►
From YouTube: Kubernetes Federation WG sync 20180620
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).
B
B
B
C
Am
currently
a
little
bit
stuck
because
here
these
support
for
sub-resources
is
really
I,
don't
know
it's
just
something.
That's
been
under
active
development
for
111
and
so
moving
to
see.
Are
these
effectively
means
that
we
would
not
be
able
to
support
anything
less
than
11,
which
is
probably
fine,
but,
as
I
mentioned
yesterday,
I
wanted
to
raise
it,
because
that
means
there's
no
ability
to
like
install
Federation
on
a
previous
version
of
kubernetes
and
have
it
work
well.
C
Literally,
do
not
work
at
all.
You
cannot
register
a
CD
that
contains
the
status
resource
and
have
it
work.
Yeah
fix
that
bug
and
111
time
frame
and
I
am
basically
building
I.
Suppose
I
could
I
could
probably
pull
alpha
binary
for
kube
API
serve,
but
building
it
from
master
and
at
least
I
could
register
a
Sierra
newest
status,
but
when
I
go
to
try
to
update
status
specifically
for
cluster
resources,
because
that's
kind
of
instrumental
or
it's
like
fundamental
to
the
propagation
is
maintaining
cluster
status
like
ready,
not
ready,
and
it's
like
sorry.
C
That
is
not
something
you
can
do.
I
can't
find
that
resource.
So
not
seeing
this.
The
person
who's
been
working
on
C,
R,
D
sub
resources
to
see
what
deal
is
because
I'm
assuming
this
is
the
feature
that
should
work
so
I'm
expecting
it.
That's
really
something
we
could
resolve,
but,
as
I
said
like
there
will
be
no
ability
to
use
this
on
110
or
previous.
It
will
have
to
be
111
before
we're
lonely,
so
that.
D
D
C
D
B
C
A
B
B
B
D
Yes,
I
would
say
that
we,
you
should
one
verify
that
there's
actually
a
bug.
If
there
is
a
bug
me
and
it's
kind
of
get
it
fixed
and
111,
you
can
talk
about
whether
it
can
be
but
I
think
that
we're
effectively
in
a
position
of
like
we
may
be
ahead
of
others
in
terms
of
our
adoption
of
certain
features.
C
C
Only
I
think
we've
kind
of
touched
on
this,
but
I
just
wanted
to
clarify,
because
if
it's
some
point
in
the
future,
you
know
stakeholders
say:
oh
yeah
I
want
to
be
able
to
install
it
on
q1
10,
one
of
mine
whatever,
and
are
surprised
when
that's
not
possible,
could
no,
no,
that
would
be
undesirable
I'd.
Rather,
people
are
sort
of
aware
of
this
before
we
commit
to
the
move.
Well,
we
finalized.
D
D
It
doesn't
raise
the
question
of
like
which
I
think
you
just
touched
on
of
what
is
the
minimum
supported
version
that
we're
really
shooting
for
and
I
think
my
own
subjective
opinion
on
that
subject
is
I.
Think
it's
fine.
The
target
111
forward,
especially
since
the
the
timing
of
this,
if
I'm
not
mistaken,
should
be
close
to
one
111
is
release.
So
it's
also
an
alpha
component
and
I.
D
Think
it's
okay
in
that
sense,
to
using
features
that
might
not
be
in
every
cluster
yet
might
need
to
do
that,
like
a
target,
cluster
might
need
to
upgrade
to
get
a
bug
fix
for
to
correctly
support.
I'd,
be
curious
to
know,
though,
what
like
what
other
people
on
the
call
might
be
targeting
in
terms
of
version,
supported
versions.
B
What
I
understand
what
I
understand
is
that,
although
we
specify
like
in
tradition
given,
we
did
state
a
soft
disclaimer
kind
of
a
thing
that
we
do
not
take
responsibility
of
backward
compatibility
and
that
kind
of
stuff
still
it
used
to
work
with
a
good
amount
of
version.
Skew
like
say,
for
example,
Federation.
If
released
bits,
1.9
worked
with
1.8
and,
in
fact,
1.7
also.
