►
From YouTube: Community Meeting, October 18, 2022
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
B
Since
nobody
else
is
raising
I'll
I'll
just
give
an
update
on
the
conversion
work
that
I've
been
working
on,
so
there's
kubernetes,
your
kcp
kubernetes,
pull
requests
and
the
kcp
pull
request
to
add
support
for
conversions
of
it
actually
would
work
for
normal
custom
resource
definitions,
but
it's
primarily
developed
for
API
resource
schemas
and
API
exports
and
API
findings,
but
you
can
define
a
new
type
called
API
conversion
and
there's
a
one-to-one
map
or
mapping
between
an
API
conversion
instance
and
an
API
resource
schema
or
crd
instance,
and
in
there
you
can
Define
conversion
rules.
B
So
you
don't
need
to
use
a
web
hook
to
do
conversions.
We
actually
don't
support
conversion
web
hooks
at
this
time,
and
you
can
also
optionally
do
cel-based,
which
is
a
common
expression.
B
Language
Transformations
I
need
to
probably
go
in
and
update
the
example
there,
because
the
syntax
has
changed
just
a
little
bit,
but
you
can
do
straight
field
to
field
translation
and
Transformations
without
needing
to
use
CDL
and
then,
if
you
do
need
to
perform
some
sort
of
CEO
Logic
on
a
field
to
be
able
to
set
it,
you
can
do
that
with
CDL.
So
I'll
get
that
example
updated
after
the
meeting.
But
if
you're
interested,
please
check
it
out.
B
If
you're
interested
in
reviewing
the
two
PRS,
please
do
so
and
hopefully
that'll
get
in
sometime
soon.
D
Yeah
I
just
want
to
give
some
very
short
introduction
of
what
this
is,
so
this
will
be
visible
at
kubecon.
So,
but
everybody
knows
how
this
is
connected
to
kcp.
So
it's
a
new
project,
so
the
majority
of
the
of
the
code
is
basically
a
bit
more
than
a
week
old
I.
Think
most
of
you
have
seen
this
pitch
at
container
days
about
binding
resources.
Everybody
here
in
the
meeting
knows
binding
right,
binding
API
exports.
D
Basically
this
is
this
topic,
but
qbind
is
basically
the
extension
of
what
we
have
built
in
kcp
into
kubernetes
clusters.
So
the
idea
here-
and
this
is
the
most
important
part
for
kcp
kcp-
is
the
natural
backend
for
this
mechanism.
So
if
you
want
to
offer
an
API
service,
for
example,
you
are
mongodb
mongodb
my
example
in
the
in
the
talk,
if
you
want
to
run
databases
for
customers-
and
you
want
to
want
to
offer
an
API
which
is
kubernetes
based
so
in
crd,
based
API.
D
Obviously,
everybody
here
in
the
meeting
would
use
kcp
to
create
such
an
API
and
controllers
running
against
workspaces,
and
this
Cube
cuddle
bind
is
a
client
side
binding
command
to
get
such
an
API,
which
is
now
just
in
kcp,
but
you
want
it
in
kubernetes
clusters,
so
it
does
something
very
similar
to
the
SIM
card
and
kcp
to
get
cids
down
into
the
local
cluster
from
an
API
resource
schema
in
kcp
parenthesis.
This
part
is
not
implemented
at
the
moment
or
example.
D
They
get
this
just
running
the
cids
in
against
any
cube
in
kcp,
but
just
City
based
there's
work
to
do,
but
Jupiter
bind
gets
a
CID,
basically
from
Upstream
from
kcp
from
a
workspace
and
we're
sync
up
objects
offset
CID,
so
in
mango
one
will
be
example:
you
will
export
and
get
square
API
service
export,
because
it's
not
the
kcp
internal
one,
but
really
really
connected
to
that.
You
will
Define
the
mongodb
and
get
that
into
your
kubernetes
cluster
by
installing
the
CID
from
PCP
into
your
kubernetes.
D
If
you
move
further
there's
a
SIM
card,
basically
we
call
it
connector,
but
it's
very
similar
to
the
SIM
card.
In
the
moment,
you
create
a
mongodb
object
in
Cube.
It
will
zoomed
up,
create
it
in
kcp
in
the
workspace,
so
the
kcp
on
Twitter
will
see
that
do
whatever
it
needs
like
deploy
a
database,
for
example,
and
sync
status
back
from
kcp
from
the
workspace
into
Cube.
D
So
it's
a
very
short
description.
There
are
details
here
about
the
design
about
the
apis
and
you
will
notice
it
replicates.
Basically
what
we
do
with
permission
claims
at
least
division
is
the
same
at
the
moment.
