►
From YouTube: Community Meeting, September 13, 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
Questions:
hey
everybody
today
is
the
kcp
community
meeting
for
September
13th.
We
have
as
always
an
issue
in
GitHub
to
use
for
our
agenda.
If
you
would
like
to
add
something
to
the
agenda,
please
feel
free
to
add
it.
Unfortunately,
I,
apparently
just
denied
varsha
access,
because
I
hit
enter.
I
was
trying
to
send
this
message.
So
if
somebody
could
tell
varsha,
please
try
to
get
back
in
for
a
letter
in
so
the
first
item
on
the
agenda
comes
from
Stefan
so
over
to
you,
Stefan.
B
All
right,
so
user
docs
I
think
there
were
a
couple
of
discussions
recently
started
in
in
the
threat
about
stock
days.
Whether
we
want
to
continue
that
my
understanding
is
user.
Docs
is
much
more
and
we
should
talk
about
that
independently
from
block
days.
So
I
think
we
all
agree,
we
need
them.
We
don't
have
much.
We
have
the
only
thing
we
have
basically
are
godox
on
the
apis,
which
are
kind
of
huge
Networks.
B
But
of
course
that's
not
enough
and
the
question
is
yeah
how
to
start
in
which
model
like
users
be
maybe
in
Epic
in
next
release,
at
least
for
the
bigger
topics
and
krishnas.
Who
will
drive
that
whether
anybody
is
interested
and
which
kind
of
publishing
will
we
use?
There's
a
different
topic.
B
Paul
has
brought
up
we'll
talk
in
a
minute
about
so
we
we
got
the
kcpio
domain,
so
we
could
do
a
docs
kcpio,
two
of
them
by
markdown
files
similar
to
other
projects,
maybe
a
Google
doxy
site,
or
something
like
that
and
start
with
publishing,
API
docs.
So
a
couple
of
generators
out
there,
so
it
would
be
a
very
pretty
easy
one-day
two-day
task
to
get
that
going.
B
That
might
be
a
good
step
to
start
with,
but
of
course
they
have
plenty
of
topics
and
I
think
we
need
a
connection
of
those
where
somebody
has
to
start
writing
real
blocks
like
tutorials
or
any
kind
of
bigger
documentation,
which
is
not
a
reference.
That's
API
Docs
some
questions
about
opinions,
how
to
start
so.
My
suggestion
would
be
to
start
maybe
with
an
epic
and
have
people
contributing
in
the
next
release,
but
of
course,
I'm
happy
to
hear
other
opinions
so.
A
I
would
be
happy
to
drive
this
I
think
we
can
start
with
soliciting
topics
that
would
make
sense
to
be
in
our
documentation.
So
an
outline
essentially
and
I'll,
probably
sick,
of
hearing
me
talk
about
cluster
API,
but
here's
the
cluster
API
documentation.
This
is
published
on
netlify,
using
the
make
docs
tool
I,
don't
care
how
we
publish
it,
I,
don't
care
how
we
generate
it.
What's
more
important
is
the
content.
So
what
I
like
about
the
work
that
we
did
here?
A
Is
that
there's
a
relatively
brief
overview
and
then
the
outline
that
shows
up
next
is
how
do
you
get
started
and
there's
a
quick
start
and
I
tried
to
model
what
I
put
in
our
readme
in
the
pr
that
I
did.
We
need
to
fix
the
formatting
but
there's
an
overview
and
then
there's
a
quick
start
and
then
anything
beyond
the
quick
start.
A
Is
here's
all
the
next
things
you
could
look
at
so
the
other
thing
I'll
point
out
about
the
the
cert
manager
docs
is
when
you
get
Beyond
quick,
start,
Concepts
and
personas
there's
what
sort
of
tasks
can
I
do
and
I
think
that's
a
good
approach
so
like
setting
up
a
sinker,
that's
a
task,
creating
workspaces
and
managing
them.
