►
From YouTube: GMT 2018-04-05 Containerization WG
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
A
Okay,
so
yeah,
thank
you
guys
for
joy
in
time
for
the
community
and
for
the
working
group
meeting,
but
I'm
denies
they
work
working
group
meeting
so
since,
like
today
we
have,
we
have
three
agenda
items
and
so
I
think
we
can
yeah.
Probably
you
finished
like
in
40
minutes
or
15
minutes,
so
the
first
one,
the
rolling
ownership
so
India
and
great,
have
been
working
on
to
figure
out
a
potential
solution
and
and
investigate
on
some
use
cases
and
this
this
the
world.
A
Honestly,
it
become
more
become
prioritized
by
by
mesosphere
side,
because
we
gonna
do
more
like
user
management
for
our
dis
us.
So
so
we
would
like
to
push
forward
a
little
bit
to
max
and
ProQuest
on
the
road
knowledge
it
search
in
chain
started
some
other
investigations
on
the
world
ownership
things,
and
this
is
the
doc
we
used
to
have
a
talk
about
the
board
and
ownership.
That
chain
started
another
one
to
cover
a
little
more
use
cases,
so
I
think
I
think
even
trendy
not
reach.
A
The
boy
like
reach
is
the
better
solution
and
I
think
he's
still
at
the
point
like
he
was
trying
to
understand,
which
is
the
best.
What
was
the
what's?
The
use
case
covered
with
the
our
crawler
ownership
and
he
tried
to
summarize
the
war
or
the
whirling
related
issues
on
missus
sure
or
from
the
customer
issues.
A
So
some
of
them
mentioned
on
the
on
alia
stock
and
some
of
them
may
be
new,
so
we
could
quickly
look
review
the
use
cases
we
have
equivalent
type
or
missus
so,
which
is
listed
here
in
this
table.
Same
lost
past
volume,
secret
volume,
persistent
volume
image
and
host
volume
and
I
think
for
the
first.
For
the
first
case,.
A
In
alia,
stone
it
is
mentioned
in
detail,
I
think
it
even
mentions
something,
but
it
is
not
that
in
detail
and
enchant.
It
goes
through,
like
the
testing
using
ratios,
let's
execute
a
test
framework
and
if
the,
if
the
executors,
if
the
agent
or
executor,
is
launched
by
the
root
user
or
say
like
if
the,
if
the,
if
the
continuous,
if
the
parent,
continuous
with
the
root
user,
but
the
try
continue,
try
to
define
symbols.
Past
volume
with
a
long
route
continue
with
another
user.
A
A
They
rely
on
the
default
security
to
to
specify
the
position
of
volume.
Right
now
we
are
doing
a
hack.
We
rely
on
this
de
persecuted,
so
basically
persistent
volume.
So
something
about
this
persistent
volumes
are
primitive,
like
user
could
reserve
some
resources
and
satisfy
a
volume
under
the
resources
which
gonna
be
handle
by
measures
and
by
among
the
persistent
volume
from
the
host
into
the
container.
A
But
if
the
persistent
one
is
specified
for
any
task
which
is
launched
in
nest
container
and
if
the
user
use
the
API
of
mesas
you
if
the
device
is
acutely,
we
have
some
tact
there,
which
we
could
not
like
support
the
persistent
all
in
primitive.
So
we
lived
on
the
same
bus
path
slowly
in
the
default
security
to
PI
mount
the
positional
from
the
parent
container
into
the
try
container.
A
A
A
This
is
also
a
tactic
which
we
don't
clearly
specify
a
volume
permission.
So
let's
say
it
will
be
the
kind
of
an
issue
in
the
future.
If
some
user
offering
work,
they
would
like
to
defy
some
role
in
any
of
the
world
and
try
to
be
shared
or
finessed
container
or
sibling
continue,
so
we
they
might
end
up
evil
ways.
So
this,
for
me
issue
also
is
also
Ted
bad.
A
We
need
to
take
into
consideration
and
yeah.
We.
Basically,
we
did
not
start
to
discuss
about
like
what's
the
potential
solution,
I
think
great,
quick
and
and
alia
might
have
a
better
idea
on
about
what
those
what
those
solution
is,
but
we
just
tried
to
like
cover
more
use
case.
It's
be
my
attack
that
we
might
want
to
address
at
this
moment
for
the
for
the
Olin
ownership
use
case
update.
