►
Description
Kubernetes Data Protection WG - Bi-Weekly Meeting - 11 August 2021
Meeting Notes/Agenda: -
Find out more about the DP WG here:
https://github.com/kubernetes/community/tree/master/wg-data-protection
Moderator: Xiangqian Yu (Google)
A
A
Can
everyone
see
my
screen,
yeah
cool
in
the
first
item
in
the
today's
agenda
is
to
go
through
the
white
paper.
We
have
made
a
little
bit
more
updates
on
it.
I
think
it
is
right
now
in
a
relatively
good
state.
We
will
go
through
it.
A
Secondly,
this
there
was
a
new
issue
created
in
the
williams
snapshot
ripple
from
humble.
I
think
he
is
here
he's
in
this
meeting.
We
will
go
through
briefly
what
this
issue
is
and
discuss
potential
solutions
for
that.
If
there's
any
open
issues,
then
we're
going
to
open
the
discussion
to
the
to
everybody.
Is
there
any
issue
over
there
yeah?
Let's
get
started
on
the
data
protection
white
paper.
Let
me
open
this
step
up.
A
A
There's
a
lot
of
content
over
there
shin
and
I
have
been
going
through
all
the
contents
in
this
documentation
or
end
goal-
is
to
publish
this
documentation
into
the
kubernetes
community,
hosted
underneath
the
data
protection
working
group
report.
A
In
that
case,
it
can
serve
as
the
kind
of
guidance
that
this
working
group
has
worked
together
to
try
to
establish,
and
it
contains
a
lot
of
content
for
those
who
have
not
got
a
chance
to
go
through
this
documentation.
Yet
please
spend
maybe
a
couple
of
ten
minutes
to
quickly
go
through
the
type
of
content.
Basically,
what
we
try
to
achieve
using
the
within
this
white
paper
are
the
four
ring.
A
So,
first
of
all,
we
want
to
define
what
is
the
scope
of
this
data
protection
working
group,
what
exactly
we're
trying
to
solve
and
what's
the
definition
of
the
of
the
data
protection
in
the
kubernetes
context,
and
then
we
have
to.
We
have
a
section
dedicated
to
rhythm
about
why
this
is
so
critical
at
this
moment.
A
Why
people?
Why
users
wants
this
and
then
a
lot
of
use
cases?
I
think
a
lot
of
people
have
been
working
on
this
section
as
well
great
huge
call
out
to
tom
on
this
section
talking
about
what
are
the
real
life
use
cases
we
are
looking
to
support
or
not
support
within
this
group
and
then
the
next
couple
sections
talking
about
existing
modules
in
kubernetes
and
the
missing
modules
in
kubernetes
to
allow
backup
vendors
to
use
kubernetes
native
apis
to
achieve
the
goals
up
to
then,
on
the
main.
A
Of
course,
there
are
many
modules
I
will
not
go
into
details
of
every
single
one
of
them.
You
guys
can
see
there's
actually
quite
a
bit
of
areas
are
being
researched,
discussed
or
even
deployed.
People
implemented
in
parallel
by
this
group.
Are
members
from
this
group
again
huge
clouds
to
those
who
are
working
on
those
specific
items
and
then
we
we
can
talk
about
what
we
think
about
is
the
workflow
to
do
a
backup
and
restoration
at
the
application
level,
because
that
that
would
be
all
kind
of
you
know.
A
First
major
milestone
to
reach
to
there
are
other
app
indexes
that
will
not.
I,
I
believe
we
don't
plan
to
put
this
into
the
white
paper
at
this
moment,
but
this
is
the
glance
of
this
documentation.
A
If
you
have
any
comments
or
suggestions,
please
please
just
comment
on
either
reach
out
to
the
section
owners
or
yeah
and
me
or
just
comment
directly
on
this
doc.
She
and
I
are
looking
to
publish
this
doc
within.
I
don't
know,
maybe
a
month
of
time
frame
or
something
like
that.
Definitely
before
when
cube
com
comes
alive
shin
do
you
have
anything
to
add.
On
top
of
that,.
B
Yeah,
so
I
do
have
some
questions
on
some
of
the
sections.
Can
you
move
down
a
little
bit
so
I
so
I
know
that
steven
and
tom
are
working
on
this
session
class.
B
C
B
C
B
Okay,
great,
thank
you.
Okay,
can
you
move
down
a
little
bit?
I
think
there's
a
another
section,
yeah
and
then
another
section
cpt.
A
B
Yeah
cbt
so
cbt,
I
think
I
copied
the
this
is
content
from
so
far.
This
is
from
your
phones
here,
right
copy
this
from
your
dog.
Do
you
have
anything?
Is
this
a
do?
You
have
any
remaining
workers?
Okay,
just
to
copy
this
here.
Is
it?
Do
you
have
anything
that
is
ongoing?
This
is
the
from
the
doctor.
E
B
All
right,
thank
you,
and
so
for
the
backup
repositories.
You
can
go
there
so
see
that
prashanto
is
there.
Can
you
so
be
so
I'm
looking
at
this
right,
so
I
I
think
this
is.
This
looks
a
little
bit
more
like
slides.
So,
like
the
the,
why
and
objective
section
can
you
maybe
add
a
little
more
make
make
it
more
consistent,
like
the
other
parts
of
the
dog?
Just
maybe
add
it
a
little
bit
more.
I
just
this
two
section
looks
yeah.
F
Basically,
you're
saying
just
to
align
it
with
the
flow
in
a
way
right.
