►
From YouTube: KubeVirt Community Meeting 2021-09-29
Description
Meeting Notes: https://docs.google.com/document/d/1kyhpWlEPzZtQJSjJlAqhPcn3t0Mt_o0amhpuNPGs1Ls/edit#heading=h.rcp2na97dbjs
A
Hello,
everyone:
it
is
september,
29th
2021,
and
this
is
the
weekly
covert
meeting.
I'm
your
host
chris
caligari.
A
A
Okay,
we
usually
give
a
minute
to
for
introductions
go
through
the
list
and
it
looks
like
everybody
here
is
tenured,
so
we
can
skip
that
section
and
get
right
into
the
agenda.
Miguel
go
ahead
and
start.
C
B
Okay,
cool
thanks
yeah,
so
I
just
wanted
to
share
with
the
community
like
design
documents.
I've
been
writing
for
the
last
week
or
something
it's
about:
hot
plugging
and
hot
unplugging
network
interfaces
from
running
virtual
machines.
B
It
has
a
kind
of
two
parts,
the
first
it
focuses
a
lot
on
on
multis
and
the
second
part
on
hubert
and
basically
tries
to
define
like
how
to
use
and
the
the
api.
I've
been
already
shopping
this
around
on
the
multisteam.
I
got
some
initial
feedback
from
them
and
the
good
news
is.
They
are
quite
okay
with
with
having
like
the
the
required
changes
on
multis
to
have
this
happen,
so
I'm
just
trying
to
get
initial
feedback
on
the
idea
from
you,
and
that
is
basically
it.
B
I've
looked
at
the
I've
looked
at
how
we
do
hot
plug
and
hot
unplug
for
disks,
and
I've
tried
to
kind
of
follow
the
the
same
approach
like
whatever
the
api.
It
is
that
we
have.
I
agree
that
we
should
make
it
consistent
across
both
disks
and
interfaces.
So
that's
kind
of
the
overall
idea
here
and
I'm
not
saying
we
should
go
over
this
right
now,
but
I
really
appreciate
some
feedback
on
on
this
thing.
B
I
mean
yeah,
it's
a
lot
of.
I
I've
unwrapped
this
like
yesterday,
so
I
wasn't
expecting
anyone
to
just
jump
that
fast
into
it,
but
I
think
okay,
excellent.
E
I'm
also
happy
to
have
a
look.
What
just
one
thing
did
I
understand
it
correctly?
That
maltose
is,
will
support
hot
plugging
and
interface
to
the
pot.
B
Around
and
they're
they're
yeah
they're
welcome
to
it
like
the
overall
idea
from
like
the
let's
say,
from
a
user
of
malta's
perspective,
it
would
be
you
you
just
update
the
annotation
on
the
on
the
network
list
annotation
on
the
on
the
pod
and
it
would
unplug
or
unplug
the
interfaces
that
you
add
or
remove
and
they're.
Okay.
With
this
approach.
A
Yeah
I
sent
out
that
email
last
week
regarding
that
customer
interested
in
running
kubert
on
openshift
dedicated.
They
were
also
very
interested
in
in
using
multis
with
coo
bird.
This
is
perfect
timing.
A
Okay,
thank
you,
miguel,
anything
else
on
the
agenda.
F
Yeah,
so
as
it's
coming
up,
I
actually
took
the
few
weeks
off
to
be
able
to
prepare
for
it
kind
of
thoroughly,
but
there
is
a
just.
There
was
a
rhino
from
last
week.
There's
a
spreadsheet,
open,
I'm
ready
to
accept
ciders
and
people
to
go
and
join
this
kind
of
multi-site
cluster.
I
have
testing
done
that
replicates
site
to
site,
so
you
can
do
cross-joining
and
it
looks
like
the
only
response
I
got.
F
The
super
thread
was,
I
know,
I'll,
add
it
to
the
channel
again
or
something
it's
from
kevin
and
he
was
having
trouble
with
the
cm4.
F
I
just
posted
a
response
saying
he
could
probably
just
you
know,
use
a
usb
since
he
was
have
worried
about
the
small
amount
of
storage
he
has
with
his
flash
cm4,
but
regardless,
if
we
can
get
more
people
to
join
the
sooner,
we
can
get
some
testing
done,
the
more
bugs
we
can
find
sooner,
as
I'm
sure
there's
gonna
be.