C
C
B
Weren't
worked
or
something
what
I
was
trying
to
establish
for
us
that
what
I
have
seen
in
in
the
environments
is
that
almost
nobody
actually
uses
the
just
release
the
or
the
latest
k
test
version,
like
the
team
that
we
have
in
huawei,
which
handles
the
features
which
they
pull
in
from
upstream
would
have
a
fair
amount
of
gap
between
what
they
are
releasing
to
their
customers
and
what
the
community
or
the
upstream
version
is.
As
of
today.
B
B
D
I
think
there's
there's
a
tension
here
like
and
not
I
didn't
are
not
socially
I
mean
I
mean
there's
a
tension
between
adoption
in
terms
of
like
you're
fond
of
my
spirit.
My
experience
is
very
similar
to
yours
in
that,
like
the
pace
that
we
developed,
kubernetes
and
released
versions
of
kubernetes
does
not
necessarily
match
at
all
the
pace
that
a
customer
or
user
adopts
tone,
because
users
have
to
how
to
use
new
features
and
there's
only
so
much
they
can
take
at
a
time.
D
You
just
don't
want
to
skip
multiple
versions
in
one
upgrade,
etc,
etc.
So
I
think
you
make
a
very
good
point,
and
that
probably
means
for
folks
in
the
community
that
are
on
older
versions
of
communities
that
were
them
they
may
have
to
like
set
up
special
areas
to
try
it
out.
If
all
that
said
this,
we
are
talking
about
an
alpha,
so
I.
D
C
D
D
So
we
need
to
investigate
it,
but
I
think
it's
pretty
important
to
have
in
the
sense
that
it's
kind
of
unlikely
to
me
that
if
we
go
with
like
a
static,
API
server
that
you
can
easily-
and
maybe
maybe
there's
a
way
that
we
could
accomplish
this
without
moving
on
to
people.
But
I
think
an
important
part
of
what
makes
Federation
be
usable,
listed
using
a
type
study
that
we're
not
aware
of
in
this
book.
B
So
so
let
me
tell
you
see
a
scenario
like
this
is
what
I
session
couple
of
other
folks,
which
are
involved
with
this
effort,
had
in
mind
so
for
examples,
especially
in
China
I
mean
most
of
who
are
these
current
users
are
actually
based
off
the
China,
so
our
engagement
is
basically
through
meetups
or
customer
conferences
or
open
source
conferences
like
coupon
and
all
for
example.
Next
week
we
have
an
open
source
corn
and
the
next
con,
that's
one
of
the
biggest
conferences
in
China.
B
That's
what
we
are
going
to
attend,
so
we
engage
with
our
users
and
customers
in
a
way
they're.
One
of
my
responsibilities
is
to
showcase
Federation
d2
to
them,
and
then
you
know
encourage
them
to
try
to
use
it
in
whatever
questions
they
are
running
it
and
that's
that's.
How
I
mean
a
path
forward
that
we
see
for
vacation?
B
In
most
cases
it
will
not
be
the
latest,
if
we
say
at
least,
is
based
out
of
CID
and
that
transform
you
on
1.1
vanity.
It
is
like
no
show
complete
no-show
right,
so
that's
where
I
mean,
but
that's
where
I'm
coming
from.
We
basically
need
to
give
something
to
them,
which
is
usable
and
provided
yeah
Meru,
as
you
mentioned,
I
mean
if
it's
really.
We
also
need
to
establish
that
whatever
you
are
observing
is
really
a
blocker.
It's
an
issue
of
it's
like
remedies
where
some
baby
I
utilizing
it
wrong.
D
A
B
D
C
B
C
Of
dependencies
like
it
seemed
to
be
really
complicated
to
do
it.
Otherwise
I
mean
I'm,
not
but
I'm.
Not
even
sure
110
is
the
issue.
I
think
we
talked
about
111
being
like
the
minimum
yeah
the
minimum
for
the
Alpha,
but
realistically
we're
not
going
to
be
releasing
until
like
113.
Oh
god
is
more
realistically
the
minimum
yeah.
D
C
D
So
I
would
say
also
there's
another
factor
that
we
need
to
consider,
which
is
that
my
experience,
developing
like
additional
API
servers,
has
been
that
the
having
a
distinct
API
server
that
requires
its
own
storage
is
basically
a
non-starter
for
many
many
users
and
a
lot
of
cloud
vendors.
