►
From YouTube: Kubernetes SIG Storage Meeting 2022-06-30
Description
Kubernetes Storage Special-Interest-Group (SIG) Meeting - 30 June 2022
Meeting Notes/Agenda: https://docs.google.com/document/d/1-8KEG8AjAgKznS9NFm3qWqkGyCHmvU6HVl0sk5hwoAE/edit#heading=h.6x6bgi4pkxh5
Find out more about the Storage SIG here: https://github.com/kubernetes/community/tree/master/sig-storage
Moderator: Saad Ali (Google)
A
All
right
today
is
june
30
2022.
This
is
the
meeting
of
the
kubernetes
storage
special
interest
group.
As
a
reminder,
this
meeting
is
public
recorded
and
posted
on
youtube.
A
If
there's
anything
that
you
want
to
talk
about
today
feel
free
to
add
items
to
the
agenda,
you
can
find
the
link
to
the
agenda
doc
in
the
calendar.
Invite
I
see
a
few
folks
have
already
put
in
items.
So
that's
perfect.
First
we're
going
to
go
over
the
1.25
planning
spreadsheet
to
get
the
status
of
the
features
that
folks
are
working
on
for
the
125
release
and
then
we'll
go
into
any
of
the
design
discussions
that
folks
have
posted
here.
A
The
upcoming
deadlines
to
be
aware
of
code
freeze
is
going
to
be
on
august.
2.
we
just
passed.
The
enhancement
freeze,
enhancement
freeze
is
where
the
features
that
we
commit
to
for
125
for
the
release
are
officially
committed,
and
now
we
are
working
on
those
features
and
they
should
be
completed
a
code
complete
at
least
by
august
2.,
and
with
that,
let's
go
ahead
and
jump
into
our
planning
spreadsheet
and
we
can
get
status
updates
on
different
features.
A
So
first
item
here
is
delegate
fs
group
to
csi
driver
instead
of
cubelet,
including
updating
end
to
end
test
fabio
or
harmon.
Any
status
update
on
this.
B
So
I
have
this
feature
as
such
is
going
to
miss
125,
but
I
have
a
csi.
Spec
change
appear
open
that
moves
this
feature
to
ga,
and
rather
than
keeping
it
alpha
because
we
don't
want
to
move
this
feature
to
gn
kubernetes,
but
while
csi
feature
is
alpha
so
yeah
and
once
that
merges
the
rest
of
the
work
in
this
one
is
pretty
simple,
so
we'll
target
it.
Next
release
got.
A
Next,
we
have
so
I'm
going
to
mark
this
as
started
here
and
do
we
want
to
change
this
so
delegate
fs
group
to
csi
driver
instead
of
cubelet,
including
etv,
and
just
put
like
csi
updates
and
then
next
time
we
can
do
the
kubernetes
updates.
C
B
So
like
because
the
the
alpha
b
alpha
to
beta
graduation
criteria,
will
change
this
release
and
we
were
asked
to
add
e
to
e.
While
the
feature
was
in
alpha
and
if
it's
being
proposed
for
beta,
then
it
must
already
have
e2e
test.
So
we
didn't
have
e2
test
for
this
feature,
so
it's
going
to
miss
125
as
well,
and
but
I
will
be
working
on
adding
some
into
it
test
for
invisibility
so
that
we
can
target
it
beta
and
next
release.
So
that's
the
plan
for
this
release.
A
A
I'll
mark
that,
as
no
update
next
item
is
determine
mount
points
without
relying
on
proc
mount
yawn.
Are
you
on
call.
A
Okay
mark
that,
as
no
update
as
well
next
item
here
is
complete,
so
we'll
go
ahead
and
skip
that.
Then
we
have
csi
ephemeral
volumes
existing
api.
The
last
status
was
kept
as
being
reviewed,
needs
prr.
I
will
take
a
look.
I
believe
this
has
been
merged
already.
A
I'll
double
check
offline,
then
we
have
local
ephemeral,
storage
resource
management.
D
This
is
jeans
yeah,
so
I
think
it
I
think
it's
just
on
track.
Cap
is
merged.
D
Cool,
thank
you,
and
also
I
I
don't
know
sorry
I
was
getting
distracted
in
the
beginning.
I
don't
know
if
you
mentioned
the
the
vlog
op
in
so
I'll
go
for
it.
I
added
well
and
say
a
feature
blog
opt-in,
so
yeah.
