►
From YouTube: Community Meeting, February 7, 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
Hello
and
welcome
today
is
February
7th.
This
is
the
kcp
community
meeting.
If
this
is
your
first
time
joining
welcome,
we
have
a
community
meeting
agenda
issue
in
GitHub,
which
is
what
I'm
sharing
on
the
screen
right
now.
Please
feel
free
to
come
in
and
add
comments
for
additional
meeting
agenda
items
if
you're
interested
and
we
do
try
to
use
Ray's
hand
in
Google
meets
so
while
we're
going
through
things.
A
If
you
have
something
you'd
like
to
say,
please
hit
the
raise
hand,
button
and
I
will
moderate
today,
so
I
suppose,
given
that
I
just
created
this,
oh,
we
already
got
some
stuff
so
awesome
to
see
that
sergius
I
will
turn
it
over
to
you
for
your
first
item
in
the
agenda
today.
B
B
Can
you
see
my
screen
looks
good
okay,
so
I've
been
experimenting
a
little
bit
not
too
much,
but
just
a
little
to
have
experimented
with
the
idea
how
we
can
get
kcp
a
little
bit
easier
installed
for
people,
so
can
they
can
just
try
it
out?
I
mean
trying
it
out
is
pretty
a
low-hanging
fruit.
B
B
There
is
work
on
going
to
reactivate
the
home
charts
and
get
them
going.
I
think
this
is
a
nose
Foundation
we
can
piggy
pack
on.
So
my
thinking
was
what
we
can
do
like
a
little
bit
of
a
step
two
and
bring
kcp.
B
Also
in
like
other
channels,
not
only
you
know
by
cloning,
the
repository
of
by
installing
huntroids,
but
obviously
like
at
least
for
openshift,
the
most
lowest
hanging
fruit
is
just
use
olm
and
if
you
have
like
the
console
here,
just
you
know,
go
operator
Hub,
just
search
for
kcp
like
here
in
openshift
and
just
have
it
installed,
but
I
have
no
logo.
Yet,
for
my
for
my
demo,
okay,
that's
fine,
and
that
is
sort
of
like
the
the
picture
of
The
strawman
Proposal,
to
have
an
installation
method.
Why
olm?
B
The
goal
is
to
avoid
redundancy,
so
I
would
like
really
really,
for
the
sake
of
you,
know,
reducing
the
maintainership
burden,
half
as
much
as
additional
manifest
as
possible.
So
the
idea
was
to
reuse
the
existing
column,
charts
that
we
are
currently
reactivating
and
the
operator
SDK
offers
a
method.
They
call
it.
The
helm
operator
to
yeah
create
a
bundle,
an
olm
bundle
which,
essentially
you
know,
bundles
Helm
and
executes
things
based
on
existing
Helm
charts.
So
the
experience
would
look
like
this.
B
You
install
an
operator
a
kcp
operator
and
that
kcp
operator
would
act
on
the
kcpc
or
these
and
that
kcp
or
CID,
if
you
see
here,
is
essentially
nothing
more
nothing
less
than
an
instance
of
Helm
variables
that
you
would
usually
create
on
your
command
line
to
create
a
hammer
installation.
So
every
held
variable
that
we
are
offering
in
our
home
charts
can
be
denoted
here
in
in
the
CRT.
C
B
Precisely
yeah
the
the
basic
idea
would
be
here
to
create
an
additional
deployment
in
the
handshots
themselves
right
and
any
parameterization
would
go
via
this
crd.
That
is
done
here.
So
essentially,
if
we
can
bring
this
additional
controller
in
the
hand
charts,
then
you
know
this
will
just
be
proxy
to
the
online
installation.
B
Just
a
little
bit
of
a
high
level
overview
because,
like
this
whole
ol,
everything
is
completely
new.
To
me,
to
be
honest,
I've
been
usually
sitting
in.
You
know
a
little
bit
lower
level
systems
so
far,
essentially,
how
it
is
being
done
is
there
is
a
Helm
operator.