So
this
this
is
not
a
choice
between
this
is
not
like.
A
one-sided
choice
of
CR
DS
of
like
hypothetically,
might
have
some
like
clothes,
or
something
that
put
a
lower
bound
on
the
version
of
kubernetes
de
Federation.
Will.
D
D
C
Like
to
point
out
that
I'm
not
sure
that
C
or
D
versus
like
API
server,
like
I,
think
to
me
like
just
working
out
the
bugs
is
the
issue.
If,
if
for
some
reason,
somebody
wanted
to
be
able
to
run
C
or
D
based
Federation
like
API,
there's
nothing
stopping
them
from
running
a
newer
cube
API.
Just
for
that,
like
the
tiny
cluster
that
just
did
Federation
and
served
older
clusters
that
were
a
different
version,
provided
we,
you
know,
don't
mind
the
skew,
but
it's
yeah.
D
Safe,
that's
that's
a
really
good
point
and
I
think
that
we
should
probably
be
explicit
about
that.
Like
no
matter
no
matter.
What,
like
the
specific,
read
that
we're
talking
for
now,
like
no
matter
how
that
resolves,
but
there's
a
very
different,
like
version
dependency
between
what
Plus,
what
version
of
kubernetes
do
you
need
run
Federation
versus
what
versions
of
clusters
can
you
target
with
Federation.
C
Being
potentially
not
saying,
it's
necessarily
something
that
we
really
want
to
do,
but
there
would
be
nothing
stopping
somebody
from
creating
a
C
or
D
for
like
an
old
version
of
a
kubernetes
resource
like
sorry,
creating
the
propagation
primitives
for
an
older
version.
If
they
really
wanted
to
I'm,
not
saying
that
we
would
want
to
do
that
ourselves,
but,
like
once,
we
go
dynamic.
There's
nothing
stopping
somebody
from
supporting
like
things
that
they
want
to
support
that
make
sense.
C
B
It
does
see
my
intention
is
simple:
we
focus
that
as
a
workgroup
we
ideally
should
have
now
is
to
have
or
to
take
steps
which
can
move
us
towards
making
this
product
I
mean
useful
and
and
and
user.
Should
we
easily
be
able
to
try
it
out,
I
mean
as
a
first
step.
They
should
be
able
to
try
it
out
easily
and
if
they
see
the
utility,
then
maybe
they'll
be
willing
to
have
have
limitations
like
it's
only
forward
compatible,
so
they
need
to
run
them.
B
It
is
cluster
locators
cluster,
even
if
it's
it
means
only
to
support
them
absolutely
deployment.
Those
kind
of
things
may
be
can
be
secondary.
I,
unless,
unless
from
the
Red
Hat's
perspective,
you
already
have
DL
customers
and
requirements
for
this.
As
of
now,
and
that's
what
you
know
satisfied,
we
have
had
some
customers
for
addition
v1,
but
because
it
did
not
basically
make
the
progress
you
were
expecting
to
make.
D
D
The
story
of
like
you
want
to
go
from
nothing
to
you've,
got
a
toy
Federation
by,
and
that
includes
between
people.
How
to
stand
up.
Thank
you
for
Federation
to
live
the
story
of
you
want
to
try
it
out.
You
want
to
target
clusters
that
you
already
have
I.
Think
it's
very
close
to
you
want
to
try
it
out,
and
you
want
to
target
some
mid
EQ
like
things
that
you
just
stood
up,
that
don't
really
mean
anything
so,
no
matter
what
I
think
we
probably
need
to
have
documentation
for
how
to
do
both
yep.
B
D
D
C
I
was
just
looking
at
at
the
test
code
and
comparing
it
with
the
code.
The
cue
builder
was
generating,
and
it
looks
like
problem
is
that
cue
builder
doesn't
generate
sub
resources.
It
generates
an
open,
API,
schema
and
I
was
kind
of
I
was
obscuring
confirming
like
that.
It
doesn't
define
sub
resources.
Okay,
I
can
help
with
that
yeah
well,
but
yeah
we'll
probably
have
to
update
the
code
generator
to
do
that.
If
you
want
manual
and
generate
sub
resources
for
all
C
or
D
dos,
like
at
least,
we
have
a
way
forward.
Yeah.
