►
From YouTube: Community Meeting, December 6, 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
Everybody
today
is
December
6th,
and
this
is
the
kcp
community
meeting
if
we
don't
have
an
agenda
for
today.
So
if
you've
got
anything
that
you'd
like
to
discuss,
please
feel
free
to
use
the
raise
hand
feature
in
Google
meet
and
and
I'll
moderate
I
did
want
to
start.
Just
by
talking
about
the
0.10
release,
we
are
hoping
to
try
and
cut
it
this
week.
The
tentative
date
is
Thursday,
although
we
may
do
it
sooner
or
later,
depending
on
closing
out
the
Milestone.
So
I
was
thinking.
A
I
could
start
by
just
running
through.
What's
in
the
Milestone
and
that'll,
give
you
all
some
time
to
think
about
if
you've
got
stuff,
you
wanted
to
chat
about.
So
let
me
just
pull
this
up.
Real
quick
in
the
screen
share.
A
Okay,
so
I
am
going
to
start
at
the
and
we'll
start
at
the
top.
So
we
have
reverse
permission,
claims
and
I
know
sergius
wasn't
able
to
attend
the
meeting
due
to
a
conflict,
but
we
do
have
one
PR
I,
think
that's
remaining
or
VPR.
A
A
Next
up
is
compute
type
management,
and
so
it
looks
like
this
got
merged
David,
so
I
think
we
can
close
this.
B
A
Okay,
please
feel
free
to
take
initiative
and
you
know
close
things
out
when
they're
ready.
So
that's
awesome
that
this
is
done.
A
A
Okay,
this
one
around
the
crud
operations
issue-
there
is
a
PR
open
for
this,
and
sergius
is
going
to
review.
D
A
B
Well,
the
thing
is:
it's
quite
old,
in
fact,
it's
it's
mainly
just
an
additional
controller,
which
is
not
part
of
the
main
kcp
which
is
just
optional
and
which
is
which
is
based
on
things
already
merged.
So
I
mean
it's.
It's
been
in
review
for
some
time.
In
fact,
I
think
it's
code
that
was
already
agreed
upon,
but
no
it's
not
critical.
It's
just
a
Pity
that
that
it's
not
merged,
but
it's
not
critical.
You
know
features
for
0.10
I
mean
so
so.
B
B
Yeah,
it's
it's
mainly
just
you
know
the
first
coordination
controller
based
on
all
the
on
the
helpers
and-
and
you
know
the
coordination
controller
and
and
the
thinker
views
that
I
showed
last
week.
So
those
the
helpers
and
and
the
main
mechanism
is,
is
merged,
and
this
one
is,
you
know,
bringing
back
the
deployment
Splitters
speaker
based
on
on
that.
So
you
know
I
mean
it's
not
critical.
It's
just.
A
Okay,
I
I
would
prefer
that
this
get
reviewed
and
I
realize
code
got
moved
around.
That's
that's
cool,
so
if
I
can
find
some
time,
I'll
try
and
take
a
look
Stefan.
If
you've
got
time.
I
know
you
probably
don't.
A
A
Next
up
is
the
this
combo
of
updating,
DNS
resources
David.
What's
the
review
status
on
this
one.
B
I
I've
reviewed
already
I
want
and
Lionel
tested
that
as
well
on
on
his
side.
I
just
wanted
to
to
do
a
last
path,
and
then
we
should
be
good.
Okay,.
A
B
B
B
Yes,
I
just
did
this
one
because
I've
been
working
recently
on
cleaning
the
core
DNS
resources
when,
when
no
results
from
for
a
given
Upstream
workspace
exists
anymore,
but
obviously
this
will
split
20.11.
But
while
doing
that,
I
just
realized
that
there
was
some.
You
know:
okay,.
A
Well,
let's
get
this
review
and
it's
it's!
This
is
small,
so
it's
cool
yeah.
B
A
So
I
think
there's
there's
not
all
that
much
left,
which
is
awesome.