Everything
we
built
here
for
this
coupon
tool
should
match
kcp.
So
kcp
is
a
natural
backend.
That
was
the
first
sentence.
D
I
said
everything
we
do
in
coupon
should
match
kcp,
API,
bindings
and
exports,
and
vice
versa,
so
that
they
are
a
good
match
and
kcp
can
do
what
you
should
do
and
and
agenda
I
linked
to
YouTube.
So
sergius
is
also
here
he
played
with
cockroach
in
the
afternoon,
so
he
he
installed
a
cockroach
Operator,
just
the
stock
operator
which
you
get
from
the
company
and
was
basically
showing
how
you
can
get
crdb.
That's
the
the
CID,
the
the
resource.
D
In
this
context,
how
you
get
it
into
a
cube
cluster,
create
a
cidb
object
for
a
database,
and
it's
a
in
the
background
which
is
here
the
lower
part,
so
the
Stark
terminals,
you
see
that
the
operators
to
their
work,
so
this
opens
up
kcp
basically
to
the
world.
That's
it
yeah.
The
fourth
speech
about
this
project
for
kcp.
D
D
Oh
yeah,
or
so
at
the
right
side.
This
shows
the
example
where
you
run
the
back
end
on
one
kubernetes
cluster.
Obviously,
isolation
is
limited.
You
only
have
namespace
Association
imagine
to
put
a
kcp
there.
I
have
another
slide.
I
didn't
paste
it
here
where
you
have
kcpe,
so
a
workspace
is
sent
on
the
right
side
and
on
the
left
side,
there's
a
consumer
cluster.
D
D
Yeah,
basically,
this
is
yeah.
The
whole
thing
is
a
book
space.
If
you
map
it
to
kcp,
so
there's
one
workspace
app,
which
is
the
upper
namespace,
basically
yeah
the
cluster
namespace.
This
is
basically
The
Meta,
Administration
and
namespace,
and
then
the
user
data
is
in
the
lower
part,
but
the
whole
thing
is
one
per
tenant
or
one
per
class.
Direction.
E
Okay,
I
don't
want
to
take
down
the
meeting
going
details
but
okay,
thank
you.
A
Oh
and
speaking
of
con
we'll
have
a
in
office
hours
for
kcp
on
Thursday
October.
Let's
see
27th
at
2,
30,
Eastern
I
believe
yep,
2,
30
p.m.
Eastern.
A
A
So
we
will
start
going
down
the
list,
so
we've
got
this
from.
F
A
C
F
B
There's
a
chance
that
pull
request.
2189
might
fix
this,
but
I
I
wasn't
specifically
attempting
to
when
I
was
working.
Yeah.
B
Okay,
thanks
Nolan,
that
was
21.89
for
the
pr.
F
This
one
is
fun:
I
created
a
basically
I
use,
export
from
sync
targets
and
both
sides,
exposing
and
consuming,
has
to
provide
identity,
and
the
result
is
that
they
get
their
identity,
hash,
double
decorated,
on
URL,
I've,
pasted,
two
log
lines
below
I
basically
turned
off
the
logging
high,
and
they
can
see
that
identity
has
appended
two
times
to
the
the
URL
and
sending
over
in
404.
A
This
fix,
did
you
put
that
in
bootstrap,
slash,
identity.go
or
both
of
them.
F
C
F
So
API
export
basically
uses
sync
targets
as
their
identities
of
different
apis
and
the
same
on
the
consuming
side.
So
I
think
these
consequences,
because
I
like
this
assumption
that
both
sides
use
the
third
part
API,
so
workloads,
TCP,
Dev
and
both
sides
provided
identity
hash
in
the
in
the
binding
and
export.
C
F
F
A
Okay
was
that
all.
F
Yeah,
thank
you.
Yep.
A
Thank
you,
okay,
I,
don't
know
if
Dan
is
on.
C
A
Look
like
so,
he
assigned
this
to
himself
right
and
this
one
doesn't
have
a
PR
with
it.
C
C
C
Oh
Mo
and
I
were
just
looking
at
the
code
there
and
there
were
a
bunch
of.
G
C
C
A
C
A
B
Hitting
it
fairly
frequently
so
I
don't
was
this
Shogun
that
I
don't
know
who
did
the
placement
test
and
code,
but
it'd
be
nice
if
whoever
it
was
could
take
a
look
at
it?
Okay,
so
David?
Maybe
you
can
delegate
this
one.
A
Okay,
another
feature
for
the
CLI
looking
for
the
Ready
or
Not
Ready
condition
on
API
bindings.
H
Yes,
I
was
assigned
at
first
then
we
have
to
look
into
these
probably
test
a
bit
more.
