
want
to
use
it
without
class
Takada,
so
meaning
I
wanna
have
control
cluster
and
I
wanna,
just
like
basically
provision
or
cluster
and
machine
resources,
machines,
etc.
Right
and
up
experiment
with
it
and
it
works.
Fine.
A
little
bit
of
AppLocker
right
now
is
actually
getting
credentials
right.
So
the
cube
config
for
the
control
plane
after
you've
ruined
it.
B
It
seems
like
the
cluster
cuddle,
at
least
for
TCP,
from
what
I
saw.
Does
a
tree
cloud
SSH
to
get
it
or
maybe
SCP
I,
don't
know
and
I
looked
at
a
couple
of
other
providers.
They
are
all
doing,
as
is
H
what
I've
seen
now.
This
is
a
bit
of
a
problem
right
for
me
because
ideally
I
want
just
like
getting
access
to
these
credentials
in
the
control
class
itself.
So
ideally
I,
don't
know,
there's
something
that
puts
it
into
the
ctrl
+
S
or
E
as
a
secret,
for
example.
B
B
Could
do
that
or
you
could
maybe
think
about
like
having
it
kind
of
like
a
callback
where
the
control
script
on
the
master
basically
puts
it
back
into
the
control
cluster
like
with
you
know,
using
like
some
service
account
that
is
very
limited.
Just
can't
create
a
secret
in
the
cluster
main
space,
something
like
that.
Assuming.
A
A
A
A
Getting
the
coop
config
from
the
master
without
SSH,
there
hasn't
really
been
any
kind
of
work
on
it,
and
160
is
the
opposite
of
that
to
fetch
the
coop
config
from
a
master
via
SSH,
which
is
all
about.
How
do
we
move
this
upstream
and
I
would
very
much
like
to
see
that
not
be
a
dependency,
because
there
are
some
systems
that
just
don't
have
SSH
enabled
all.
B
B
Sorry,
I
really
don't
want
to
I
checked
this
meeting,
but
so
many
questions
I
mean
you
so
my
control
cluster
is
not
a
cluster
api
created
cluster.
My
bootstrap
cluster
I
just
use
a
different
cluster
right
and
I
can
create
master.
No,
it's
just
fine,
previously
IDs
I've
problems
with
the
with
the
FO
notes,
with
the
worker
notes.
The
reason
being
is,
it
seems
like
right
now
we're
using
cube
a
cube
ATM
to
do
that,
and
we
need
to
get
like
some
sort
of
token
inner
I.
B
Don't
exactly
why
and
what's
going
on,
I've
looked
into
it,
but
that
depends
on
the
fact
that
there
is
a
cube
config
on
that
note
where
the
con,
where
the
controller
runs
that
has
that
the
cube
ATM
can
access
to
which
isn't
at
EDC,
kubernetes,
config
or
whatever
or
admin.
Admin
comes
now
in
my
cluster
that
doesn't
exist
right
because
it's
not
a
cube,
ATM
strapped
cluster.
So
yet
another
thing
of
like
the
question
of
like
are
you
aware
of
this-
has
has
come
up
before
anything.
I
can
look
at.
A
A
Theoretically,
you
are
able
to
use
other
bootstrapping
mechanisms
other
than
cube
ATM,
but
most
of
the
providers
have
an
assumption
that
you
are
using
cube.
Atm
is
to
bootstrapping
mechanism
and
have
a
lot
of
that.
Hard-Coded
in
I
know
that,
for
example,
the
OpenStack
provider,
which
is
what
everybody
also
on
this
call
works
on.
A
We
have
in
all
of
our
scripts.
You
know
that
we're
using
to
start
up
the
machines,
etc
heart
dependencies
on
cab,
ATM,
there's
some
stuff
in
the
code
that
depends
on
Kubb,
ATM
and
even
until
recently
depended
on
you
running
a
specific
version
of
Ubuntu
and
so
we're
just
trying
to
slowly
abstract
those
out
and
find
out
where
we
can
draw
those
lines
and
what
things
we
actually
need
to
have
available
as
abstractions.
A
So
it
certainly
is
something
that
we're
aware
of,
but
it's
definitely
something
that
we
would
like
to
see
more
visibility
raised
to.
So,
if
you
do
come
across
those
things,
it
would
be
fantastic
if,
in
the
specific
provider
that
you're,
using
or
even
in
the
cluster
API,
if
you
can
go
ahead
and
open
issues
around
those
kind
of
things
so
that
we
have
more
visibility
on
these
are
the
things
that
we're
running
into
the
other
side
of
the
coin.
A
B
A
Yeah
absolutely
there's
quite
a
bit
of
work
with
the
COO
baby
stuff
going
on
right
now,
trying
to
move
some
of
that
stuff
out
of
the
tree
as
well.
So,
for
example,
bootstrapping
tokens
is
a
thing
that
the
code
lives
in
cuvee
DM
right
now,
even
though
it's
technically,
he
is
interacting
with
kubernetes
kubernetes,
so
we're
working
on
extracting
that
code
and
enabling
other
providers
other
than
cuvee
DM
to
use
that
that
same
bootstrapping
token
mechanism.
So.