You
know
some
various
amount
of
things
that
we
have
to
handle.
F
A
It
does
not
look
like
it:
okay,
sam
I'm,
I'm
for
sure
going
to
be
helping
you
with
this
starting
like
after
this
meeting.
A
And
if
anybody
else
wants
to
participate,
this
is
our
offings
open
live
demo.
A
We
will
be
speaking,
it's
probably
going
to
be
a
virtual
conference.
A
They've
been
flip-flopping
back
and
forth
between
having
to
be
in
person
versus
virtual,
but
pretty
sure
that
they're
they're
finalized
on
virtual
presentation
and
attendance,
so
we
won't
have
to
travel
across
the
country
and
then
that
should
make
running
the
demo
easier
since,
like
ours,
our
physical
machines
can
be
right
next
to
us
and
so
yeah.
If
you
want
to
participate,
find
this
the
super
thread
and
slack
and
and
join
us.
A
All
right,
let's
move
on
to
david
talking
about
pvcs,
go
ahead
david.
D
Hey
so
I
kind
of
have
a
larger
initiative
here,
and
that's
that
I
want
to.
I
want
to
make
virtual
machines
more
usable,
and
I
want
to
make
it
more
apparent
when
problems
occur,
and
I
want
to
get
earlier
user
feedback,
so
they
can
kind
of
determine
why
their
virtual
machine,
for
example,
hasn't
started,
and
I
have
a
centralized
way
of
kind
of
figuring
out.
D
Why,
like
on
the
virtual
machine,
object
or
coming
back
into
events
or
the
earlier,
we
can
give
this
feedback
the
better
as
well.
So
maybe
there's
some
things
we
can
do
even
when
they
post
updates
to
a
virtual
machine
or
post
the
virtual
machine
for
the
first
time
so
one
of
like,
if
I'm
just
kind
of
narrowly
looking
down
at
one
of
the
problems.
D
We
have
it's,
that
somebody
can
post
a
virtual
machine
today
with
a
clone
source,
a
pvc
clone
source,
and
if
that
ppc
clone
source
does
not
exist,
we
will
accept
the
virtual
machine.
That's
consistent
with
the
kubernetes
ecosystem,
but
the
virtual
machine
can't
start
until
that
clone
source
is
created.
D
So
I
wanted
to
kind
of
go
through
the
thought
experiment
of.
If
it
would
make
sense
for
us
to
begin
sending
back
warnings
to
a
user
at
creation
or
update
time
in
our
validation
hook,
to
tell
them
hey,
you
know:
you've
posted
this
virtual
machine
has
a
clone
source
that
doesn't
exist,
so
the
virtual
machine
is
not
going
to
be
able
to
start
until
this
pvc
exists.
D
Should
we
be
providing
feedback
like
that,
so
we'll
still
accept
the
virtual
machine,
update
or
creation,
but
should
we
be
providing
feedback
of
warnings
and
things
like
that
to
the
user
at
creation
and
update
time,
or
would
that
be
confusing?
I'm
trying
to
get
some
thoughts
there.
A
I
think
we'd
want
that
a
lot
of
times
this
there's
not
gonna,
be
hands
behind
these.
These
processes
right,
it's
gonna,
be
there's
going
to
be
automation,
tools
like
ansible
or
something
else
who
knows
that's
possible.
Yeah
like
I
would
I
would
want
is.
I
would
want
to
be
as
verbose
as
possible
so
that
loggings
can
either
be
viewed
manually
with
human
eyes
or
sent
out
to
some
kind
of
log
analyzer.
H
Not
mean
that
that
the
watch
already
shows
me
the
source,
pvc
for
the
validation
webhook.
So
we
would
really
have
to
query.
D
You're
saying
that
the
user
posts
a
virtual
machine
and
the
source
pvc
at
the
exact
same
or
like
really
close,
I
mean.
E
It's
what
it's
it's
one
file
and-
and
you
have
a
yammer
where
you
first
have
the
first
pvc
and
below
that
you
have
the
target
pvc
and
if
a
post,
if
you
post
them
with
one
cube
cut,
will
apply,
you
will
very
likely
see
an
error
because
the
watcher
didn't
catch
up.
Yet,
if
you
do
it
just,
for
instance,.
D
D
You
would
still
get
your
yeah.