Those
are
tasks,
so
we
can
work
as
a
community
to
try
and
put
together
an
outline
and
then
have
folks
assigned
to
working
on
the
individual
pieces.
C
Next,
just
raising
general
interest
to
contribute
I'm
fine,
with
whatever
who's
taking
the
lead.
I
generally
feel
earlier,
when
I
was
wandering
through
the
documentation
of
this
project,
parka
about
continuous
profiling
that
you
know,
whenever
I
land
on
kcp
I
would
love
to
understand
the
sort
of
all
the
general
architecture,
the
use
cases.
The
project
is
supposed
to
help
me
and
then
what
what
you
just
said
and
the
like
for
for
the
given
use
cases
have
like
a
little
tutorial
on
how
to
set
things
up
so
yeah.
C
Consider
my
driving
and
feel
free
to
assign
me
documentation
tasks
as
well.
I
think
you
know
we
have
a
good
start
with
the
markdowns
in
the
Ducks
folder,
but
like
this
should
be
continued
and
be
on
the
lending
page
yeah.
A
A
Thank
you,
Steve
you're.
Next.
D
A
D
A
C
Next
I'll
say:
I
love
that
idea,
because
that's
how
I
was
able
to
contribute
some
docs
without
being
an
expert
in
the
field.
I'd
also
want
to
mention
whether
or
not
we
want
to
have
this
as
part
of
our
definition
of
done
for
an
epic
and
the
other
thing
that
I
want
to
do
is
put
MJ
on
the
spot,
because
he
was
just
telling
me.
He
started
using
kcp
on
a
side
project
and
I
had
some
thoughts
on
Docs.
E
So
it's
a
sync
targets
and
things
like
that
which
I
get
a
point,
but
this
kind
of
was
challenging
to
get
the
relationship
between
like.
What's
if
I
don't
want
to
use,
let's
say
kcp
as
a
as
a
unified
control
plane,
but
just
simply
as
a
cluster
registry
to
be
able
to
access
multiple
clusters
through
the
single
entry
point,
which
could
be
one
of
the
use
cases.
E
A
Yeah
definitely
we
know
that
the
terminology
is
confusing.
Sync
Target
used
to
be
called
workload
cluster
for
what
it's
worth,
but
we
didn't
want
to
make
it
sound
like.
If
you
created
a
workload
cluster
resource,
you
would
have
a
cluster
that
was
usable.
So
that's
why
we
renamed
it
but
yeah,
please,
you
can
create
discussions
in
GitHub.
If
there's
a
place
where
you
want
to
chat
about
things,
I
would
encourage
folks
to
create
discussions.
A
We
haven't
really
highlighted
it
much,
but
if
you
go
to
the
project
there
is
a
discussions
tab
up
at
the
top
and
that's
a
good
place
to
chat
about
things
before
deciding
that
something
is
actionable,
and
then
we
can
make
it
an
issue.
A
Forward
yeah,
so
I
think
publishing
API
reference
stocks
is
actionable,
so
that's
something
where
we
can
file
an
issue
and
have
a
brief
discussion
about
what
tech
we
want
to
use
and
then
get
it
assigned
to
somebody
and
then
the
the
more
the
broader
topic
is
all
the
other
stuff.
So
a
minimal
top
down
left
nav
outline
approach
for
what
sort
of
content
we
want
to
have
and
then
starting
to
actually
write
that
content.
So
I
guess
Step
One
is
deciding
where
we're
going
to
host
it
and
how
and
with
what
like?
A
How
are
we
going
to
generate
the
content,
and
then
we
can
do
the
other
two
things:
the
reference
docs
and
the
the
actual
site
generation
yeah.
B
Concrete
level
process
should
be
starting
epic
and
put
subtasks
there
and
assign
people,
or
do
we
wait
until
the
next
Milestone?
What
would
we.
A
Do
there,
let
me
think
about
that:
a
little
bit
more
I
mean
I,
think
we
can
create
an
epic
but
I
wouldn't
expect
it
to
get
finished
for
0.9.