D
C
B
D
Useful
info,
okay,
so
Lindsay
and
I
have
been
working
on
doing
an
automated
build
of
a
canary
image
from
master.
So
what
that
means?
Is
that,
like
every
time
there's
a
commit,
it
was
in
the
master,
well
rebuild
this
image
and
that's
really
like
the
first
step
that
we
need
to
make.
So
you
start
moving
away
from
you
need
you
need
to
actually
have
API
server,
builder
or
builder
CLI
on
your
machine
to
deploy
Federation
so
Lindsay.
D
A
I've
been
able
to
build
a
container
with
API
server
and
controller
manager,
which
is
what
the
API
server
builder
was
doing
well,
part
of
it.
So
now,
I
just
have
to
work
on
capturing
those
in
the
employee
Federation
script
and
make
sure
that
it's
working
not
locally
as
well.
I
guess
yeah
I.
B
I
have
a
question
for
this.
So
currently
we
have
it's
not
a
CI
job.
It's
like
a
free
Commission
jobs
which
runs
on
each
new,
commit
on
a
PR
yeah
and
we
don't
have
a
CI
like
a
periodic
job
kind
of
a
thing
or
a
post
commit
which
will
be
once
a
PR
is
like
much
after
that,
a
job
worker.
And
so
what
exactly?
Would
this
be
done
on
and
so.
D
On
the
ideas
that
we
turn
on,
so
we
we
do
have
a
master
bill
that
right
now,
even
though
the
script
is
called
pre-commit
playing,
it's
just
a
scripted
that
does
the
the
binary
builds
and
runs
the
go
tests
right
now
is
basically
what
it
does.
It
falls.
Live,
go
format
verify
okay,
okay,
the
idea
and
we
run
that
same
flow
when
there
is
a
commit
to
master.
So
this
is
a
travesty.
I
badge
on
the
federation.
D
So
when
it's
being
run
and
master
and
the
pre-commit
script,
that
does
all
that
stuff,
like
the
CI
tracks,
works
correctly,
that
will
build
a
container
image
and
push
to
the
quay
dot.
Io
image
repository
that
I've
set
up
so
she's
got
a
working
locally.
The
next
steps
are
to
like
actually
enable
it
to
happen
in
trance.
Does
that
answer
your
question.
B
D
B
C
A
D
So
what
barring
any
changes
in
terms
of
like
CIU,
we
can
also
make
Travis
do
this
for
release
tags
eventually,
so
that
when
you
do
a
tag
like
when
you
do
a
github
release,
it
will
kick
off
a
Travis
job
that
builds
that
tag
associated
with
the
release
and
I
can
push
tag
images
out
too,
which
is
actually
what
we've
done
in
service
catalog.
It
seems
to
work
pretty
well
so.
B
D
Yeah
we
we
shouldn't
probably
start
thinking
about
how
we're
going
to
run
ETS
in
yeah,
I
I.
Think
that's
a
pretty
big
question,
but
it's
it's
worth
starting
to
think
about.
We've
talked
about
that
before,
but
I
don't
think
anybody's
begun
working
it
all
on
running
with
Rio
of
kubernetes
clusters
or
like
holo
holo
API
servers
in
yeah
yeah.
No.
B
D
B
C
B
D
A
B
B
B
D
B
Kind
of
put
down
some
draft
about
that
feature.
It
is
not
thank
you,
but
I
will
try
to
take
you
along
this
feature
how
it
is
implemented
right
now,
and
we
do
if
you
have
any
questions
yeah,
you
can
interrupt
me
in
between
so
currently
in
v2.
Now
we
are
trying
to
implement
a
new
API,
that
is,
that
a
multi
cluster
service,
DNS
API,
which
contains
all
the
information
required
to
program
the
DNS
for
this
particular
culture,
so
yeah
I
think
I
need
not
go
through
in
detail
about
what
all
it
contains.
B
Think
it
is
already
there
in
the
code
right
now.
So,
okay,
there
is
a
two
options
we
thought
of
for
implementing.
One
first
option
would
be
maintaining
the
same
status
for
as
the
v1,
where
we
had
developed
our
DNS
provider,
library
and
already.
We
supported,
therefore
DNS
providers
in
that.
So
there
are
certain
differences
between
this
Florida
library
and
new
approach.