So
if
folks
want
to
and
have
review
time,
there
are
four
open,
PRS
so
feel
free
to
to
come
in.
If
you
are
interested
and
hopefully
we'll
be
able
to
do
the
release
later
this
week,.
A
Okay,
so
Stefan,
do
you
want
to
chat
just
a
little
bit
about
what's
to
come
after
0.10
with
the
workspace
refactor
just
to
give
a
bit
of
an
overview.
E
Yes,
I
can
do
that
thinking.
What's
the
best
way
is
so
we
have
a
watch
in
it's
actually
here
in
kcp
called
main
mind
or
something.
If
you
can
go
there.
E
Two
blonde
launches
one
of
those
and
he
has
a
rebates,
one
that
one
yeah
and
that
one-
and
we
actually
have
a
follow-up
of
that-
and
this
is
basically
like
a
feature
of
orange-
and
you
see
the
number
of
commits
pretty
massive
what
we
are
doing-
and
we
talked
about
that
before
we
are
changing-
how
authorization
Works,
how
lots
of
controllers
check
for
what
a
workspace
actually
is
like
they
look
into
the
workspace
and
there's
a
new
object,
called
this
workspace
at
the
moment,
and
we
will
likely
rename
that
again
to
luxury
clusters.
E
We
have
planned
originally-
and
lots
of
controllers
are
touched
in
here
to
to
look
on
this
object
inside
of
the
workspace
no
lookup
anymore
into
the
parent,
and
the
change
is
massive
because
basically
every
controller
is
touched
and
every
controller
gets
a
new
Informer
and
or
at
least
the
the
old
workspace
Informer
is
replaced
by
the
this
workspace
Informer.
E
We
have
inverted
how
workspace
and
cluster
workspace
relate.
So
the
real
object
in
this
branch
is
a
workspace.
Cluster
workspace
is
still
there,
but
just
to
get
us
over
the
line
in
end-to-end
tests.
So
we
will
remove
the
projection
of
class
Network
spaces
when
we
have
moved
everything
to
workspaces.
So
it's
much
simpler.
There
is
no
airbag-based
filtering
anymore,
so
this
whole
part
goes
away.
E
What
else
one
goal
was
to
decouple
the
hierarchy
like
the
salute
colon
Something,
Something
Something
from.
E
The
actual
logical
cluster
contact,
so
we
were
with
this
Branch
we
will
have
more
than
one
root.
Workspace
or
woodwork.
Space
is
a
long
version,
so
the
woodwork
space
is
a
root
right.
It's
a
root
of
hierarchy.
That's
why
it's
called,
but
there
will
be
more
hierarchies
more
trees
if
you
want
so,
for
example,
your
home
workspace.
E
Yes,
exactly
and
you
could
even
run
it
without
any
hierarchy,
so
API
bindings
and
exports
we
work
with
outside.
But
if
you,
if
you
have
the
tenancy
API
Group
enables
you
have
to
follow
us,
that's
the
point.
So
home
workspaces
will
live
in
the
whole,
their
own
subtree,
basically
in
this
house
and
also
an
organization
for
example,
could
also
be
a
subtree
or
you
can
have
multiple
root
workspaces,
replicas
of
that
so
much
much
more
flexible
and
the
last
thing
which
we
do
to
make
this
work.
E
We
so
we
changed
what
the
lottery
cluster
is.
So
today
we
check
forward
and
we
check
for
system
colon.
So
there
are
just
a
few
prefixes
which
are
allowed
in
the
future.
That
can
be
basically
every
so
we
call
it
a
cluster
name,
so
the
workspace
has
an
lcdp.
Everybody
knows
that.
So
there
are
those
keys
and
let's
see
the
prefix
is
basically-
and
we
put
we
have
put
in
the
past-
the
hierarchy.
Parts
like
wood,
colon
workspace
colon
through
something
like
that.
We
have
put
into
that
the
required
that's
in
prefix.