B
That's
that's
part
of
the
so-called
operator
SDK,
it's
a
little
component
that
reconciles
a
crd
that
you
provide
as
part
of
your
olm
bundle
and,
as
you
see
here
for
the
ones
that
you
know
the
operator,
SDK
generator
creates
something
that
has
preserved
unknown
Fields,
true,
which
is
not
nice.
So
essentially,
what
we
could
obviously
do
in
the
Cod
is
to
spell
out
all
the
Helm
variables
that
that
we
offer
in
the
installation.
B
And,
based
on
the
crd,
the
hand
operator
reconciles,
alongside
with
the
existing
Helm
chart,
that
we
provide
yeah,
it's
the
current
the
pull
request
here.
Okay,
you
did
F1
charts
reconsides
using
the
existing
Helm
charts,
a
yeah
help
operator
deployment
and
that
one
creates,
you
know
all
the
necessary
bits
that
are
necessary
for
kcp
to
run.
What
is
nice
about
this
thing?
B
As
I
just
mentioned,
you
can
override
and
you
essentially
also
need
to
create
a
kcp
custom
resource
before
kcp
itself
is
being
deployed,
so
you
can
tweak
your
settings
and
have
things
like
TMC
turned
on
or
like
I,
don't
know,
and
a
route
created
or
not,
but
the
end
result
looks
a
little
bit
like
this
that,
at
the
end
of
the
day,
what
we
have
in
the
home
charts
now
the
title
operator
would
create
some
hcd
instances
would
create
the
kcp
from
proxy
and
what
other
little
control
components
we
need
to
do:
additional
reconciliation
and
additional
initialization.
B
What
do
we
need
to
do?
Well,
we
need
to
create
a
new
repository.
Unfortunately,
you
know
I
just
created
one
here
in
my
under
my
name
kcp-olm
and
essentially
what
this
thing
is.
It's
you
know
like
a
little
bit
of
a
boilerplate
around
the
helm.
Chart
currently
I
merge
the
helm
chart
in
three
of
this
bundle.
This
could
be
out
of
three,
which
is
because
we
have
a
working
version
and
that
bundle
would
essentially
create
contain
a
little
bit
of
metadata
would
create
crd.
B
It
would
contain
the
Cod
declaration
that
I
just
mentioned.
That's
the
yeah,
that's
the
custom,
research
definition
and
except
of
version
bumps
famous
last
words,
I,
don't
see
much
additional
maintainership
being
done
here.
Most
of
the
work
is
being
done
and
they
have
Choice
themselves
but
yeah.
If
we
want
to
go
this
route
with
offering
Hunters
via
olm
VIA
the
operator
to
have
a
nicer
online
installation
story,
we
do
need
a
little
bit
on
the
project,
so
that
would
be
on
us
to
create
and
a
little
bit
of
maintainership
in
terms
of.
B
Bombs
and
dimitro,
for
instance,
I
guess:
yes,
there
will
be
it
from
my
site.
Any
questions
Frederick.
E
Yeah
I
just
wanted
to
say
that
and
it's
the
end
of
Facebook,
maybe
to
increase
the
visibility
of
kcp
for
a
larger
community.
So
in
this
case
openshift
we
we
can't
publish
a
few
of.
D
E
Bundle
in
the
openshift
Capital
Community
catalog,
and
we
could
also
consider
that
there
is
a
artifact
at
dot
IO,
where
we
could
also
publish
the
M
chart
there
and
do
you
think
that
would
make
sense
and
other
obviously
is
not
production
development,
but
just
to
get
people
having
Hands-On
kcp
and
trying
it
in
three
minutes.
B
Yes,
precisely
I
would
say
so
this
is
you
know
like
like.
This
is
something
that
we
should
discuss
like
how
we,
how
do
we
frame
this
on
our
installation?
Right?
Obviously,
it's
for
me,
the
the
motivation
was
just
have
something
to
give
people
at
hand.
So
they
can.
You
know,
try
things
out
quickly
for
workspaces.