We
thought,
but
still
for
a
time
moment.
I
think
it
would
be
make
sense
to
hold
on
to
this
feature
for
little
more
until
we
mature
the
other
option.
B
So
so,
okay,
this
is
the
option
one.
So
this
is
what
comes
from
the
fetish
encounter
plane
or
the
fettuccine
pace
ever
so
this
DNS
year.
It
acts
like
a
tool
which
will
monitor
for
this
API
objects
and
then
program
the
DNS.
These
are
the
Global
DNS
options.
We
have.
We
have
a
DNS
provider
library
so
through
through
that
the
Federation
DNS
programs,
this
the
NSS
okay.
B
So
the
option
two
we
have
is
almost
similar,
but
instead
year,
Federation
DNS
will
be
acting
like
a
service
like
a
server
and
external
DNS
is
a
tool
in
kubernetes
incubator,
which
can
be
configured
to
use
any
source
right
now,
so
it
will
be
requesting
the
endpoints,
which
needs
to
be
provided
that
Federation,
DNS
server
will
be
providing
that
endpoints
list
and
external
DNS
will
take
care
of
programming
the
DNS
external
circular
by
Daniels.
So
there
are
quite
a
lot
of
form.
B
Dns
servers
already
supported
by
external
DNS
and
kind
of
they
are
taking
care
of
lot.
Many
details
related
to
the
NS
which
we
are
not
taking
care
in
our
implementations,
so
I
think
it
makes
sense
in
the
long
term
to
go
along
to
use
the
external
DNS.
So
only
thing
is:
maybe
we
need
to
see
further
to
mature
it.
So
for
now
we
have
both
options,
but
going
forward.
We
will
go
with
this
option
and
okay.
B
So,
okay,
just
in
like
this,
is
the
endpoint
which
in
point
structure,
which
is
expected
by
the
external
DNS,
the
name
and
the
targets,
what
type
of
record
TTL
and
the
labels
for
that
endpoint.
So,
okay,
so
further
to
that
and
now,
okay,
how
the
service
discovery
happens
through
this?
Okay,
it's
a
very
brief
I
am
going
to
explain
if
you
are
not
able
to
follow.
Maybe
you
can
interrupt
here
so
from
external
clients.
B
How
it
happens
is
there
is
a
balance
which
is
basically
one
of
this
dns
providers,
as
shown
here
so
so,
this
global
dns
is
going
to
be
programmed
by
our
Federation
DNS
tool
to
add
the
DNS
records
to
match
the
status
of
the
federated
service,
so
any
any
external
client.
For
example,
you
will
use
this
naming
scheme
to
to
access
the
service
federated
service,
the
protected
service
name,
the
namespace,
and
this
is
the
Federation
meal,
and
this
is
a
keyword,
and
this
is
the
domain.
B
So
so
we
can
form
a
Federation
of
multiple
clusters
and
Federation
has
a
mean
and
it
is
associated
with
our
domain
and
each
cluster.
There's
a
prerequisite
like
each
cluster
should
have
a
region
and
zone
labels,
so
so
any
client
who's
accessing
this
there
is,
you
will
find
a
DNS
record
for
this
and
there
could
be
a
multiple
targets
for
this
particular
DNS
record
and
it
further
to
this.
B
It
really
depends
on
the
capabilities
of
the
local
DNS,
how
what
records
it
is
giving
to
the
clients,
so
it
can
be
doing
a
round
robin
or
it
could
be
doing
the
maybe
the
nearest
service,
shark.
Okay,
based
on
this
on
the
capabilities
of
Google
DNS,
it
can
be
done
right
now,
yeah.
There
are
many
options
right
now
for
the
global
DNS
shashi
interrupting
I
have
a
question
about
this
petition.
Name
yeah:
where
is
it
I
mean
in
what
form
is
it
stored?
B
C
B
B
Coming
to
internal
client,
so
I
have
put
it
as
a
local
DNS
right
now
it
is
Cuba
DNS,
but
there
are
efforts
towards
moving
towards
Co
DNS
as
a
local
DNS
service.
So
what
had
happen
is
any
internal
client.
He
can
use
a
short
form
like
the
Federation
service
namespace
and
the
Federation
name.