E
What
we
do
now
in
this
branch
is
basically
adding
another
concept,
which
is
probably
just
a
random
base.
36
ID
we're
just
distinct
problems
from
the
path
in
the
hierarchy,
so
workspaces
will
get
something
called
a
cluster
ID
or
cluster
name.
That's
a
thing
that
India's
highlighting
here,
so
this
ID
can
be
used
in
logical
cluster
path
to
go
down
into
sub
workspaces.
E
So
you
can
take
the
ID
of
your
homework
space,
which
is
some
random
number,
random
character,
combination
and
then
say
colon
software
space,
and
this
is
a
valid
launching
cluster
string
as
well.
So
they
are
much
more
flexible
and
by
doing
that
in
the
future,
we
can
do
things
like
moving
workspaces
in
the
hierarchy
somewhere
else
and
this
identity.
This
cluster
name
will
not
change
when
you
do
this
operation,
because
the
hierarchy
path-
and
this
is
existing-
it's
a
different
thing.
E
Yeah
there
are
examples,
so
there
are
different
ways
to
address
one
workspace
higher
level
better
level.
This
branch
is
breaking
so
disruptive
in
a
sense
that
the
apis
change
like
cluster
workspace,
basically
goes
away.
It's
still
there,
some
protection,
but
approximately
so
you'll,
think
API
bindings
will
only
reference.
Those
new
cluster
IDs
cluster
names.
That's
a
that's
a
change!
We're
just
creaking.
E
E
So
there
are
a
couple
of
things
which
will
change
so
the
CLI
is
mostly
compatible,
but
I
also
think
you
need
a
new
version
for
that.
Maybe
last
word
before
I
finish.
The
motivation
of
that
is
starting.
So
if
we
want
to
horizontally
scale
it's
bad
to
look
into
the
parrot,
that's
basically
there's
a
whole
reason
why
we
do
all
of
that,
so
this
Branch
will
change
that
it
will
allow
us
to
operate
workspaces
when
everything
else
of
the
kcp
system
is
down.
E
So
it's
important,
there's
basically
No
Way
Around
such
a
change,
and
we
combine
a
couple
of
things
here
because
we
are
breaking
anyway.
So,
of
course
we
fix
things
we
had
plans
because
we
have
to
wipe
at
the
defaults
out
if
you're
running
at
City
deployment,
there's
kcp.
The
data
must
be
right.
A
Stefan
and
so
I
would
add,
once
0.10
is
released,
the
main
branch
is
basically
going
to
be
on
hold,
while
Stefan
and
lukash
and
others
finalize
this
feature.
Branch
we'll
get
the
feature,
Branch
merged
main
will
open
back
up,
and
then
there
will
be
probably
numerous
follow-ups
to
you
know
fix
whatever
sort
of
Ripple
effects
that
we
didn't
catch
in
the
in
the
initial
merges.
C
A
If
we
can
get
0
10
this
week,
that
would
be
ideal
and
then,
when
Stefan
and
Lucas
are
ready
for
the
branch
so
like
we
could,
we
could
accept
some
PRS
to
come
in,
but
they
would
need
to
be
probably
reviewed
just
at
a
high
level
to
make
sure
that
they
don't
have
a
Major
rebase
Impact
for
the
feature
branch.
E
Yeah
I
think
a
week
or
so
it's
another
time
frame.
So
it's
compiling
it's
starting
up.
We
haven't
done
a
last
step,
so
we
have
to
replace
a
couple
of
libraries
or
bump
them
to
the
next
version
clusters:
okay,
Machinery,
the
clients.
E
We
are
integrating
that
right
now
and
we
have
to
do
another
walk
through
all
the
end-to-end
tests
and
fix
them.
So
we
had
some
green.
Everything
was
green
already,
but
we
have
we
hadn't
done
the
cluster
ID
circulation.
Yet
that's
the
next
step
and
yeah.
Maybe
a
week.