E
D
Think
about
the
error
message
and
what
it's
saying
it's
saying
this
pvc
can't
this
virtual
machine
can't
start
until
the
pvc
exists.
So
I
guess,
if
so
you're
saying,
if
we
have
a
file,
it
has
the
source
pvc
in
it.
It
has
a
virtual
machine
in
it.
We
post
that
we're
going
to
see
that
error.
That
says
the
virtual
machine.
Okay,
so
then
the
user
would
see
that
error
and
wonder
what's
going
on
and
they
would
investigate
it
would
waste
their
time
because
everything's
fine.
E
For
instance,
although
it's
kind
of
you
can
also
end
up
in
situations
in
ultimate
in
automation,
where
you
will
always
see
the
warning,
then-
and
you
have
to
ignore
it
so,
but
what
what's
definitely
fine
from
my
perspective
and
does
not
have
these
consequences
is,
if
you
have,
for
instance,
events
which
say
then
yeah
delaying
the
start,
because
the
search
gpc
is
not
there
or
not
created
right.
I
mean
that's
fine,
that
are
events,
they're
known
to
be
asynchronous
and
yeah.
D
What
you're
saying
makes
total
sense
to
me
and
that's
the
same
sort
of
training
thought
that
I've
gone
down
until
I
think
about
how
virtual
machines
are
actually
used
and,
if
I
think
about
the
common
use
case,
it's
not
posting
a
source
pvc
that
we're
then
going
to
clone
for
our
virtual
machine
right
away.
It's
that
a
source
ppc
and
like
a
pre-populated
image
exists
somewhere
and
we're
creating
a
virtual
machine.
That's
referencing
that
as
our
boot
source.
D
E
Yeah,
I
mean
I
mean
you
can
probably
come
up
with
use
cases
where
you
think
it
might
be
handy
to
get
the
warning,
but
I'm
not
sure
if
we
shouldn't
even
bring
us
into
the
trouble
considering.
If
that
you
will
immediately
see
the
warning
event
anyway
so,
and
this
is
something
which
yeah
what
I'm
always
struggling
with
for
me.
It's
always
like
in
kubernetes.
E
You
just
have
to
look
at
your
desired
states
and
your
warnings
or
in
your
events
I
mean
that's
just
what's
telling
you
what's
happening
and
the
rest
is
not
telling
you
what
is
happening,
and
I
mean
we
can
probably
create
an
insufficient
list
of
extra
cases
where
we
do
additional
warnings,
but
it
still
does
not
free
you
from
the
requirement
to
check
events.
For
instance.
That's
that's
my
main
concern
also
like
uis
in
the
shift
ui.
E
For
instance,
you
would
immediately
see
you
created,
you
create
the
vm
and
you
wouldn't
immediately
see
an
even
a
warning
associated
to
the
object
directly
on
your
web
ui.
So
from
the
web
ui
perspective,
I
don't
think
that
for
users
it
would
even
make
a
difference.
If
you
get
the
warning
there
or
not,.
I
Let's
see
what
you're
getting
at
so
is
it
sufficient
to
yeah,
but
just
one
more
last
thing,
one
last
thing:
I
really
think
that
we
are
really
bad.
E
Yeah
some
weird
feeder
or
34,
but
with
my
microphone
it
just
turns
down
the
volume
yeah
I
just
wanted
to.
One
last
thing
I
wanted
to
say
is
that
I
think
that
we
are
definitely
bad
incubate
right
now
regarding
to
creating
always
events
in
the
right
moments
and
to
updating
the
stages
of
vms
and
vmis
properly,
always
with
the
most
needed
information.
That's
definitely
something
where
we
should
improve,
but
I
really
think
that
these
are
the
better
patterns
and
for
many
uses
it
would
not
even
make
a
difference.
E
J
Yeah,
I
agree
with
roman
and
then
there
are
other
I
mean
just
for
this
specific
case
too.
You
know
they
may
be
referencing
a
pvc
in
a
name
space
that
doesn't
exist.
So
that's
another
thing:
it
would
a
parameter
the
warning
or
something,
and
also
they
may
not
even
have
permission
to
clone
from
that
like
pvc
in
a
different
name
space.
D
How
would
we
get
information
that
the
user
doesn't
have
permission,
for
example,
to
clone?
Is
that
something
that
exists
in
cubert
or
is
that
something
that
exists
in
cdi.
J
No,
so
cuber
cougar
actually
does
some
checking
too
yeah.