B
Looked
at
this
document
we
had
a
sync
with
James
peach
and
she
also
to
chat
about
the
user
name,
space
work
and
if
there
was
any
any
coupling
between
these
two
features,
and
we
were
I,
think
we
were
in
general,
preferring
the
ACL
approach,
the
Linux
ACL
approach
and
at
our
last
meeting
we
kind
of
left
it
where
we
were
going
to
invest,
though
the
one
question
we
had
was
whether
or
not
Linux
ACLs
would
be
supported
in
all
filesystem
types
that
we
might
be
interested
in
supporting.
B
So
the
last
meeting
that
we
had,
we
were
going
to
investigate,
investigate
that.
Just
kind
of
confirmed
that
Linux
ACLs
would
be
viable
would
be
supported
on
all
file
systems
that
we're
interested
in,
and
that
was
a
while
back
I
think
I
got
distracted
by
other
priorities.
I
haven't
spoken
with
Ilya
recently
about
it,
so
I
think
we
need
to.
B
We
still
have
yet
to
do
that
due
diligence
and
sync
on
it.
So
perhaps
Ilya
myself
and
Shion
should
have
a
meeting
go
over
jion's
document,
because
I
haven't
had
a
chance
to
look
at
that
yet,
but
our
at
least
with
the
use
cases
that
we
were
considering
before
I
haven't
looked
at
Chen
stock.
But
we
thought
that
Linux
ACLs
verge.
It
was
just
a
cleaner
solution
in
general
because
you
don't
have
to
worry
about
dynamically,
creating
and
cleaning
up
these
supplementary
groups,
so
that
does
the
current
status.
A
Yeah
I
think
I
think
I
think
Chen
is
not
did
not
reach
deploy
on
investigating
solutions.
Yet
because
he's
trying
to
work
on
life
totally
understand
what
what
issues
or
use
cases
we
should
cover
with
the
appalling
ownership,
but
but
I
remember
yeah
this.
This
has
been
proposed
for
a
long
while
by
Ilya
like
last
summer,
and
there
we
used
to
have
Microsoft
folks,
I
guess
I,
don't
think
any
Microsoft
first
joy.
The
meeting
today
that
they
have
been
I
believe
they
have
preferred
more
on
the
ACL
solution.
A
I
am
assuming
Michaels
windows
have
son
like
corresponding
ACL
support
on
Windows
machine
and
for
Sacramento
I
guess
it
is
pretty
like
Linux
specific,
and
they
also
train
also
like
the
Tucson
investigation.
Unlike
was
the
doctor
and
kubernetes
solution,
and
basically,
both
of
them
rely
on
the
supreme
military
group
to
support
the
volume
ownership
with
different
user,
and
basically
it
is
not
an
implementation
of
kubernetes
because
I
they
don't
have
a
general
interface
tube
to
manage
the
ownership
or
user
in
kubernetes.
A
So
they
just
basically
used
what
docker
demon
provided
so
because
the
hood
demon
it
provide
with
some
argument
some
options
to
specify
the
corresponding
sub
romantic
groups.
So
they
it
just
rely
on
the
docker
feature
and
I
think
they
prefer
super
military
coup
because
of
lag,
and
some
reason
and
I
did
see
some
like
open
issues
with
their
volume,
ownership
management,
so
yeah
I,
guess
we
it's
very.
A
It's
very
good,
starting
from
India
and
Gregg
to
figure
out
like
we
have
some
more
some
ACL,
some
other
potential
solution
and
then
I
feel
like
it
might
be
likely.
If
you
want
to
support
Windows
well
on
with
the
rolling
ownership,
we
might
end
up
if
the
ACL,
but
more
investigations
needed,
because
we
need
to
like
figure.
A
What's
the
trade-off
between
these
two
cuz,
we
also
while
they
why
we
are
trying
to
like
discussed
like
what's
the
potential
if
I
sum
this,
if
I
an
open
issue
from
the
cabinet
from
kubernetes
about
like
the
volume
ownership,
with
some
strict
permission
this.
Basically,
this
this
discussion
is
about
like
discussion.