If
anyone
wants
to
write
a
blog
for
the
feature
that
you
are
targeting
this
release,
we
can
add
it
here.
So
I've
added
a
few
to
get
started.
A
A
Next,
we
have
volume
group
api
previous
status
was
cap,
was
updated,
we'll
take
a
look
at
patrick's
proposal
on
dynamic
resource
provisioning.
Anything
new
here.
D
A
Sounds
good
so
we'll
go
ahead
and
review
that
afterwards
next
two
items
have
been
marked
as
complete,
so
we'll
go
ahead
and
skip
those.
Then
we
have
provision
volumes
from
cross.
Namespace
snapshot.
Pvc.
A
This
was
waiting
for
reviews.
Anything
new
here.
D
C
Format,
crossout.
D
Before
going
to
that,
jonathan
has
an
update
in
the
chat
he
said
the
csi
inaudible.
He
has
some
mic
issues.
A
Cool
thank
you
for
that
update
and
with
that
we
can
jump
to
csi
volume,
health,
additional
metrics.
D
Right
so
this
one
also
because
I
don't
have
e3
test
yet
couldn't
make
it
yeah
so
got
delayed
to
next
release.
G
A
All
right
so
then,
next
item
is
volume,
populator
data
source,
add
metrics
and
support
and
testing
mostly
out
of
tree
work.
Here.
H
H
No,
no,
there
is
no
no
work.
It's
beta
we're
not
going
to
ga.
We
should
probably
write
a
note
to
that
effect.
C
A
D
D
Last
week
we
got
a
few
new
people
from
microsoft
or
suicide
they're
going
to
help
with
some
of
the
tasks.
So
I
just
want
to
get
the
invitation
updated
based
on
the
the
new
cap
and
trying
to
get
those
submerged
before
the
1.25
code
phrase.
A
Yeah,
that
would
be
great,
so
good
call
out
for
folks
on
this
call
if
you
are
interested
in
participating
and
helping
out
somewhere.
Cozy
is
a
exciting
new
project,
they're
trying
to
effectively
be
the
csi
for
object,
storage,
which
is
a
tall
order,
but
pretty
exciting,
and
their
meeting
is
immediately
after
this
call.
So
if
you
rejoin
the
same
zoom
call
after
at
10
am
pacific
time,
you
will
be
able
to
join
the
cozy
meeting
and
you
can
participate
in
helping
there
all
right
with
that.
We'll
go
ahead
and
move
to
change,
block
tracking.
D
It
did
not,
so
the
cap
didn't
make
it
yeah,
so
it's
going
to
be
next
release.
We
did
have
some
meetings
discussing,
but
yesterday
we
had
we
have
a
discussion
at
the
data
protection
group
talking
about
the
how
this
would
work
with
aggregated
api
server,
but
I
think
we
still
have
not
sorted
out.
We
still
need
to
talk
more
to
understand,
more
details
and
then
compare
that
one
with
the
existing
proposal.
A
Got
it
I'll
go
ahead
and
drop
this
for
tracking
for
this
cycle,
and
then
we
can
pick
it
up
next
cycle.
I
Yeah
this
one
it
the
the
cap,
didn't
make
it
this
release
and
mainly
the
status
is.
We
are
working
on
the
feedback,
which
was
to
come
up
with
a
proxy
model
for
micro,
vm
runtimes
and
basically
incorporate
that
design
into
the
cap.
So
right
now
we
are
working
on
the
car
aside
to
make
sure
that
actually
works
out
and
will
like
integrate
the
findings
from
that
exercise
into
the
cap
and
can
re-review
it
as
part
of
125.
A
A
Next
item
is
csi
proxy
for
windows,
transition
to
privileged
containers.
Anything
new
here.
A
I
think
we
were
looking
for
volunteers
to
help
drive
this.
Anyone
end
up
picking
this
up
designed
for
this
cycle.
A
Okay,
well,
if
anyone
on
the
call
is
interested
feel
free
to
reach
out,
and
otherwise
I'm
going
to
go
ahead
and
mark
this,
as
not
tracked
for
this
cycle.
D
Yeah,
so
this
one
is
pretty
much
done
just
to
need
the
blog
and
documentation.
A
Cool,
that's
great
news.
Note
expansion,
secrets
so
I'll
go
ahead
and
leave
it
open
for
now,
but
we'll
wait
for
the
remaining
blog
and
documentation
to
be
completed.