B
Another
use
case,
which
I
think
Stefan
mentioned
would
be
awesome
to
have,
is
the
TMC
stuff
and
the
singer
stuff
that,
because
that's
going
to
be
a
pretty
common
use
case,
so
yeah
so
I
would
say
as
a
first
step.
Just
have
something
like
a
hell,
hello,
world,
kcp
deployment
with
some.
You
know
additional
nice
batteries
and
then
later
on,
we
can
see
what
what
goes
from
here.
F
By
the
way,
maybe
it
could
be
also
an
option
to
have
you
know,
after
your
main
kcp
crd,
to
also
have
a
thinker
or
think
target
crd.
You
know
something
that
mainly
just
drives
you
to.
You
know
a
secondary
crd
that
that
allows
adding
string
targets
inside
the
The
Host
server
or
something
like
that.
I
mean
it
could
be
done
step
by
step
as
well.
B
B
Well,
he
wanted
to
contribute
and
I
would
say
once
we
have
something
that
you
know
just
works
trademark,
then
I
would
say:
let's
transfer
it
and
the
other
prerequisite
I
would
say,
is
shortest
PR
in
the
entrance
repo,
because
yeah
I
need
that
one,
and
as
soon
as
you
know,
things
converged
against
stability
a
little
bit
in
the
next
days,
I'm
totally
fine
with
creating.
A
A
G
Sure,
hopefully,
they're
minor
first
one
is
just
to
update
I
wanted
to
understand
where
we
are
and
we're
still
expecting
to
go
in
terms
of
versioning
the
documentation
I
see.
Now,
if
you
go
to
the
kcp.io
documentation
site,
the
releases
menu
now
has
two
entries,
Main
and
V
0.10.0
technically
main,
isn't
to
release
but
I.
Suppose
that's
not
much
of
a
big
deal.
G
If
I
pick
the
V
0
to
10.0.
However,
it
tells
me
there's
no
content.
A
Yeah
still
on
my
to-do
list
and
I
know:
avanal
have
been
helping
out
as
well.
So
much
appreciated
there.
H
A
Got
sidetracked
by
a
pretty
nasty
watch,
cache
storage
bug,
so
I
will
get
back
to
that
as
soon
as
I
can.
G
G
G
Keep
going
all
right,
then.
My
other
item
was
about
Ides
and
go
workspaces.
I'm
kind
of
curious
as
to
what
other
people
are
doing.
I
tried
to
do
something
that
I
thought
would
be
fairly
obvious
last
week,
but
it
ran
into
several
problems,
so
I
drew
the
conclusion
that
people
aren't
doing
it
so
I'm
kind
of
curious.
What
I
tried
to
do
was
make
a
good
workspace
that
let
me
look
at
both
kcp
and
kcp's
Fork
of
kubernetes
and
Edge
MC.
Also
I.
Think
the
edgemc
wasn't
really
the
problem.
G
The
problems
are
coming
from
kcp
and
it's
Fork
of
kubernetes.
There
are
a
bunch
of
things
that
my
IDE
complained
about:
I've
been
using
visual
studio
code
and
I
tried
to
point.
You
know,
create
that
go
workspace,
Point
Visual
Studio
code
edit
and
had
several
problems.
So
my
question
is:
what
are
people?
What
do
other
people
do
you
know
and
and
kind
of
more
forward?
Looking,
you
know
what
should
be
supported
and
what
should
we
expect
to
work.
A
So
I
can
tell
you
my
workflow
and
setup,
so
I
tend
to
use
so
I
use,
go
land
and
I
have
one
project
that
is
all
the
kcp
repositories
and
if
I
need
to
be
working
on
kubernetes
as
well,
I
have
that
in
a
different
project
and
if
I'm
trying
to
run
kcp
on
my
laptop
with
changes
to
kubernetes,
then
I
run
a
subscript
to
update
my
go
mod
so
that
all
of
the
replace
statements
that
you
see
in.
H
A
Way
so
that
all
the
replace
statements
that
are
down
here
all
the
way
at
the
bottom
for
all
the
kids.io
are
just
relative
links
to
to
files
on
my
laptop
and
that
tends
to
work
pretty
well.