No,
no!
That's
what
I
supposed
to.
B
A
B
I
think
it's
Bose,
there's
a
technical
part,
and
but
the
outline
feature
I
think
it's
an
effort.
We
will
do
now.
It's
the
one-time,
epic
we
will
have.
Eventually
there
will
be.
The
discussions
at
Paulo.
I
think
was
mentioning
that.
How
is
this
part
of
definition
of
done?
Is
it
something
we
have
an
epic
and
there
was
a
discussion?
Maybe
it's
also
something
we
can
talk
about
now.
B
So
Upstream
has
this
stock
deadline
I'm
not
sure
when
it
is
exactly,
but
it's
before
the
release?
Basically,
but
after
code
freeze,
maybe
something
like
that
would
work
as
well.
D
A
D
B
A
That
works
for
me,
so
we
should
add
that
to
the
contributing
doc.
D
Ument
and
I'd
say
if
we
have
an
epic
I,
think
it
probably
makes
sense.
If
there's
a
bunch
of
mechanistic,
things
need
to
happen
to
actually
get
this
bootstrap,
but
then
like
once
it's
going.
It's
not
really
like
a
sensible
question
to
ask
like
how
are
docs
going.
This
is
you
know,
because
it's
kind
of
intertwined
with
the
work
in
general,
so.
D
A
Okay,
I
will
I
pick
an
action
item
to
get
the
contributing
doc
updated
to
talk
about
docs
and
feature
dates
and
whatnot,
so
I
think
we
can
asynchronously
organize
and
epic
and
or
tasks
to
get
this
started.
Does
that
sound,
fair
and
I'm
happy
to
start
that?
A
Okay
Paul?
You
have
the
next
topic.
C
Yeah
I
think
it
just
kind
of
dovetails
into
what
we
were
mentioning
before.
We've
got
the
kcpio
domain,
we're
going
to
make
some
decisions
on
how
and
what
we
want
to
dock
there.
So
it
probably
relies
on
that
technology
discussion.
So
just
general
awareness
and
if
you
have
opinions
on
things
that
worked
well
for
you
in
the
past.
There's
some
comments
on
here.
A
Cool
thanks,
Paul
Stefan,
TBD
Milestone
to
project
backlog,
yeah.
B
This
is
maybe
something
we
could
just
try
out
when
we
talk
about
the
new
issues,
so
some
people
in
the
project
try
to
work
with
projects
in
GitHub,
because
I
got
much
better
and
they
are
usable.
So
some
of
the
features
like
TMC
features
say
experimenting
with
that
we
realized.
So
what
we
have
now
that
every
new
issue
is
going
to
the
project,
like
the
sap
project
automatically
there's
a
GitHub
action
doing
that
and
we
have
a
way
now
to
yeah.
B
We
are
seeing
it
here
to
see
incoming
issues
in
the
project
and
what
is
missing
is
that
we
don't
use
this
TVD
Milestone,
but
we
have
to
use
something
which
is
compatible
to
to
projects
here
to
to
get
just
issues
which
are
not
some
subtask
of
an
epic
and
the
proposal
is,
if
you
go
to
some
issue,
maybe
you
click
on
one,
there's
a
state
of
an
issue
on
the
right.
If
you
click
on
where
this
no,
the
state
which
is
new
yeah,
yes,.
B
Basically,
the
status
new
and
there's
also
a
status
called
backlog
and
the
should
correspond
maybe
to
a
column
in
the
yeah
board.
We
can
make
that
so
we
don't
misuse
the
Milestones
anymore,
but
use
this.
A
Okay,
yeah
I,
like
that
idea,
so
we'll
figure
out
the
Mechanics
for
the
best
way
to
do
this.
I,
don't
think
we
need
to
take
up
everybody's
time
to
you
know,
show
me
fumbling
through
GitHub
projects,
but
yeah
I,
like
that.
D
But
super
quickly,
though,
if
you
go
back
to
that
issue,
that
you
had
open,
there's
a
plus
one
more
on
the
right
side,
and
so
that's
where
we've.