The
the
name
will
be
auto
expanded
to
the
fully
qualified
domain
name
by
local
DNS,
but
before
doing
that,
it
will
search
for
a
local
chart
if
there
is
any
and
if
there
is
any
it
will
resolve
to
that.
B
Otherwise,
if
there
are
no
shorts
here,
so
that's
when
it
will
be
contact
the
global
DNS
with
this
naming
scheme
like
a
zone
and
then
the
region
and
then
the
domain,
so
the
global
DNS
will
be
able
to
resolve
this
to
the
nearest
zone.
If
not,
then
the
nearest
region,
if
that
also
is
not
possible
the
global
service.
B
Ok,
so
that's
how
it's
getting
resolved,
so
the
design
is
pretty
much
the
same,
which
we
had
for
quite
long
and
now.
So
this
is
how
it
happens.
The
resolution
starts
with
this
name,
and
if
it's,
if
it
contains
any
targeting
points,
it
would
resolve
to
that,
and
if
there
are
none
there
will
be
a
cname
record
which
will
be
pointing
to
region
level
of
a
name
and
the
similar
thing
happens
until
there
is
a
global
level.
If
there
are
now
no
service
shall
occur,
set
global.
The
global
level
then
service.
B
B
B
B
So
it
creates
a
separate
namespace
for
those
services.
Now
this
CT
CT
is
for
codeine,
assign
in
this
example
I
am
using
accordion
s
as
an
example
and
ET
city
is
kind
of
acting
like
a
back
end
for
that
code.
Enos.
So
please
don't
get
confused
for
the
CT
CD
and
there
is
a
code
eNOS
I
am
installing
through
help
chart.
B
Then
there
is
a
slight
configuration
we
need
to
do
for
the
kuba
DNS,
which
is
the
current
local
DNS,
are
to
fall
back
to
the
global
DNS
that
we
do
using
the
stuff
down
a
stub
zones
feature
within
the
kuba
DNS.
So
that's
done
for
all
the
classes,
so
we
have
created
a
config
map
in
a
federation
so
that
we'll
configure
those
config
map
in
all
the
Federated
cluster.
Now
now
the
Federation.
Now
the
DNS
is
all
set
up.
There's
a
code
eNOS
deployment
which
is
working
and
there
is
a
federation
DNS
which
is
working.
B
B
These
are
the
Han
chart
values
the
Federation
name,
I'm
using
is
Galactic
and
then
domain
name
is
T
zone
dot,
IO,
okay
and
then
provider
is
co
DNS,
and
this
is
a
container
image
which
I
have
already
pre
pushed
built
and
pushed
already
okay,
I'm.
Coming
to
Cordy
honest,
it
is
using
the
backend
etcd
back-end
and
the
zone
name,
it
will
be
authoritative,
is
DS
zone,
not
I,
go
and
the
endpoint.
This
is
the
etcd
endpoint,
the
backend
so
through,
which
will
be
programming
it
now
coming
to
the
demo.
So.
B
So
we
have
a
federated
replica
set
or
placement,
so
it
will
be
deployed
in
c1
and
c2
and
federated
replica
set
template,
which
is
a
simple
container
which
will
just
print
out
some
environment
variables.
Ok
and
then
we
have
a
federated
service
placement.
It
is
also
the
plot
on
c1
and
c2
both
and
we
have
a
federated
service
here.
There
is
a
load
balancer
service,
okay,
which
will
be
matching.
The
replica
sets
which
we
deployed
and
there
is
a
multi
twist
ourselves,
the
NS
object.
We
need
to
create
with
a
similar
name.
B
B
B
B
B
Idly
this
is
the
situation,
it
should
happen,
but
I'm
I'm
unable
to
set
it
up
the
DNS,
the
coordinates
which
is
there
right
now.
It
is
not
a
publicly
resolvable
service
right
now
so
and
I
was
not
able
to
set
it
up
in
a
result
chain.
So
currently
this
particular
part
is
not
working,
so
we
have
an
issue
with
that.
Actually,
but
it
can
resolve
if
we
are
giving
a
complete
domain
name.
B
Yeah,
it's
there's
a
slight
delay
mostly
what's
happening.
Is
it
resolves
to
see
one,
but
there
are
no
back
in
pots,
then.