A
It
is
about
like,
if
the,
if
the
kubernetes
user
launch
apart
as
root
and
it
changed
the
user
in
the
application
and,
for
example,
it
finished
if
initially
it
finished
some
something
in
the
application,
as
richness
group
permission
and
up
stir
is
done
for
safer
security
or
sexy
issue,
a
reason
it
changed,
the
user
to
non
root
user
and
the
non
rule
user
might
change
a
couple
time.
So
so,
in
that
case,
we
need
to.
A
A
For
the
photograph
we
have
everything
larger
container
with
depends
on
a
bit.
We
depend
so
on
the
can
in
the
info,
and
we
have
the
user.
We
get
the
user
from
the
comment
info
and
then
that
one
gonna
be
take
into
the
ACL
list,
but
we
only
have
one
user
in
that
case.
So
if
we
want
to
like
a
multiple
UID
and
GID
into
the
ACL,
we
probably
need
to
change
the
the
API
we
have
for
the
user.
But
basically
those
are
the
detailed
implementation
detail.
A
So
you
just
have
some
like
very
brief
discussion
yesterday
about
like
if
we
really
want
to
like
a
last
allow
framework
to
specify
multiple
UID
GID
to
the
list
to
allow
to
allow
a
container
with
multiple,
your
ID
in
the
ACL.
So
we
probably
need
to
happen.
I
see
how
a
general
ACL
interface
as
a
new
component
amazes,
and
we
may
also
like
to
adjust
some
tack
there.
A
So
we
used
to
have
some
discussion
like
to
move
the
command
info
user
to
continue
info,
because
user
is
continuous
Pacific
we
used
to
have
the
user
income
and
info.
It
is
from
the
legacy
reason
because
Pierce
lis
four
years
ago,
we
when
we
have
still
have
the
command
tasks,
and
then
we
have
the
command
info.
We
don't
have
to
continue
info
yet,
and
it
seems
to
me
ideally
if
even
we
don't
specify
an
image.
We
have
just
have
a
comment
through
launcher
comment
has
or
missus
to
share
the
file
file
system.
A
We
could
still
have
some
partial
isolation
depends
on
how
you
to
configure
isolator
is
so,
which
means
like
the
command
is
still
running
inside
of
a
container,
but
the
container
may
or
may
not
have
this
own
round
names
based
in
evidence-based,
exact
sure.
So
so
it
seems
to
me
this
is
a
legacy
issue
and
I
do
leave
ish
we
may
want
to
like
by
the
command
info
and
continue
in
field
together
in
the
other
tasks.
All
we
are
we
the
best
the
best
seems
to
me.
A
A
That's
something
people
have
mentioned,
something
have
been
doing
that
with
kubernetes
and
and
we
could
not,
we
couldn't
not.
It
is
basically,
it
is
a.
It
is
not
bug,
it
is
a
feature.
People
expect
it
and
I'm
not
sure
how
many
people
who
expect
to
behave
in
that
way
or
mazes,
but
I'm
assuming
this
is
various
security
concern.
If
they
care
about,
like
the
application,
might
do
something
dangerous,
which
my
modified
they're,
persistent
wall
and
data
data,
or
they
should
have
limited
access
to
some
volume
which
might
change
the
host
host
file
system.
A
A
This
is
something
I
I
I
I
I
would
take
that
as
a
potential
use
case
and
to
cover
this
use
case,
and
it
depends
on
how
complicated
we
decided
we
may.
We
may
be
able
to
cover
that
case
in
this
side,
when
we
designed
the
implementation
detail,
and
then
we
decided
api,
so
I
will
take
that
as
effective,
but
it's
just
like
spill
still
as
to
be
determined.
We
don't
know
we
cannot
go.
That
way.
B
A
So
yeah,
so
basically
just
a
brief
discussion
and
and
as
far
as
I
can
see
if
we
change
the
API
the
way
that
I
move
the
user
or
adding
some
new
or
move
the
common
info,
it
will
be
breaking
change,
which
is
like
hardly
happens
because
we
have
to
like
guarantee,
develop,
compatible
so
idly
in
the
field.
So
right
now,
then
people
launched
at
a
stick.
A
It's
not
necessary
to
specify
any
container
info,
and
if
you
want
to
specify
you
have
some
user
in
container
info,
which
means
user
to
for
the
pierced
user
they
have
is
somehow
I
can
change
for
them
if
they
want
to
honor
the
user
from
the
container
info.
So
but
I
do
in
the
future.