D
And
so
then,
at
the
top
of
this
you
see
the
views
potentially
that
view
or
that
view.
If
you
click
on
those,
then
you
can
see
progress
inside
of
it,
and
you
can
also
see
obviously
progress
across
the
whole
project
yeah,
which
is
precisely
what
is
in
TMC,
for
example,
the
next
tab.
Now
it's
the
the
project
is
filtering
issues.
D
C
A
C
C
Can
I
ask
if
we
had
a
take
away
from
that
for
the
next
meeting,
and
can
we
kind
of
demonstrate
here's
how
we
would
use
those
views
to
talk
in
this
meeting
or
other
meetings
or
what
our
expectations
would
be
for
a
contributor
versus
an
approver
or
a
viewer
or
how
it
lines
up
to
the
roles
and
what
they
would
use
it
for
yeah.
B
B
It's
just
a
different
way
to
this,
and
it
allows
other
teams
like
Focus
teams,
to
use
projects.
That's
all
I
think.
C
A
A
So
I
guess
we
don't
really
need
to
stick
these
in
the
TVD
Milestone
anymore,
but
we
can
at
least
go
through
wow.
We
have
a
lot
go
through
the
28
here.
If
we
want
or
not
what
do.
B
A
Yes,
so
I
still
want
to
do
proud,
post,
submit
jobs
that
run
against
Main
and
release
branches
because
something
may
merge
or
the
pull
request
test
may
pass
and
then
it
merges
and
I
don't
know
something's
weird.
So
unless
there's
objections,
I'm
going
to
put
this
in
the
backlog.
A
I
mean
this
is
more
about
user
experience,
so
I
think
at
the
very
least.
We
need
to
address
each
item
here
with
a
comment.
So
I'll
put
this
in
the
backlog.
A
So
one
request
that
I
have
is
that,
if
there's
a
request
for
multiple
things
that
we
make
each
thing
its
own
issue
instead
of
having
one
issue,
that's
got
four
different
things
in
it.
So
let's
David
can
I
give
this
to
you
to
basically
close
and
replace
with
any
work
that
we
do
want
to
proceed
with.
D
Why
not
I
I
would
be
you
know
on
PTO
tomorrow
evening,
so
so
okay,
never
mind
then,
but
if
I'm
not
mistaken,
yeah
there
are
part
of
them
are
more
core
related
and
part
of
them
are
more
TMC.
I
think
you
know
I.
A
D
A
I'm
going
to
leave
it
in
here
all
right
this
one,
so
this
was
a
request
that
when
the
Sinker
deployment
and
its
related
resources
are
deleted,
the
requester
was
asking
that
all
of
the
synced
content
on
the
physical
cluster
get
deleted.
A
A
This
was
a
proposal
to
donate
a
shell
command
that
can
install
start.
Stop
install
the
Sinker
and
clean
up
stuff
and
I.
Put
a
comment
in
here
about
I.
Think
here's
my
development
process,
but
maybe
we
start
with
some
make
file
targets
to
start
and
stop
the
kcp
server.
If
wanted,
but
I,
don't
know
I'm,
not
particularly
interested
in
trying
to
maintain
a
shell
script
that
could
do.
B
There's
a
there's
a
root
of
some
which
I
try
to
apply
a
shade
script,
which
well
could
be
something
where
you
can
copy
out
the
commands
and
execute
them
manually
in
the
moment.
It's
about
hiding
complexity
and
I
think
this
is
about
that
yeah.
It's
over
the
line,
basically.
B
And
we
try
to,
we
try
to
be
pretty
strict
about
that.
Once
the
test
servers,
yeah
I
mean
I,
won't
script
too
much
yeah.
A
Yeah
I
mean
basically
Iran
test
server
I
hit
Ctrl
C
to
stop
it.
I
manually
run
RM.
If
I
want
to
delete
stuff,
I
use,
make
install
to
install
things
and
I
use
our
commands
to
get
a
sinker
installed.