F
A
Okay,
next
is
vsphere.
D
Yeah
so
much
so
I
do
have
a
question
about
the
blog,
because
I
think
last
time
humble
is
trying
to
write
a
blog
for
csm
migration,
seth
rpd,
but.
E
D
Think
there's
some.
There
are
some
comments.
When
the
you
know
the
talk
team
is
reviewing
this.
They
are
saying
why
you
know
why
don't
you
just
write
one
like
general
cs
migration
block
instead
of
just
for
one
driver.
I.
C
A
A
Yeah,
I
think
what
we
could
do
is
do
a
core
moving
csi
migration
to
ga
blog
post
and
then
underneath
that
have
a
section
for
each
one
of
these
and
where
their
status
currently
is.
D
D
Right
so
I
think
last
time
we
talked
about
this,
we
decided
that
we
want
to
go
ahead
and
dedicate,
and
I
need
I
need
to
check
with
tumble
because
he
said
he's
going
to
send
out
an
email
about
this.
One.
D
So
we
did
just-
I
think
we
mentioned
that
on
the
slack,
but
we
should
also
send
an
email
to
the
mailing
list
because
he
actually
sent
an
email
some
time
ago,
like
maybe
a
year
ago,
about
this,
so
he
can
just
you
know,
go
back
to
that
same
thread
and
saying
that
now
we
have
made
a
decision.
I
want
to
deprecate.
This
got
it.
D
A
D
We
are
not
yeah,
because
this
is
this:
maybe
this
one
can
press
out
for
now.
Okay,.
C
D
A
A
D
D
A
A
D
Otherwise,
how
would
you
wait
for
one
more
release,
because
now
we
need
a
ptp
test
marches.
First.
F
And
then,
let's
just
suffix
this
with
e2ejs.
D
Yeah,
so
this
is
the
same
situation
because
we
couldn't
get
the
e3
test
in
got
it.
Okay,
so
yes,
stay
in
your
office.
I
can
change
the
background,
notes.
Yeah.
D
Oh,
the
cap
is
merged.
Actually
it
is
targeting
yeah
he's
talking
beta,
so.
D
A
B
K
F
B
Owner
has
not
formally
volunteered
in
this
in
here,
but
there's
he's
proposing
some
changes
I'll
ask
him
to
yeah,
but.
B
Yeah,
so
this
guy
right
in
the
chat,
so
I
put
it
in
there
and
I
have
to
review
the
clip.
There
is
one
outstanding
thing
there
is
like
like
when
me
and
charlie,
the
previous
owner
we
talked
about
with
charlie,
took
some
notes
and
I
was
hoping
to
get
those
in
a
public
document,
because
we
discussed
the
issues
that
will
come
up
and
those
were
not
captured
in
the
enhancement.
Just
like
my
bad,
but
either
like.
If
charlie
doesn't
put
it,
then
I'll.
A
All
right
next
item
is
better
default.
Storage
class
cap
was
merged
here,
anything
new,
no,
no
update,
yeah
like
that
is
kept.
A
Distracted
this
one
is
handle
per
volume,
csi
driver
capabilities.
H
Oh
yeah,
we've
we've
been
unable
to
get
michelle
in
the
meeting,
so
we
need
to
skip.
We
need
to
pick
a
time,
that'll
work
for
her
and
everyone
else,
but
we
have
a
we're
starting
to
converge
on
on
a
tentative
strategy,
but
it's
one
that
she
was
opposed
to.
So
we've
got
to
get
her
buy-in
or
or
decide
to
do
something
else.
A
Okay,
cool.
Thank
you
for
that
update.
I'm
gonna
go
ahead
and
mark
this
as
started
since
the
discussions
are
underway,
and
this
is
design
all
right
so
with
that,
I
think
we're
done
with
the
125
planning
updates.
Thank
you.
Everyone
for
the
status
update,
so
next
up
carter
has
a
pr
skip.
Mount
point
checks
one
possible
during
mount
cleanup
carter.
You
want
to
talk
about
this.
M
Yeah,
so
this
one
has
been
around
for
a
while.
I
think
jing
is
the
kind
of
the
main
reviewer
on
it
and
I
don't
think
they're
on
the
call.
So
I
think
what
I
really
just
want
is
the
sig's
input
on
whether
this
is
reasonable
to
get
merged
before
the
turbulence
around
the
125
release.