A
F
So,
a
couple
weeks
ago,
we
merged
the
enhancement
for
process
improvements
and
we're
going
to
need
some
owners
for
those
going
into
the
the
first
of
the
of
the
year.
So
there
are
a
couple
little
timelines
in
there
that
were
just
suggested,
and
that
was
that
we
could
have
an
enhancement
proposal
process
in
q1.
F
In
that
we
would
do
the
initial
level
level
of
effort
analysis
on
splitting
workloads
from
control
plane
in
q1,
with
an
eye
to
execute
on
it
in
Q2.
So
we're
looking
for
some
owners
for
both
of
those
pieces.
Anyone
who
might
be
able
to
help
out.
A
Thanks
for
bringing
that
up,
Paul,
so
yes
plus
one,
we
would
love
folks
to
volunteer
to
help
out.
So
you
know
how
to
reach
out
if
you're
interested
we're
on
slack
we're,
we
have
the
kcp
dev
Google
group,
there's
GitHub
discussions
and
issues
so
any
any
way
that
any
one
of
those
is
valid
for
for
reaching
out.
C
A
There
any
to
become
a
reviewer,
that's
a
good
question,
so
anybody
can
review
anything
the
we
follow
the
same
two-step
process
that
you
may
be
familiar
with
in
kubernetes,
where
anybody
who's
an
organization
member
in
GitHub
can
lgtm
and
only
certain
people
can
approve,
and
you
need
both
an
lgtm
and
an
approval
to
have
a
PR
merch.
A
So
if
you
are
interested
in
having
more
reviews,
sent
your
way
for
a
certain
package
or
portion
of
the
repository,
certainly
start,
you
know
looking
at
PRS
that
touch
those
areas
and
reviewing
them
and
then
at
some
point
you
are
more
than
welcome
to
open
up
a
PR
to
add
yourself
as
a
reviewer
to
any
of
the
owner's
files
and
if
there's
an
owner's
file
in
a
directory
that
or
if
there's
a
directory
that
doesn't
have
an
owner's
file
and
you
would
like
to
be
considered
a
reviewer
there,
you're
more
than
welcome
to
create
an
owner's
file
in
that
directory
and
add
yourself
as
a
reviewer.
A
Know
yeah:
when
you
talk
about
the
owners,
but
oh
yeah,
we
do
have
those.
So
if,
if
this
text
is
not
sufficient,
let's
maybe
chat
offline
and
see
if
we
need
to
make
any
changes
here,
but
we
strongly
welcome
and
encourage
anybody
to
review
everything.
A
It's
a
good
way
to
get
familiar
with
different
areas
of
code
and
the
more
you
review,
the
more
you
contribute.
Eventually,
you
can
become
an
approver.
A
Okay
right
so
again
just
call
for
help
if
anybody
is
interested
and
anything
else
on
this
topic
at
Paul.
F
I
guess
the
only
other
thing
I'll
mention
is
that
we're
not
looking
for
someone
to
create
something
from
scratch.
There's
many
enhancement
processes
that
you
guys
have
already
used
guys
and
gals
have
already
used
so
using
some
of
that
prior
art
to
Beach
campus
I
think
is
what
we
would
be
looking
for
on
that.
One
I
just
suspect
that,
folks
that
work
in
the
workloads
project,
the
most
are
going
to
be
the
folks
that
are
most
able
to
to
work
on
the
split
of
it.
F
A
And
we
do
I
I
took
the
minimal
step
of
creating
put
it
in
here
somewhere
I
did
create
the
enhancements
repository,
which
is
basically
blank.
A
It
has
the
boilerplate
templates
in
here,
but
the
remaining
work
for
moving
this
file
from
the
kcp
repo
to
the
enhancements,
repo
and
and
the
other
things
in
here
and
have
not
been
done
yet.
So
that's
what
we're
looking
for
help
on
all
right
I
did
want
to
welcome
a
new
member
to
the
community.