For
me,
I
I
tried
going
to
go
workspace
route,
but
you
have
to
issue
replace
statements
for
for
all
of
these,
in
the
workspace
or
in
the
go.work
file.
F
Sorry
I'm
using
vs
code,
and
you
have
a
just
an
example-
maybe
in
kcp
contrib
kcp.code
workspace
mainly
I,
have
two
folders.
F
So,
as
you
can
see
here,
there
is
the
three
back
so
mainly
at
the
right
place
in
your
go
source,
the
rights
relative
place,
and
once
you
have
that
it's
quite
easy
to
to
manage
that
with
the
you
know,
workspace
with
two
sub-projects
in
vs
code
and
then
to
switch
between
using
the
kubernetes.
As
you
know,
normal
Imports
or
all
the
kubernetes,
which
is
you
know
that
you
have
locally,
you
can
just
choose
also
this
script,
which
is
in
contrib
a
local
Dev,
local
dev.sh.
F
This
does
the
switch
and
mainly
changes
everything
in
your
go
mode.
To
use
this.
You
know
local
kubernetes
installer
at
the
right
place,
at
least
that
works
for
me.
Okay,.
G
Maybe
a
relevant
Factor
here,
and
this
may
be
a
bit
of
a
vs
code.
Foo
is
I
know
when
I
open
something
in
vs
code.
I
have
to
tell
it
whether
I
trust
it
or
not.
G
I've
been
telling
it
for
kubernetes
for
everything.
I,
don't
write
actually
for
everything
that
Imports
a
bunch
of
stuff
I
say
that
I,
don't
trust
it
because
I,
don't
trust
all
the
things
that
it
Imports
and
that
makes
it
kind
of
crippled
so
right
right.
The
thing
that
set
me
off
on
this
is
when
I'm
working
on
kcp
right,
the
def,
a
reference
to
something
in
in
its
Fork
of
kubernetes
I
can
follow
the
from
the
use
to
the
definite
follow
from
the
use
in
kcp
to
the
definition
in
the
kubernetes
fork.
G
So
first,
just
a
factual
question
here
when
I
tell
vs
code
did
I
trust.
Something
am
I,
am
I
not
telling
it
that
I
trust
everything
it
Imports.
A
Don't
know
I
I,
that's
a
good
question,
it's
probably
something
we
could
Google.
Let's
pause
on
that
for
just
a
second
I
know:
Daniel
has
his
hand
up
so
Daniel.
I
Yeah
I
was
just
gonna,
mention
kind
of
like
a
slightly
different
perspective
when
you're
trying
to
import
all
this
as
well.
If
you're,
trying
to
use
kcp
as
a
library
in
another
project
being
able
to
find
all
the
right
import
replaces
is,
is
pretty
challenging,
especially
when
you
already
have
a
lot
of
the
Upstream
dependencies.
I
You
know
I.
Think
one
area
where
this
particularly
came
up
recently
was
with
some
of
like
The
Logical
cluster
V2
to
logical
cluster,
V3,
stuff
and
just
kind
of
like
when
that
was
updated,
and
the
API
Machinery,
the
kcp
API
Machinery
repo,
the
kubernetes
fork
and
then
the
controller
runtime
Fork.
I
So
I
wonder
if
there
is
like
this
is
only
sort
of
somewhat
tangential
to
what
we're
talking
about
here.
But
if
there
is
like
a
a
canonical
place
where
we
could,
you
know
test
to
make
sure
that
third
parties
are
like
consuming
kcp
in
a
different
repository
if
they
can
point
to
it.
Maybe
it's
the
controller
runtime
example:
that's
not
actually
using
kcp
as
a
library
right.
I
It's
just
a
controller
that
runs
against
kcp,
but
if
that
is
something
that
other
folks
are
trying
to
do
as
well,
I'm
happy
to
contribute
some
of
my
my
pain
to
make
a
repo
that
makes
that
a
little
easier
yeah.
It's
definitely.
A
A
goal
we're
just
we're
not
there
yet
we're
on
the
road
through
it
yeah.