M
It's
been
kind
of
difficult
for
me
to
get
eyes
on
it.
Nothing
significant
has
changed
on
the
pr.
In
probably
like
six
weeks,
it's
just
been
kind
of
hard
to
get
the
final
thumbs
up
from
folks.
M
N
M
Okay,
great
well
yeah
I'll
respond
to
your
comments.
I
mean
it
sounds
like
we
still
have
several
weeks
to
polish
this
up
before
kind
of
the
final
code
freeze.
So
it
sounds
like
we
should
be
able
to
reach
that.
M
E
N
A
All
right,
thank
you
both
with
that
we'll
move
on
to
design
reviews.
First
up,
shang
questions
regarding
volume
group
cap:
you
want
to
talk
about
this
one.
D
Yeah
so
with
the
volume
group
cap,
I
think,
went
back
and
forth
on
that
one
initially.
I
only
have
this
one
running
group
object.
So
I
don't
have
this
like
one
group
and
modern
group
content
like
the
pvdbc
or
snapshot
ones,
natural
content
model,
but
then
I
thought
okay.
Normally,
we
should
have
two
if
we
want
to
support
static
provisioning,
so
I
updated
the
cab
and
now
in
the
existing
cab.
There
are
basically
it's
a
kind
of
same
pattern.
D
One
group,
one
good
content,
but
then
apache
has
this
a
new
cap
that
just
got
merged
for
1.25,
but
he
is
proposing
this
dynamic
resource
allocation.
This
is
a
owned
by
signaled,
so
in
in
here
there
is
a
resource
class
and
there's
a
resource
claim,
but
there's
no
resource.
So
it's
like
there's
a
pv
c,
but
there's
no
pd,
it's
kind
of
like
that.
There's
no
binding
because
they
just
have
business,
just
one
object
per
resource
and
then
the
the
resource
handle
is
stored
in
the
resource
claim
status.
D
So
talk
to
him,
he
said
the
resource
handle
cannot
be
confidential.
If
you
know,
if
this
is
confidential,
then
the
driver
will
need
to
do
something
about
it
to
maybe
somehow
hide
it
only
display
the
one
that
is
allowed
to
be
shown
by
you
know
to
others
now.
D
So
if
we
are
using
this,
so
just
wondering
if
we
could
use
the
same
order
to
simplify
the
design,
if
we
can
have
one
object
in
step,
two
that'll
be
easier,
but
the
question
is:
are
we
fine
with
the
you
know,
showing
a
volume
handle
which
is
which
is
supposed
to
be
a
handle
on
the
solid
system
in
the
volume
group
status,
which
is
a
this
one
group
supposed
to
be
a
user
namespace
object?
I
guess
that's
the
question,
so
this
is.
L
D
That's
also
supported,
so
basically,
it's
up
to
the
driver,
so
kubernetes
will
not
be
the
one
who
is
worrying
about
that.
So
if,
if
you
already
have
some
resource
that
is
already
there
so
yeah
so
so
it's
not
going
to
be
like
a
declarative
model
anymore.
So
it's
not
like
you
put
that
in
your
stack.
You
don't
put
your
resource
handling
stack,
but
when
you're
doing
dynamic
provisioning,
but
if
you
already
have
some
existing
resources
available,
then
the
driver
can
just
pick
one
that
is
already
there.
D
L
K
L
D
Less
important
for
people
to
really
imported
it
for
a
specific
one:
oh
okay,
yeah
yeah.
That
makes
sense
so,
but
for
the
for
the
rewarding
group
case,
though,
since
we
still,
it's
still
like
wrapping
around
those
volumes
right,
so
you
are
so
the
static
provisioning
of
a
volume
was
to
be
supported.
It's
not
like.
If
we
just
have
one
more
group
object,
we
cannot
support
p
static,
provisional
period.
That's
that
one
is
still
there.
It's
just
like.
You
cannot
have
something
like
you.
D
Let's
say
you
know
a
volume
group
handle
that
is
on
your
three
system.
You
cannot
just
say
hey.
I
want
that.
I
wanna
handle
import
that
one.
I
cannot
do
that
anymore
right.
So
that's
the
thing
so
to
be.
Is
that
important?
I
guess
that's
the
question.
D
So
if
we
don't
right
so
basically
is
it
important
to
support
the
case.