Actually,
I
think
we
do
reject
it
now
if
they
don't
have
permission
to
do
it.
That's.
D
Okay,
so
we
we
have
warning
events
we
can
send.
We
also
have
conditions,
and
where
do
we
separate
the
two
like
where
what
belongs
to
the
warning?
What
belongs
as
a
condition,
how.
E
Does
yeah
and
we
have
the
printable
status,
which,
let's
me
implement
it?
I
think
yes
also
and
yeah.
The
printable
status,
I
think,
is
very
important
important
when
you're
on
cube
couple,
for
instance
and
yeah
in
the
events
of
course,
for
cute
cuddle
describe
regarding
to
conditions.
I
think
most
of
our
objects
have
one
catch-all
condition
like
sync
error
or
something
where
we
can
just
put
in
yeah.
There
is
something
wrong
and
because
of
that,
it
doesn't
progress.
D
Yeah,
but
this
isn't
necessarily
a
sync
error,
so
somebody
could
let's
say,
for
example,
on
this
warning:
if
we
want
to
give
feedback
to
the
user
that
the
pvc
doesn't
exist
and
the
virtual
machine
is
not
able
to
start,
I
yeah
we
would
still
post
the
data
volume
and
everything
we
would
just
be
waiting
on
the
data
volume
to
complete
and
it's
not
really
an
error.
It's
just
waiting
yeah.
I
guess.
E
That
that
yeah,
I
guess
the
printable
status-
would
be
important
here
and
I'm
not
sure
if
we
needed
to
reflect
it
on
the
status.
Even.
D
D
Yeah
and
then
it
gets
tricky,
so
we
can
send
a
warning
event
when
this
occurs
and
say:
hey
we've
create
that
that's
fine,
but
then
we're
saying
that
there's
three
places
a
user
looks
they
look
at
their
their
events.
They
look
at
the
air
condition
and
they
look
at
this.
Printable
status
get
a
high
level
overview
of
where
in
the
launch
flow
there
they're
at
can
we
remove,
but
am
I
I
mean
how.
E
Can
we
yeah,
I
mean,
like
also
like
the
pod,
doesn't
have
a
lot
of
sub
conditions,
but
I
don't
know
there
is
one
condition
which
success
port
not
ready
or
something,
and
then
you
have
different
reasons
for
that,
like
still
pulling
creating
initial
sandboxes
or
creating
sandboxes
feed
or
whatever
there's
all
kinds
of
things
there
which
are
not
necessarily
bad,
but
just
what's
going
on,
because
it's
just
saying
why
it's
not
progressing,
I
guess
yeah
just
having
a
location
like
this.
Maybe
we
even
have
it
right
now.
D
Yeah,
what
about
progressing
something
like
I
don't
know
that
would
make
a
lot
of
sense.
Progress.
E
D
D
E
E
I
would
expect
it
to
save
when,
for
instance,
the
epic
reference
pvc
does
not
exist,
and
I
think
we
just
edit
this,
then
it
will
tell
you
waiting
for
pvc
whatever
and
if,
if
to
be
created,
and
if
this
the
clone
target
is
there,
I
would
expect
that
it
kind
of
tells
me
waiting
for
the
clone
pvc
to
be
created.
E
D
E
E
As
verbose
for
users
as
like,
we
wanted
to
be
with
the
printable
status
one.
I
just
wanted
to
make
the
point
that
I
still
think
that
the
printable
one
has
the
advantage
that
it
can
really
just
show
show
the
user.
What's
the
most
interesting
thing
right
now,
no
matter
if
this
make
is
technically
interesting
or
not.
D
Yeah,
it
just
lacks
the
fidelity
to
give
any
real.
So
if
we
had
a
printable
status
with
a
reason,
then
that
would
satisfy
this,
except
if
we
did
that,
then
we're
coming
up
with
new.
It's
not
a
condition.
It's
not
a
warning
event.
It's
something
completely
new!
It's
not
a
condition.
I
guess
so.
D
Okay,
I'm
still
not
convinced
this
helps
the
user
experience
much
because
who's
looking
at
like
conditions
and.
E
E
Okay,
but,
but
maybe
I'm
not
sure
what
the
warning
I
I
would
basically
just
expect
to
permanently
see
the
most
accurate
running
minute
to
cuddle,
get
my
vmi
or
or
might
get
my
vm
where
in
of
the
object
like
I
get
it
for
for
parts
where
it
tells
me
image,
pull
back
off
or
where
it
tells
me
pulling
image
or
something
like
this.