C
Make
sure
Library
usage,
this
pattern
keeps
working
I,
think
it's
helpful,
so
either
inside
of
kcpe,
but
then
the
goal
mod
problem
is
not
solved
right,
maybe
even
a
different
repository
in
inside
of
kcp
Def,
which
renders
kcp
and
that
we
can
get
CI
somehow
or
tests
that
anything
of
those
along
those
lines
would
be
what
I
think.
A
Maybe
it
would
be
worthwhile
to
try
and
find
some
other
folks
who
are
using
vs
code
and
working
on
kcp
and
kubernetes
together
and
just
kind
of
have
a
little
hacking
session
to
like
you
could
screen
share
and
show
what
you're
trying
to
do
and
folks
could
try
and
help
out.
Maybe
yeah.
H
A
So
I
will
leave
that
to
you
to
try
and
see
some
folks
and
get
that
organized
I
know
David
used
to
use
vs
code,
maybe
posting
slack
and
see
if
you
can
get
some
other
people
to
chime
in
and
then
you
all
can
self-organize
very
good.
Thank
you
awesome.
A
Any
other
thoughts
on
this
topic.
Before
we
move
on
to
the
next
one.
J
The
some
of
us
met
last
week
and
we
talked
about
workspace
in
Easy
Edition,
and
then
we
came
up
with
this,
drawing
that
Stefan
Drew,
basically
I'm
not
going
to
try
to
explain
this.
Drawing
with
Stephen
wants
to
do
it
feel
free.
But
then
the
next
step
is
to
turn
this
into
like
a
proposal
right,
so
I
need
to
digest
it
and
and.
A
Can
you
give
some
background
on
what
you're
trying
to
do
here?
Yeah.
J
So
probably
I
should
open
an
issue
like
an
epic
that
was
the
scenario
I
shared
on
slack.
Maybe
you
have
it
in
the
previous
community
meeting.
J
That
was
the
scenario
where
you
have
a
workspace
exporting
a
pi
like
Kafka
API
and
then,
when
creating
or
binding
to
creating
a
new
workspace
binding
to
the
to
the
Kafka
workspace.
Then
the
the
service
provider
wants
to
create
a
secret
in
the
in
the
user.
Workspace.
That's.
A
Okay,
yeah
I
definitely
think
an
issue
of
some
sort
is
worth
having
Stefan
was
there
anything
you
wanted
to
say
about
this
now
or
just
to
hold
off
until
later.
A
Okay,
so
yeah,
if
you
can
create
a
GitHub
issue
for
your
use
case,
I
think
that
would
be
a
good
starting
point.
A
I,
don't
see
any
other
things
in
here.
Anybody
have
any
additional
topics.
A
If
not
I'm
going
to
take
a
look
at
our
new
issues
that
need
to
be
looked
at
all
right,
so
the
first
one
here
is.
D
Can
explain
if
required
so
when,
when
we
try
to
run
the
workload
sync
command
using
hyphen
hyphen
resources
with
just
the
resource
name
like,
for
example,
if
you
take
Route,
if
you
just
do
a
hyphen,
hyphen
resources
and
then
route
and
try
to
sync
that
resource
to
the
physical
cluster,
we
always
get
error,
saying
that
there
is
a
sar
check
that
is
being
failed.
Instead,
if
we
use
the
group
and
version
that's
this
route
dot
routes.rout.openshift.io,
then
everything
works
fine
and
we
don't
see
any
such
error.
F
My
understanding
of
this
is
that,
obviously
in
during
the
API
import
mechanism,
we
are
not
strict
enough,
in
fact-
and
you
know,
as
soon
as
it
finds
a
Javier
that
seems
to
correspond
to
what
is
requested
either
by
singular
name
or
plural
name,
then
it
takes
its.
But
the
problem
is
that
now
we
we
require
additional.
You
know
permissions
Downstream
and
those
permission
then
are
put
on
the
wrong
name.
A
Okay,
I
mean
we
do
have
this
issue.
That's
you
know
history
file
as
well.
That
I
remember
talking
to
her
about
this
back.