I
would
like
to
say
like
if,
for
any
of
the
framework
they
have
battery
for
any
of
the
task.
They
specify
continue
and
comment
inkle
together,
and
that
will
be
easier
if
we
have
some
other.
A
other
protocol
fails
which
we
need.
B
A
Yeah
cuz
depends
on
the
release
policy
question
out
on
the
mailing
list.
We
can
reduce
the
money
list,
which
means
this
is
like
the
early
of
me,
and
we
can
show
me
gonna
cut
around
cutting
the
one
six
around
like
by
the
end
of
this
month
for
the
RC
one,
so
yeah
I,
don't
think
I,
don't
think
we
could
finish
this
four
one.
Six
and
another
thing
it
is
like.
There
are
many
tack
that
related,
not
only
the
user
here.
A
A
If
we
want
to
resort
that,
however,
we
need
to
introduce
to
some
refactoring
and
then
introduce
some
new
agent
API
to
support.
That's
the
container
persistent
volume
primitive
so
that
one
I
can
there
was
some
discussion
before
and
it
might
take
like
four
to
six
weeks,
which
is
no
not
a
trivial
task,
so
yeah.
How
many
should
we
should
we
should
we
should
schedule
meeting
next
week
and
discuss
like
what's
the
best
way
to
to
adjust
the
tech
there
and
then
support
the
pursuit.
The
oil
industry
feature.
C
Haga
what
james
unfix
was
talk,
it
looks
very
interesting.
Can
you
I
didn't
see
a
link
to
it
anywhere?
So
oh
yeah.
A
C
A
C
D
C
It
only
has
apples,
windows,
apples
are
totally
sane
and
makes
sense.
So
if
we
want
to,
if
you
want
to
do
anything
with
Windows,
it
has
to
be
a
call
oriented
fault
default.
Linux,
a
calls
are
different.
Semantically
detective
way
from
Windows
apples,
which
is
a
problem,
is
it's
a
patch
set
which
is
then
landed
on
all
the
rich
Michael's
patch
set.
This
brings
Windows
like
a
call
semantics
to
to
Linux.
So
if
you
went
down
that
path,
you
could
have
a
consistent
at
work
model
problems.
C
A
A
Yeah
and
I
think
I.
Think
probably
you
should
like
I
think
we
don't
have
a
juror
to
put
every
of
this
use
case.
A
bug
together
in
one
place,
maybe
I
should
create.
Maybe
I
should
create,
create
JIRA,
create
an
epic
to
check
on
all
this
issue,
and
and
for
the
reply
about
you
mentioned
you,
you
guys
encountered
this
morning.
It
is
it.
Is
it
like
you
guys,
sharing
somewhere
in
four
different
continents,.
C
A
Okay,
yeah
for
the
for
the
for
the
volume
permission
things
I
think,
apart
from
the
except
persistent
volume,
we
did
not
do
anything
for
all
the
others,
so
yeah,
but
that
person
is
a
missing
part
of
missile
site
and
I.
Think
from
the
API
in
the
volume
in
the
volume,
the
volume
in
the
prototype,
we
have
the
rig
only
already
write
permission
things.
It
started
with
the
doctor
support
initially
and
the
position
volume.
So
we
only
have
this.
A
A
C
Wonder
whether
me
sources
even
should
really
even
be
doing
anything
here,
because
these
missiles,
the
storage
manager,
do
we
really
want
this
kind
of
implicit
behavior?
So
maybe,
with
some
of
these
things,
where
we
have
historically
in
these
features,
where
we've
done
something,
we've
done
the
permission
changes
already.
Maybe
we
should
just
think
about
whether
that's
the
right
thing
to
do
in
the
first
place.
A
Think
I
think
it's
really
up
to
mrs..
Mrs.
should
not
like
her
code.
It
some
of
those
some
of
those
like
morning,
permission
things,
but
Mrs
Cooper
buys
such
an
API
in
Cocoa
Puffs
to
allow
users
to
do
that.
I
think
that
might
be
they
might
be.
That
might
make
sense
because,
like
let's
say,
I've
been
operated,
it
has
they
have
like
thousands
of
container
and
then
they
have
thousands
of
tasks
they
they.
They
could
still
rely
on
some
external
tools
to
manage
the
voting
permissions.