I
see.
D
Okay,
so
we
would
have
a
status
field
that
and
how
would
we
aggregate
okay,
but
I
think
we
have
to
have
already
isn't
it
the
printable
status,
it's
a
printable
status.
It's
just
gonna
give
us.
E
E
E
I
mean
that's
what
normally
you
do
on
kubernetes
all
the
time
you
see
something
in
the
status
is
not
right.
You
do
cuddle
get
my
objects,
then
you
get
this
overview
with
the
status
where
it
tells
you.
This
is
not
really
right
and
then
do
you.
You
do
a
cube
cut
to
describe
on
the
object.
Then
you
get
all
the
details
with
warning
events
and
everything.
E
E
E
C
E
E
F
Yeah
so
there's
this
field
in
the
vmi
called
status.com
and
it's
you
know
you
can
use
it
for
like.
If
a
vm
is
paused,
for
example,
maybe
it
would
go
there,
I'm
just
I'm
just
looking
over.
I
was
looking
over.
You
know
how
you
know:
rook
does
it
with
rb
declining,
and
it
means
it
updates
the
status
object
of
the
pvc
which
in
this
case
you
know
the
keyboard
wouldn't
really
have
control
over,
because
I
would
just
go
with
a
pending
bound
scenario
of
the
volume,
but
I
guess
the
most
appropriate
area.
F
Would
you
probably
be
fast
black
conditions?
I
could
be
wrong,
but
it
sounds
like
you're
kind
of
wanting
to
replicate.
What's
done,
like
you
know,
openstack
or
aws,
for
example,
where
you
have
this
state
where
you're
loading
an
iso
onto
a
block
device,
and
it's
in
that
kind
of
cloning
phase.
I
guess
that
would
be.
Is
that
right.
D
Yeah,
so
the
printable
status
on
the
virtual
machine
gives
us
kind
of
a
high
level,
something
similar
to
what
you
described.
It
just
doesn't
give
us
anything
other
than
a
one
word
string
today,
so
we
wouldn't
get
like
a
reason
for
why
it's
waiting
in
this
state
or
anything
like
that,
but
yeah.
That
is
what
we're
at
least
what
I'm
shooting
for.
F
What
about
the
dot
like,
if
you
have,
if
you
reference,
satisfied
conditions,
dot
message
and
you
have
like
a
pause?
For
example,
it
will
give
you
a
message
saying:
vmi
was
paused
by
user
as
like
the
short
string
of
a
message.
If
you
ask,
if
he's
asking
for
more
like,
I
I'm
kind
of
seeing
the
use
case,
but
I'm
kind
of
not
like
I'm
between,
but
I'm
just
trying
to
figure
out
where
it
would
fit
in
to
help
out.
D
Yeah,
so
we
can,
for
example,
in
the
when
we're
registering
our
crd,
we
can
say
we
want
printable
status
and
we
can
say
what
fields
we
want
and
one.
This
could
be
a
condition
and
we
could
actually
print
the
reason
or
information
in
a
condition
when
somebody
does.
D
We
could
add
more
information
there.
The
thing
is,
we
don't
have
a
consistent
condition
that
gives
us
information
about
all
of
this.
That's
kind
of
the
problem,
I'm
saying
okay,
so
there's
one
condition
that
tells
us
if,
like
an
error,
occurred
during
the
reconciliation
of
the
virtual
machine,
but
that
might
not
always
give
us
the
broad
category
of
everything
that
we
need.
So
there's
sometimes.
K
D
D
E
C
L
D
I
kind
of
like
that
idea
of
a
car
like
the
equivalent
of
the
classes,
so
it's
a
list
of
things
that
we
want
to
report
back
to
the
user.
When
we
do
reconcile
that's
interesting,
it
might
not
always
even
be
anything
most
of
the
time.
It's
not
going
to
be
anything,
but
it
gives
us
the
ability
to
send
it
back,
multiple
things,
so
it
might
not
just
be
one
thing:
it
could
be
multiple
things
and
we'd,
essentially
just
override
it
on
every.
If
it
changes
on
every
reconciled,
loop
yeah,
maybe.