F
A
This
may
be
old,
but
so
David
I'm
gonna,
assign
this
to
you
and.
A
If
you'd
like
and
then
I'm
also.
A
A
A
feature
request
to
be
able
to
set
this
in
the
past,
but
if
there's
some
way
that
you
know
this
is
somebody
else
asking
for
it.
So
maybe
we
do
need
to
do
it.
We
just
want
to
make
sure
that
we
do
it
in
a
way
that
you
can't
really
break
things
or
if
you
do
break
things,
you
know
what
you're
doing
so
David
do
you
have
any
thoughts
on.
F
Yeah,
it
seems
I
mean
to
me
it
seems
a
bit
like
you
know
the
thing.
The
features
requested
for
age,
for
example.
So
really
it's
very
specific
use
cases
where
we
would
like
to
disable
some
of
the
some
other
things
and
standard
limitations
that
that
we
are
doing
in
the
Sinker
for
obviously
for
security
purposes,
and
then
we
want
to
disable
such
things.
F
For
you
know
here,
you
know
enable
to
be
able
to
to
you
know,
accommodate
a
specific
namespace
for
tests,
so
I'm
not
very
fond
of
this,
at
least
as
a
normal
feature,
maybe
having
a
sort
of
you
know,
dedicated
way
to
start
the
Sinker
or
with
an
option,
or
you
know,
even
even
have
a
dedicated
binary
for
things
like
that.
I
don't
know,
but
once
again,
I'd
rather
keep
this
as
a
really
special
case.
Instead
of
adding
this
feature
in
the
normal
use,
cases.
K
A
K
I
was
trying
to
perform
it's
just
on
kcp
like
how
resources
are
getting
synced
using
kcp,
and
you
know
how
usually
an
application
runs
on
a
self-managed
cluster.
So
in
that
process,
I
tried
to
sing
some
deployments
from
kcp
workspace
to
myself
manage
cluster
and
during
the
process
I
noticed.
Like
you
know,
namespace
was
issued
with
our
ansible
operator.
So
later
we
decided
not
to
you
know,
run
those
benchmarks
and
try
out
something
different
I
mean
try
some
different
kind
of
tests
to
do
that.
So
that's
not
a
big
period.
A
A
Next,
oh,
this
is
a
docs
thing.
H
A
This
is
an
interesting
one:
oops
I
lost
it.
Resources
sync
to
the
workload
cluster,
are
deleted
and
recreated
when
computer
enters
sleep
mode.
David
I'm
gonna
give
this
one
to
you
to
take
a
look
at
well.
F
But
it
you
know,
it
might
be
that
that
it's
some
somehow
expected
at
least
the
way
it's
currently
implemented,
because
you
know
when
you
shut
your
when
you
close
your
laptop
and
it
starts
sleeping
the
times.
The
time
still
goes
flies.
A
F
And
so
you
know
when
you
open
that
back
the
heartbeat
obviously
has
not
been
updated.
So
that's
you
know
we
have
to
think
effectively.
If
it's
a
bug
or
not
I,
don't
know
yeah.
A
L
L
A
high
check
has
fail,
I
think
when
we
start
TCP
just
starting
kcp,
we
Mark
everything
as
fail
if
they
are
past
the
time
that
they
should
be
so,
we
should
wait
for
the
thinkers
to
be
able
to
reconnect
to
kcp
and
and
update
the
hair
bit.
Something
like
you
know
it's
something
that
we
can
Implement
like
having
a
10
second
delay
or
20
second
delay,
depending
on
the
here
with
time.
F
Yes,
if
there
would
be
something
to
fix,
it
might
more
be
at
the
at
the
restart
of
of
kcp,
for
example,
but
there
you
know
when
you
yeah.
F
That's
a
bit,
you
know
quite
a
strange
situation
and
not
sure
you
know
initially
in
in
in
the
initial
issue
from
from
Polo,
there
was
also
obviously
the
feeling
that
it
was
failing
when
there
is
a
disconnection,
but
in
fact
it
was,
you
know
a
wrong
failure.