Well,
I
want
to
say
I
want
to
import.
The
specific
one
group
handle
is
that
something
that
is
important
or
also
like
voting
group
snapshot,
handle
same
thing
right.
Similarly,
is
that
important,
or
can
we
just
have
the
dynamic
way?
Would
that
be
sufficient?
I
guess
that's
the
question.
A
I
can
imagine
a
use
case
where
you
know
somebody
has
an
existing
volume
group.
You
know
maybe
they're
it
was
created,
they're
recreating
their
cluster
or
something
like
that.
G
D
D
L
It's
been
a
while,
since
I've
it's
been
a
while,
since
I've
seen
the
volume
group
proposal,
but
is
it
that,
would
you
create
a
volume
group,
you
specify
like
a
bunch
of
pvcs,
that
you
want
to
be
part
of
that
group.
D
So
the
currently,
so
what
we
are
supporting
is
we
have.
We
are
not
going
to
be
so
what
we
are
going
to
do
is
you
you
will
be
creating
a
new
empty
one
group,
and
then
you
can
say
I
create
one
pvc.
Then
you
give
the
one
group
name,
you
basically
kind
of
create
a
new
pvc
and
add
into
the
group
one
at
a
time.
D
So
that's
that's
the
that
that's
one
floor
and
that
another
one
is
to
you
should
be
able
to
add
an
existing
pvc
to
your
group
so
that
one
seems
to
be
kind
of
more
important.
Well,
you
have
you
know,
pvcs
already
created
you
want
to
add
them
to
and
to
group
and
can
can.
D
L
Okay,
I
mean
so
I
guess
I'm
so,
I'm
just
trying
to
brainstorm.
L
L
D
Right,
I
think
we
did
have
that
we
actually
because
there
was
some
people
had
some
comments
on
here.
You
know
if
you
we
support
this
one,
adding
one
of
the
times
so
efficient.
We
did
actually
say
that,
like
the
like,
as
at
least
in
the
csi
space,
we
would
be
able
to
allow
multiple
volumes
to
be
added
or
removed.
Basically
there's
a
list
of
that.
D
So
we
did
say
that
and
then
in
the
like
a
controller
that
one
google
controller
will
be
the
one
who
is
kind
of
looking
at
the
list
of
them.
I
believe
a
bit
of
saddle
saying
it's
going
to
be
like
the
attached
to
that
controller
kind
of.
D
Right,
it's
going
to
be
like
a
yeah
bulk
ad
or
remove
something
like
that
yeah.
So
it's
not
like
so
yeah.
We
do
handle
multiples.
N
D
L
I
I
mean
yeah,
I
I
think
if
I
mean
we
do
have
to
consider
the
security
things
like
if
there's
any
potential
security
issue
of
exposing
that
handle
to
users.
N
A
Yeah,
it
sounds
to
me
like
there
is
an
import
use
case
here,
right
shanks
and
it
doesn't
work
if
you
have
a
single
object
right.
So
if,
if
I
want
to
take
an
existing
volume
group
that
has
a
set
of
pvcs
already
defined
yeah
import
that
into
my
cluster,
if
I
have
a
single
object,
then
it's
up
to
the
driver
to
try
and
figure
out
which
volume
group
I
was
talking
about
correct
yeah.
We
don't
yeah.
If
we
don't
have.
D
Two
objects,
then,
there's
no
deterministic
way
right.
D
I
I
think
it
definitely
would
be.
That
definitely
is
a
valid
use
case
for
people
who
want
imported
just
can
can
we
do
it?
Without
I
mean,
can
we
just
somehow
where
you'd
be
okay,
if
we
just
kind
of
using
this
dynamic
provisioning
way,
but
have
a
driver
to
figure
out
which
one
to
pick?
Would
that
be
good
enough?
I
guess
that's
the
question
right.
D
So
the
name
well
so
that
because
because
normally
we
do
have
the
name
that
we
give
at
provision
in
time
is
not
necessarily
a
name
that
is
returned
by
driver.
So
maybe
there's
a
way
for
driver
to
kind
of
translate
that,
and
then
I
don't
know,
I'm
just
thinking.
A
Yeah,
if
there's
some
way
that
we
can
get
away
with
you
know
the
driver
automatically
selecting
the
right
thing,
then
I
think
the
use
case
is
fulfilled
and
it
should
be
fine,
but
I
suspect
this
will
be
something
that
requires
kind
of
user
intervention.