We've
got
Vince
who
started
yesterday
on
the
team
here
at
red
hat,
so
Vince
welcome
happy
to
have
you
on
board
and
Vince
comes
to
us
from
working
on
cluster
API
and
other
community
efforts.
A
So
if
you
see
him
opening
PRS
and
getting
involved,
we're
super
happy
to
have
him
here.
So
please
say
hi
when
you
get
a
chance
and
Vince.
If
you
wanted
to
say
anything,
feel
free,
don't
want
to
put
you
on
the
spot,
though
so.
D
Oh
yeah
I
mean
I'm
super
excited
to
be
here.
This
project
has
been
on
my
radar
for
a
while
and
I
can
say
that,
like
from
the
outside,
like
a
like
claustrophic
Community
was
super
excited
to
just
like
look
into
kcp,
so
I'm
happy
to
contribute,
and
if
you
have
any
tips
or
things
to
get
started
like
feel
free
to
reach
out.
A
D
A
Any
other
comments
on
the
enhancements
before
we
go
into
other
topics
or
look
at
the
issue.
Triage.
A
Nope,
okay,
so
I
think
I
opened
a
board
there
we
go
so
we
have
just
three
new
issues
that
have
come
in
in
the
past
week,
so
the
first
one
was
about
creating
too
much
stuff
in
our
Universal
workspace,
so
backlog,
I'm,
guessing
on
this
one,
Steve,
yep,
okay
and
and
the
ddsif
manually,
managing
informers
I
think
can
also
be
backlog
unless
you
want
to-
or
you
started
working
on
this,
but
it
didn't
pass
because
of
the
indexing
right.
A
C
C
A
B
If
I
remember
correctly,
I
might
be
wrong
because
it's
quite
long
ago,
but
selling
things
at
to
Jason
was
also
related
to
wasn't.
It
related
to.
C
B
They
not
everything,
I
mean
all
the
all
the
standard
objects
like
deployments
and
the
others.
They
are
expected
on
the
client
side
to
be
native
types
in
Cube
in
cubectl
for
example,
though,
in
fact
they
are
served
from
crds
and
I'm,
not
sure
if
it's
not
related
to
that,
but.
A
Yes,
but
anyways
I
think
we
should
try
setting
accept
content
types
to
Proto
and
Json
and
see
what
happens
because
it
should
improve
performance
at.
B
Yeah
anyway,
it
was
related
to
some
limitations
on
the
very
early
prototype
of
kcp.
So
probably
possibly
this
changed
now.
A
A
Alrighty,
so
that's
all
the
new
issues
for
triage
last
call
for
topics.
D
F
I
had
a
thought
that
it
might
be
fun
to
end
the
year
on
next
Community
call
with
maybe
just
an
issue
of
statement.
The
same
things
like
in
2023,
kcp
should
or
we'd
be
crazy.
If
we
don't
type
of
thing
to
help
us
kind
of
see
where
people
are
thinking
about
kcp
and
it's
future
of
the
next
year,
where
they
would
advocate
for
change
and
have
something
in
2023
to
actually
look
back
on
and
say.
Where
did
we
go
curious
if
folks,
like
that,
idea
or
think
it's
silly.
A
I
think
that
sounds
cool.
Maybe
we
start
a
discussion
in
GitHub
now
and
let
folks
fill
it
in
over
the
next
week
and
then
we
can
take
a
look
at
it
at
the
next
meeting.
F
A
And
speaking
of
end
of
the
year,
the
next
next
week
would
be
the
13th,
and
then
the
following
would
be
the
20th
which
is
creeping
up
on
the
holidays.
Do
we
want
to
just
say
that
next
week
is
our
last
meeting
of
the
year
or
do
we
want
to
try
and
do
one
on
the
20th
as
well
foreign.
A
All
right,
well
not
hearing
anything
else,
have
a
great
rest
of
your
Tuesday
everybody
and
see
you
in
a
week.
Yeah
take
care,
see.