A
But
if
it's
good
to
have
this
support
on
basis
and
just
up
to
them
read
or
not,
they
want
to
use
it,
and
it
is
just
like
the
secret
bro
I
oaken
show
and
those
things.
If
for
any
of
the
see
group,
it
can
be
managed
by
the
operator
from
the
colonel
site
and
to
configure
some
control
file
of
the
season
from
ioc
books,
but
lasers
could
also
put
by
the
api
I'm
further
back
so
user
they
can.
They
can
they
have
to.
A
A
C
I
am
not
convinced
that
my
sauce
should
arrange
that
sharing
to
work
force
before
all
arbitrary
configurations.
So
one
way
to
make
one
way
to
make
sure
way
to
work.
One
obvious
way
is
just
to
say,
run
your
tasks
as
the
same
user
and
that's
how
everyone
gets
sharing
to
work
in
in
storage
today
and
that's
well
understood.
Whereas
if
one
should
miss
us
handle
arbitrary
configurations
and
still
guarantee
that
sharing
it's
going
to
work
in
the
way
that
you
expect,
that
seems
potentially
possibly
unreasonably
hard
yeah.
A
That's
a
good
point:
there's
definitely
a
good
goodbye,
because
this
is
yeah.
We
have
some.
We
have
some
discussion
yesterday
about
this,
like
for
the
sharing
for
sharing
for
sheer
volume.
This
is
sorry
basis,
does
not
have
a
very
concrete
desire
or
primitive
or
such
a
concept
of
a
to
allowing
words
to
defy
a
sharing
share
volume.
A
It's
really
up
to
the
framework,
so
framework
will
used
like
the
same
host,
pass
all
the
same
same
vast
path
to
defy
and
then
to
specify
different
more
about
even
continue
and
accept
the
persistent
volume
in
persistent
volume
we
could
specify
involving
to
be
shared
for
all
the
others.
Missus
does
not
have
this
like
management
of
life,
how
to
defy
boring,
to
be
sure
and
for
those
straight
volume
measures,
because
missus
does
not
divide
that
so
missus
does
not
have
the
responsibility
to
make
sure
every
continue
with
different
users.
A
A
A
Yeah,
we
I
think
I
think
we
I
need
something
a
little
deeper
about
this
first,
it
might
not
be
like
something
nieces
two
should
manage
it.
Might
it
might
be
a
mess
it's
like
if
we
missus
really
like
manage
some
of
them,
but
and
then
the
user
end
up
with
I
could
not
change
it
when
the
continuous
running,
so
it
might
be
something
missus,
that's
wrong,
yet
I
think.
C
C
B
B
C
A
A
A
Okay,
so
yeah
we
don't
have.
We
are
running
out
of
time.
I
want
to
finish.
They
are
the
two
islands
today,
so
maybe
we
could
we
could
come.
You
could
come
up
with
an
email
threat
to
discuss
the
ownership
things
and
oh
yeah
and
epic,
so
we
could
create
we
create
we
could
attach
any
corresponding,
but
our
core
issues
underneath.
So
it's
better
to
follow
up
and
okay.
A
So
the
next
one
it
is
the
minimum
Linux
kernel
version
supported
on
basis.
So
right
now
in
ASIS,
documentation,
I
think
I
think
this
one
Android
has
done
a
great
job
on
this
one
to
do.
Investigation
on
some
issue
with
the
Freesat
stable,
so
right
now,
right
now
be
the
minimum
Linux
version
we
support
it
in
our
documentation
is
2.6
23
and
we
want
to
pump
it
to
2.6
to
the
a
for
the
preset
for
the
nested,
preset,
stable
support
and
sorry
entry.
I
did
not
mention
this
to
you
seriously.
D
D
A
Basically,
basically,
a
did
a
very
good
job
on
investigating
and
a
bracket
test,
and
this
firket
test.
It
comes
from
the
race
when
we
run
in
multiple
agents
simultaneously.
If
you
and
now
apply
the
race
between
like
creating
and
removing
testicle,
the
reason
of
this
theoretical
reason
is
created.
This
is
a
it
is
a.
It
is
a
bug
in
the
unit
test
it
exposed
the
son
of
son
check,
it's
an
STC
will
check
in
our
in
our
sieve
library,
so
in
Linux
secret
of
cpp.
A
Every
time
we
initialize
an
agent
we
can.
This
is
some
background
about
this.