D
So
it
would
be
essentially
we'd
have
printable
status
and
then
we
would
have
something
like
printable
status
reasons
or
that's
a
terrible
name,
but
something
along
those
lines
where
it
would
be
a
list
of
reasons
for
the
status
being
where
it
is.
D
Okay,
I
think
I'll,
I
think
I'll
run
with
that,
a
little
bit
we'll
see
where
that
goes.
That's
a
that's!
A
cool
idea.
D
D
F
A
Okay,
any
pull
requests
that
you
have
pending
and
you
would
like
to
talk
about.
A
All
right
no
pull
requests.
Let's
just
take
a
quick
look
at
the
mailing
list
to
make
sure
all
conversations
are
handled.
A
G
A
A
Just
not
today,
bug
scrub,
who
is
not
using
fedora
34
and
can
share
their
screen
and
run
bug
scrub.
D
Go
and
click
on
it.
I
created
it.
Let's
see
what
the
response
has
done.
I
signed
it
to
lubo.
I
don't
know
if
he's
looked
at
it,
it
has
to
do
with
our
non-route
test
lanes
are
not
okay,
no
more
information,
we're
good.
We
can
yeah.
A
A
Are
we
triage
accepting
that.
A
We
just
want
to
look
at
unlabeled
issues,
correct.
D
Ones
that
haven't
been
looked
at
yeah,
so
we
have
a
okay,
so
somebody
says
kind
bug
I
I
don't
know
if
that.
Well,
let's
see
what
this
is.
E
E
D
E
So
for
the
usb
redirect
one,
I
think
victor
torso
would
be
happy
to
have
a
look
here
and
help
out.
C
E
C
E
I
think
I
already
see
the
issue
there.
This
is
probably
copied
again
from
an
outdated
post.
I
think
his
issue
is
the
access
mode
read,
write
once
command
line.
It
should
be
regret.
Many.
G
G
D
E
E
M
J
Yeah,
it's
weird
and
it
shows
the
pod
is
running.
Although
okay,
it's
a
timed
out
waiting
for
the
pod
to
be
ready,
so
yeah.
C
E
A
And
also,
if,
if
there's
any
related
blog
that
this
is
coming
from,
let
me
know
and
and
we
can
revalidate
that
blog
pretty
sure-
I
know
where
this
is
what
blog
post
this
is
referring
to.
E
J
E
J
B
One
thing
about
this
bug:
there
was
another
very
similar
bug
that
got
kind
of
fixed.
I
think,
like
the
fix
got
merged
yesterday
and
we
probably
yeah
leave
it
there,
like.
We
probably
want
to
cherry
pick
that
that
pull
request.
I
mentioned
here
the
6465
to
the
release.
B
E
What
kinds
of
functionality
does
it
exactly?
What
is
it
fixing
exactly
how
many
vms
would
be
affected
by
this?
Is
it
windows,
vms
only
or.
B
Oh
now,
it
affects
pretty
much
all
the
vms,
but
like
the
face
just
enforces
that
the
dhcp
v4
server
replies
with
a
with
ipv4
addresses
like
support
that
he
was
replying
with
ipv4
addresses
in
ipv6
format.
B
B
E
E
E
A
I
don't
know
david.
This
looks
like
one
of
your
pvc
status
messages.
D
Oh
yeah,
you
can't
migrate
what
so,
what
create
more
than
three
vms
at
a
time,
one
got
stuck
at
schedule:
status.
D
No
that
doesn't
make
me
good
migration.
Can
you
go
up
to
the
top
a
little.
D
D
This
must
be
an
old
version,
because
we
don't
even
have
the
volume
status
or
are
we
looking
at
the
vm
object
rather
than
the
you
know.
This
is
the
bmi.
K
D
K
First,
one
there's
like
an
undefined
or
whatever.
E
K
D
D
Ask
ryan
if
this
is
something
that
is
like
this
is
something
you
all
are
going
to
contribute
affects
to.
This
is
something
you
all
are
interested
in
and
I'll
say
that
nicely
backboarding.
D
A
All
right,
five
bugs
that's
pretty
good
7
57,
I
think
we'll
call
it
a
meeting
I'll
give
you
guys
give
you
all
three
three
minutes
to
relax
and
move
on
to
your
next
meetings.
A
Move
on
to
your
day.
Thank
you
for
attending
and
see
you
all
next
week.
Thank
you.
Everyone
all
right
bye,
but.