User
knows
exactly.
D
D
Two
objects,
you.
Basically
in
the
let's
say:
if
volume,
let's
say
one
good
content,
is
your
admin
space
object
in
in
there?
You
can
provide
the
name
right.
That's
just
like
today
how
we
do
static
provisioning.
You
give
the
handle
name
like
one
handle
or
special
handle
name.
You
put
that
in
in
there.
D
N
Well,
yeah:
maybe
it's
there's
not
enough
time
for
of
us
to
think
the
the
rule,
so
maybe
we
can
think
offline
and
then
see
whether
it's
resolvable
and
think
about.
N
A
Okay
cool,
so
thank
you
for
the
discussion.
Let's
move
on
to
the
next
item,
abhishek
as
if
I
I
wanted
to
discuss
kep
661
and
see
if
anyone
is
still
driving
it,
if
not
I'd
like
to
help
with
it
abhishek
are
you
on
the
line.
O
D
G
D
D
But
I
think
command
also
is
going
to
review
that
one
believe
there
are
probably
still
issues
that
are
not
addressed
yet
from
previous
reviews.
But
if
you
can't.
O
Yeah,
no
I'd
love
to
be
in,
like
the
discussions
and
I'm
happy
to
help
with
implementation
as
well
like
we're
active
users
of
stateful
sets,
and
I
would
really.
E
B
Yeah,
I
think
we
might
have
to
set
up
some
calls
to
discuss
the
design.
It
would
be
good
to
hold
those
meetings
in
zoom
call
like
in
this
six
storage
zoom
and
get
this
recorded
and
I'll
talk
to
this
other
person
and
see
if
they
are
interested
in
driving
it
because
it's
it's
not
one
of
those
small
features
that
you
just
go
and
do
it
it's
tricky
and
it
requires
like
some
so
I'll
see
if
we
can
do
some
follow-up
call
and
do
design
and
and
drive
this
this
feature
to
completion.
O
B
O
Need
to
add
myself
to
this,
I
think
the
slack
as
well,
but
I
will
keep
a
note
of
it.
At
least.
A
Awesome,
thank
you
abhishek
for
volunteering
and
thank
you
vermont
for
the
follow
up
there.
Next.
O
A
Have
gary.
E
Yes,
of
course,
a
call-
I
was
waiting
for
all
day
came
in
this
second,
but
I
thought
we'll
be
right.
Back
yeah,
there's
a
a
something
called
a
special
resource
operator.
That's
going
to
be
renamed
into
something
called
autotree
modules
for
installing
kernel
modules
into
a
openshift
cluster,
but
it's
going
to
be
merged
into
the
sig
node.
I
believe
in
general,
kubernetes,
brett,
thurber
and
clinton
brought
said
if
you
guys
would
be
interested
they'd
be
willing
to
give
a
five-minute
overview
of
it.
A
Yeah,
I
think
installing
kernel
modules
is
an
important
thing
for
a
lot
of
csi
drivers,
so
yeah.
That
would
be
very
useful.
E
Great
so,
should
I
just
suggest
say:
join
the
next
meeting.
A
E
E
It
basically
it's
it's,
it's
a
it's
a
it's
an
operator
that
leverages
something
called
node
feeder
feature
discovery,
no
future
discovery,
labels
all
the
nodes
with
the
kernel
version,
and
so
that's
a
separate
operator,
and
so
what
sro
does
and
modules
does?
Is
it's
a
way
to
look
at
that
and
then
look
at
all
the
nodes
and
it
dynamically
creates
some
daemon
sets
that
have
permission
to
install
kernel
modules
and
sends
it
the
right
information
for
the
kernel
module.
So
you
install
the
right
kernel,
module
how's
that
right.
H
H
H
E
E
It
also,
it
also
will
will
automatically
update
if
it
notices
that
you
know
the
kernel
module's
been
updated.
E
Yeah
we
we
needed
to
do
it
for
a
major
financial,
so
so
it
makes
them
happy.
So
it
makes
my
life
easier.
A
Yeah,
that's
awesome
thanks
gary
and
feel
free
to
add
something
to
the
agenda
for
the
next
time.
If
you
get
confirmation
from
them.
A
Sounds
good.
Thank
you
all
right.
We
got
about
six
minutes
left.
Anyone
else
have
any
topic.
They
want
to
bring
up.