So
every
time
we
initialize
an
agent,
we
will
create
a
testicle.
A
nested
has
the
group
underneath
the
freezer,
missus
recency
book
and
pretty
and
remove
an
emitter
three.
So
this
is
just
for
test
to
make
sure
for
occur.
No
for
okra.
No,
that's
not
support
this
feature
in
not
star
measures.
We
will
not
start
aging,
so
it
is
something
heck
is.
A
This
is
a
hack
down
by
timidly
like
more
than
four
years
ago,
so
for
now,
most
of
the
use
that
they
do
not
the
not
on
that
Oh
Colonel
ready.
So
I
think
it
is
entry
and
and
I
think
I
like
it
might
not
no
longer
be
an
Indian
issue
for
our
missus,
so
we
just
like,
if
we
have
multiple
agent
to
doing
this
in
might
end
up
with
a
race,
because
every
time
you
remove
a
single
pin,
you
will
verify
the
single
existed.
If
not,
we
will
return
an
error
instead
of
skip
it.
A
A
Think
is
this
one
yeah
it's
this
one,
so
it's
initially
implemented
in
two
by
six.
Twenty,
a
and
maysa
rely
on
this
feature
to
support
a
free
system
and
that's
the
future
sticker,
and
so
that's
the
reason
why
we
want
to
bump.
There
is
a
minimal
kernel
version,
a
little
bit
from
two
623
to
to
628
and
I,
think
I,
think
NJ.
We
should
send
out
an
email
to
the
to
the
tab
and
user
list
and
to
see
if
there's
any
objection
about
this
change,
and
then
we
could
like
change
the
documentation
course
and
tolerant.
A
D
C
A
A
A
Okay,
so
yeah
we
still
have
ten
minute,
so
I
would
like
to
do
a
little
bill
together
on
the
feature
spreadsheet
cuz
be
last
time
we
did.
That
is
I,
think
it's
two
months
ago
and
we
were
trying
to
defy
what
we
want
to
accomplish
for
1/6
+
I
square
mentioned
earlier.
In
this
meeting
we
can
cut
the
one
six
around
the
end
of
this
month
and
hopefully
we
we
talk.
A
We
should
for
the
release
around
the
early
of
me,
so
we
have
a
lot
of
islands
divided
for
different
coats
and
this
feature
I'm
not
sure
we
could.
I
I
just
want
to
have
a
bigger
understanding
like
what's
the
progress
of
this
these
targets
and
then
how
many
of
them
he
could
try
to
land
for
one
six,
because
we
only
have
one
month
left
and
and
then
we
can.
We
can
do
some
change
corresponding
me
to
and
make
sure
we
could
lens.
A
At
least
we
could
learn
something
then
some
feature
into
one
six:
okay,
so
for
the
first
one,
documentation,
I
think
this
is
ongoing
issue.
It
should
be
ongoing
and
we
we
did.
You
have
to
dock
down
in
the
past
six
months
and
then
we
do
a
lot
of
change
on
documentation.
I
think
this
one.
We
do
have
a
lot
of
progress
for
one
six
and
for
the
second
one
it
is
use
a
namespace,
so
so
James
I'm,
not
sure
you
have
cycle
to
on
this
one
is
this.
Do
unfortunately,.
C
A
A
C
A
Yeah
so
for
the
second
supports,
basically,
there
was
some
decide
to
set
aside
up
from
Jay
and
he
implements
some
patches
which
I
did
not
learn
yet
because
I
I
think
we
could
do
a
little
improvement.
But
if,
if
you
next
one,
he
can
pick
those
patches,
maybe
the
math.
If
you,
if
you
survey
the
professor
a
little
bit
but
most
likely,
we
can
make
no,
not
gonna
blend
this.
For
once
inside
this
yeah.
C
A
A
Yep
yeah,
have
you
update.
I,
will
update
the
target
release
version
and
yeah
and
for
unified
artifact
store
I
think
we
I
do
not
have
a
chance
to
start
at
this
side
out
yet,
and
this
design
yeah
I
should
I
should
have
done
that
earlier
and
not
if
a
sighing,
verification
and
volume
on
this
ship
so
yeah
promoting
ownership
we
already
discussed
so
hopefully
we
can
get
the
decide
to
get
the
decide
down
by
1/6,
but
for
the
other
two
I
think
it
will
be.