It's
obviously
some
bugs
in
the
scene
command
that
prevent
restarting
it
correctly.
So
there
is
some
analysis
to
do
here.
F
C
Sorry
we
broke
this
out
yeah,
it's
like
one
of
the
last
as
probably
missing
context.
I
can
fill
more
in
later.
Okay,.
C
Just
we're
missing
an
epic
for
tenancy
I
think
so!
No
okay,
not
yet.
C
A
Same
deal
with
the
the
Epic
link,
there's
no
epic
board
for
it.
Yet,
yes,.
C
C
G
Nolan
yeah
just
one
question:
I'm,
sorry,
because
I
unfortunately
need
to
run.
There
is
a
issue
that
was
raised.
I
asked
a
Community
member
to
raise
it's
these
trimsy
operator,
I.
Believe
it's
608
don't
hesitate
this
one
to
me.
I
asked
the
person
to
submit
an
issue
just
FYI
I'm,
sorry
to
interrupt.
Oh
yeah.
G
And
since
this
involves
the
Sinker,
maybe
somebody
who
has
synchron
knowledge
could
help
me
a
little
bit
in
debugging.
H
B
B
I
don't
know
if
this
is
something
we
should
just.
We
can
keep
it
open
and,
if
folks
run
into
it
make
you
know,
link
references
in
future
comments,
but
there's
not
anything.
We
can
do
with
it
other
than
maybe
upgrade
go.
C
B
C
H
Yes,
I
think
we
can
put
that
in
in
backlog
and
keep
in
mind
as
well
that
we
are,
in
the
middle
of
you,
know
quite
big
epic
well
at
the
end
of
a
quite
big
epic,
about
compute
type,
and
so
probably
this
would
be
an
opportunity
as
well.
At
the
end
of
the
Epic
to
we
will
have
to
check
all
the
docs
and
to
ensure
that
everything
is
consistent.
H
C
A
H
Yes,
it
seems
like
I
think
we
can
put
that
in
in
backlog
and
also
I'll.
You
can
just
assign
it
to
me
or
something
like
that,
because
it
seems
to
me
that
it
would
have
we
would.
We
would
have
a
similar
requirements
for
other
parts
of
the
TMC,
especially
for
interest
or
networking
for
also
security
related
aspects
of
TMC
networking.
We
will
have
to
be
able
to
select
a
number
of
Downstream
namespaces
according
to
the
corresponding
that
stream
workspace.
H
C
D
H
Well,
in
this
issue,
it's
still
allowing
promoters-
or
you
know
such
third
party
tools
to
select
them
so
I
mean
depends
on
what
you
call
user.
Yes,.
H
So
yes,
yes,
I
know
so
I
mean
that's.
Why
I
said
we
have
to
to
have
to
put
some
30
in
this
and
and
to
have
some
design,
because
it
seems
to
me
that
initially
precisely
the
change
to
switch
from
Sing
Target
name
to
sim
Target,
you
know
hash
and
all
this
was
also
to
you
know:
office
guilt,
obfuscate,
part
of
it
for
the
end
user.
So
yeah.
We
have
I
think
to
to
think
about
this
more
more
deeply
and
really
Define
the
the
limits
and
the
scoping
of
what
we
want
here.
A
This
is
no
intention
for
this
to
go
on
an
epic
right
now.
B
C
H
Yes,
this
one
you
you
can
put
in
in
a
backlog.
Yes,
we
we
just
analyze
the
the
thing
with
yeah
and
Anthony.
You
know
one
hour
ago,
okay,
and
it
seems
just
an
inconsistency
between
between
the
the
AP,
what
is
defined
in
the
API
in
the
comment
and
and
the
implementation
of
it.
C
B
For
for
PNF
Mike,
do
you
know
what
Jamie's
made
any
progress.
B
I
mean
ideally
I
would
say,
given
that
kubecon
is
next
week
and
folks
aren't
going
to
have
time
to
really
focus
on
reviewing
it
kind
of
needs
to
get
pushed
for
review
in
the
next
day
or
two
or
it's
going
to
slip
to
0.11..
Well,.
E
Let
me
just
text
him
right
now
and
can
I
get
back
to
you
in
a
couple
minutes
sure
thing:
okay,.
H
Yeah
this
one
was
not
closed,
but
mainly
is
being
superseded
by
the
by
the
next
one,
compute
type
and
multi-sinker
thinking
so
multi-sync
Target
thinking.
C
A
I
think
that's
it.
Anybody
have
anything
else.
E
Well,
I
am
Jamie's
typing
his
reply
to
me
right
now.
Oh
okay,.
C
E
Negative
yeah:
he
won't
be
ready
this
week.
Okay,.