Finally,
it
is
falling
back
to
sea
yeah,
sometimes
equal
to
c1.
Sometimes
it
goes
to
C.
Okay,
this
sorry,
okay!
So,
but
this
case
won't
happen
if
you
are
using
the
public
publicly
accessible
DNS
service,
so
okay,
hopefully
that
should
be
solved
and
okay
and
that's
it
any
questions.
B
B
B
B
Rest,
all
steps
are
same,
and
so
here
in
v2
we
are
running
the
sectional
DNS
service
has
on
chart,
and
these
are
the
values
default.
Values
which
we
are
doing
from.
Adder
is
co
DNS
and
there
is
a
new
connector.
It's
called
connector
sources
actually
sources,
and
this
will
be
the
service
name,
Federation
DNS
80,
and
this
will
be
our
80
City.
You
are
L
through
which
external
DNS
will
be
able
to
program.
B
B
B
It
just
has
the
names,
it
has
the
dns
names
and
it
has
a
targets
where
to
and
the
record
type
and
that's
all
it
needs,
so
it
is
already
taken
kappa
federation
DNS,
but
if
you
have
or
adding
that
in
the
user
supplied
object,
I
suspect,
like
this
is
the
dns
name,
and
then
this
this
will
be
the
thing.
So
that
is
what
will
be
supplied
to
the
endpoints.
B
B
B
Yeah,
these
are
the
DNS
records
which
it
is
getting
so
when
we,
when
we
do
this,
it
would
be
redirected
to
another
name
in
the
region
and
that
will
be
redirected
to
global
region.
Right
now
we
have
only
two
classes,
and
that
has
these
endpoints
right
now,
okay,
so
one
of
them
would
be
it
and
right
now
it
is
showing
the
other
records.
Also,
probably
this
is
because
of
the
caching
DNS
caching,
which
is
happening.
B
B
B
C
B
D
C
C
D
B
Yeah
and
I
have
one
more
request
like
do
you
think
you
guys
think
that
next
end
of
June
then
be
like
next
weekend?
I
guess
is
really
reasonable
time
to
cut
out
this
release,
or
maybe
a
week
further
might
be
better
like
I
did
not
get
the
point
Maru
was
making
Kate
is
also
is
releasing
1.11
end
of
June.
What
was
the
point
that
you
made
move.
B
C
C
C
B
C
Be
honest,
I'm
not
really
sure
we
want
to
be
targeting
something
like
that
for
alpha.
Just
because
we're
kind
of
down
to
the
wire
I
mean
granted
you're
kind
of
blocked
on
me,
getting
the
cube
builder
stuff
and
to
hammer
on
other
things
like
getting
usability.
Maybe
up
for
fixing
whatever
issues
but
I.
Think
right
now
or
I
would
suggest
a
feature
freeze
in
order.
A
B
C
B
We
I
mean
I,
had
this
thought
also
like
we
can
go
ahead
and
release
our
alpha,
like
the
first
cut
version
of
I
mean
like
end
of
June
or
whatever
time
that
we
actually
did
decide,
and
we
can
do
some
follow-up
releases
after
that.
Maybe
subtly
this
kind
of
a
thing.
There
are
two
weeks
something
like
that:
I.
C
Think
that
sounds
good
I
mean
how
it
doesn't
mean
we
ceased
it
all
I
mean
thank
all
suggesting
that
we
do
like.
Maybe
we
want
to
release
weekly
or
you
know,
bi-weekly
or
whatever
it's
more
than
you
get
to
a
point
where
we
want
to
start
we've,
given
it
to
a
point
where
people
can
start
using
it
and
have
a
reasonable
degree.
So.
D
D
If
you
find,
if,
if
you
want
to
extend
it,
you
can,
if
you
find
bugs,
will
work
with
you
on
it,
but
like
it's
basically
a
stamp
of
like
you,
don't
have
to
be
a
developer
to
use
it
right
now
you
need
the
development
dependencies
that
you
made.
It
look
like
it,
but
I'm
hopeful
that
we'll
get
more
people
using
it
and
trying
it
out
and
collaborating
with
us
and
that
will
have
a
relief,
a
reason
to
release
so
I
yeah
I
would
hope
that
alpha
is
the
start
of
things
accelerating
and
not
vice
versa.