So
yeah
I
don't
know.
Do
we
want
to
just
close
this,
or
do
you
want
to
maybe
do
simple
make
targets
for
starting
and
stopping
the
test
server.
B
B
What
we
have
we
start
test
servers,
even
the
shorted
one
and
so
on.
I,
don't
think
we
have
a
way
to
not
start
tests.
Basically,
this
is
missing
if
you
split
it
up
in
two
steps.
Maybe
we
have
what
is
possible
here.
A
Claiming
permissions
on
a
sar,
subject,
rules
or
for
class
fails.
I
know,
sergius
is
looking
in
this,
but
he
had
to
go.
A
Oh
wait.
No,
this
is
different.
This
is,
can
we
just
expose
SAR
and
subject
rules
review
to
all
virtual
workspaces?
What
does
it
mean?
It
means
that
the
controller
that's
going
through
a
virtual
workspace
can
issue
a
sar
and
get
a
response.
I
think
it's
reasonable.
What
what
does
it
mean
s.
A
B
B
A
B
A
A
So
backlog
PR
title
indicators,
I'm
working
on
don't.
A
Well,
I
put
mine
in
process.
I'll
go
look
at
the
at
that
one.
A
A
I
filed
this
one,
so
we
have
in
our
Cube
control
plugin
when
you
try
to
create
a
workspace.
If
you
get
a
not-found
response,
there's
a
comment
in
there
that
says
there
may
be
some
cash
slowness
and
so
or
something
along
those
lines,
and
this
is
the
error
message
that
we
return.
We
need
to
revisit
the
error
message,
because
this
is
not
accurate
anymore.
D
A
Clock
we
go
thank
you
for
letting
Help
Wanted
on
there,
although
I
think
we
need
some
direction
for
what
we
say
before.
Just
having
somebody
work
on
it.
A
Oh
yeah
dead,
so
I
don't
know
if
this
is
more
prevalent,
but
when
I
was
trying
to
debug
some
deletion
issues,
the
kcp
API
binding
controller
string
is
used
as
the
user
user
agent
for
three
or
four
or
five
different
controllers
and
I
think
we
probably
wanna
in
a
unique
user
agent
for
each
controller
when
it's
making
client
requests
so
I
filed
this.
For
that
one.
A
Oh,
this
isn't
good
what
was
this
Steve?
Oh,
this
was
yeah
okay
yeah.
So
this
was
fun
where,
if
you
manually
do
a
find
and
replace
against
a
crd,
that's
in
our
repo
and
the
corresponding
generated
API
resource
schema.
But
you
don't
change
the
prefix
name
for
the
schema
and
you
don't
go
update
the
API
export
from
the
new
prefix
name.
You
basically
break
kcp,
where
you
can't
start
up,
because
it's
trying
to
do
an
update
on
an
existing
schema
instead
of
a
create.
A
Deleting
a
deployment
from
workspace
doesn't
delete.
The
Pod
I
know
there
was
some
a
little
bit
of
back
and
forth
Stefan
you
had
looked
at
this
I
think
it
probably
needs
more
investigation
because
we
know
this
works,
so
there
must
be
something
specific
here.
That's
breaking
in.
A
Okay
site
is.
A
This
one
I
know
David,
you
had
some
back
and
forth
on
this.
Did
you
try
to
reproduce
it
yourself
or
were.
D
The
issue
I've
already
sent
some
and
then
I
was
asking
for
more
especially
I'm
was
suspecting
yeah
things
you
know
in
the
API
export
or
also
Shima,
some
so
I
I.
You
know,
probably
I
have
to
look
at
the
last
answers.
Obviously,
just
now
it's
the
last
okay,
so
this
is
still
under
investigation.
D
A
Right,
can
you
make
sure
somebody
takes
over
yeah.
A
Okay,
so
that's
it
because
I'm
gonna
come
back
to
this
one,
so
any
other
topics
or
shall
We
End
early.