It
was
just
because
the
connection
was
only
on
one
side
on
the
Sinker
and
not
on
the
kcp
side.
You
know
just
due
to
the
way
the
connection
was
was
down
on
the
network.
F
So
in
fact,
in
fact,
the
the
error
with
the
connection
is,
is
not
there.
It's
it
works
properly.
So
the
only
thing
that
that
is
still
is
still
not
working
is,
is
the
sleep
mode,
but
I'm
not
sure
we.
We
should
consider
that,
as
you
know,
a
book,
or
at
least
a
critical
one,.
A
F
A
Next
up
is
a
flake.
This
is
yeah.
A
If
y'all
need
help
with
this
one,
let
me
know
I
I
wasn't
sure
if
I
should
try
to
deflake
Sinker
tests
or
wait
until
you
have
some
more
PR's
merge.
So
what's
the.
I
F
A
One
I
think
I
fixed
this
one
I'll
double
check.
A
Give
this
one
to
you
all
to
triage.
Have
you
seen
this
one
before.
F
Yes,
maybe
I
don't
know
if
casterly
wants
to
explain.
I
can
do
it
in
fact,
yeah
sorry
go
in.
If
you
want.
D
Yeah,
when
trying
to
think
anything
related
resources
into
the
physical
cluster
I
see
that
the
physical
cluster
is
an
openshift
cluster.
Where
I
am
trying,
where
I've
installed.
K
native
and
trying
to
sync
some
resources
from
kcp
to.
K
D
Physical
cluster
I
see
that
openshift
is
creating
some
few
other
resources
other
than
the
k-native
Eventing
service,
which
gets
synced,
and
the
singer
thinks
that
the
resources
are
being
synced
from
the
kcp
and
it
tries
to
sync
it
back.
And
then
the
logs
are
flooded
with
error
messages
and
also
the
resources
that
are
being
coming
up
in
the
physical
cluster
has
the
annotations.
F
Yeah,
so
that
that
raised
an
interesting
question
of
you
know
when
you
have
some
controller
on
the
downstream
cluster
on
the
physical
cluster,
that
many
create
derives
objects
from
initial
objects
that
were
synced
from
Upstream,
typically
K
native
controller.
Here
it
many
creates
derived
objects
with
the
same
levels
and
annotations
as
the
main
object.
F
So
the
key
native
Service
will,
you
know,
give
its
levels
and
and
annotations
to
the
cube
service
and
Cube
deployment
and.
J
A
I
think
the
issue
here
is
that
a
a
controller
or
service
such
as
k-native
would
need
to
be
modified
to
work
against
kcp
properly.
I
would
imagine.
F
Yes,
I
think
sure
these
type
of
things
will
happen
when
we
don't
use
the
typical
nominal
way,
which
is
you
know,
have
the
higher
level
objects
at
the
kcp
level
and
only
think
the
typical
Cube
resources
to
kcp
to
the
physical
cluster.
But
on
the
other
hand,
it
seems
to
me
that
it
would
be
still
good
to
fix
that,
possibly
just
by
checking
if
the
object
has
an
owner
reference,
because
everything
syncs
directly
by
kcp
doesn't
have
any
owner
reference.
A
Right
well,
I've
got
this
in
the
backlog
for
you
all
I'm.
H
A
And
thank
you
for
filing
this
Dan.
H
I
Yeah
I'd
love
to
go
ahead
and
just
copy
over
what
I
have
I
I'm
up
for
discussion
around
like
what
we
want
the
example
to
do.
It
could
just
be
like
you
know
something
simple
like
using
kcp
as
a
library
and
like
a
post
art
host
that
just
adds
some
controller
to
watch
things.
I
mean
that
could
be
a
good
starting
place.
I
A
All
right,
thank
you
for
following
that,
I've
got
this
set
up,
and
that
is
all
for
triage.
Okay,
so
I
think
we've
come
to
the
agenda
the
end
of
our
agenda
for
today.
Thank
you,
everybody
for
joining
and
participating,
and
we
will
see
you
next
week
have
a
great
rest
of
your
week.