F
This
was
more
around,
you
know
what
should
the
backup
repository
be
capable
of,
and
you
know,
kind
of,
if
you
see
the
objective
below
you
know
the
requirements
for
creating
a
backup,
repository
and
the
kind
of
guidance
in
terms
of
which
one
to
use
when
again,
it's
a
little
high
level
at
the
moment,
the
overall
consensus
was
that
you
know
once
we
kind
of
have
the
rest
of
the
document
flowing
we'll
kind
of
touch
up
on
the
depositories
a
little
bit
more
right.
Now,
it's
getting
at
least
gold
down
shin
or
chanting.
F
B
A
It's
it's.
It
is
actually
stephen
to
answer
your
question.
It's
the
storage
system
itself
will
sit
outside
but
we'll
likely
sit
outside
the
kubernetes
council,
as
you
said,
but
we
do
want
to
have
a
kubernetes
representation
of
those
those
artifacts
right,
for
example,
the
ongoing
quality
effort.
It's
part
of
that.
G
A
G
F
B
Yeah
so
stephen
everyone
else.
If
you
have
any
comments,
please
add
it
here.
So
presenter
can
address
your
comment.
B
And
then
the
the
section
about
rolling
popular
so
so
then
I
I
think
I
found
you
have
a
paragraph
from
an
older
doctor.
So
I
copied
to
here
is
this:
is
this?
Is.
B
That's
fine,
if
you
know,
if
you
think
this
is
fine.
H
A
H
B
We
are
so
I
think
here
we
are
really
talking
about
what
we
need
right.
We
don't
we
don't
really
go
to
the
design
so,
like
whatever
you
have
in
your
cap,
we
don't
really
need
the
details
of
that
api
design.
Basically
we're
just
talking
about
why
we
need
this
like
the
functionality.
What
what
can
you
achieve
so,
but
probably
you
can
probably
expand
more
a
little
bit
more
yeah.
It
looks
a
little
short
compared
to
other
others.
H
Yeah
yeah,
I
I
I
wrote
this
like
six
months
ago
and
then
it
disappeared
and
I
couldn't
find
it
to
add
more
to
the
we
could.
We
could
expand
this
for
sure.
C
H
Well
it
it
would
be
a
subsection
of
like
an
actual
volume,
backup
api,
because
it
would,
it
would
mean
it
would
be
meant
to
fill
in
the
restore
part
of
any
sort
of
a
backup,
workflow
or
at
least
to
be
the
glue.
You
can.
B
H
B
Okay,
yeah,
maybe
if
you
can
just
take
a
look
at
the
one
backup
section
so
so
sean
chen,
is
anyone
going
to
modif
update?
That
section?
Is
that
all
set?
I
think
that
section.
Actually
it's
pretty
well
written
already.
I
don't
know
if
you
have
someone
updating
this
yeah.
A
I
do
I
do.
I
took
a
look
at
the
worm,
backup
section
actually,
the
other
day
I'll
ask
noah
to
take
another
look,
but
I
do
feel
we
want
to
probably
shorten
this
for
a
little
bit.
This
is
too
long.
B
A
Yeah
because,
okay,
I
need
to
switch
between
my
my
own
account
to
do
that.
Don't
worry
I'll!
Take
care
of
that.
B
Okay,
so,
okay,
now
can
we
go
to
the
end?
Just
I
just
want
to
see
like
what
we
want
to
include
right.
Can
we
just
take
a
quick
look
and
see
if
we
can
make
a
decision?
Of
course
we
we're
not
going
to
include
the
pendings,
I
think,
but
the
section.
So
I
think
this
one
we
are
not
quite
ready,
yet
the
application
backup
and
the
last
section
before
appendix
do.
You
think.
J
B
A
A
A
I
I
think,
would
better
have
it,
because
we
do
talk
about
a
lot
of
the
use
cases
in
the
application
level.
B
B
G
I
I
was
just
scanning
through
it,
so
so
the
use
cases
is
more
again
hey.
This
is
what
needs
to
be
done
that
that
workflow
and
I
think
we
could
clean
it
up,
but
the
workflow
does
read
more
as
look
like.
Here's
almost
the
step-by-step.
You
know
actions
right,
so
it
is
a
bit
different
than
the
others.
So
it's
a
really
useful
part,
but
but
yeah
I
I
I
would
tend
to
agree.
G
We
could
probably
probably
could
use
a
an
edit,
but
but
I
like
the
idea
behind
it
because
it
it
doesn't
keep.
G
E
K
B
G
B
Okay,
yeah
that
no
problem
yeah,
that
maybe
we
just
cleaned
up
and
then
yeah
include
that
high.
A
A
B
A
B
C
I
did
them,
I
think
there
might
be
some
useful
content
to
add
to
the
appendix,
because
a
lot
of
the
what
we
mentioned
in
the
document
is
at
various
stages
of
progress.
Okay,
and
so
I
was
thinking,
maybe
for
all
the
things
that
are
we're
putting
into
kubernetes
or
proposals
for
kubernetes.
You
know
from
cozy
to
the
popular
stuff
to
cbt
to
container
notifier.
C
B
Those,
I
think,
probably
we
can
probably
add
that
at
the
end
of
like,
for
example,
in
the
section
about
the
class
right
at
the
end,
we
can
probably
just
add
a
link
to
the
cap.
I,
instead
of
going
through
details
of
that.
B
Something
like
that,
you
know
like
a
copy
submit,
submitted
or
something
some
of
those
like,
for
example,
bans.
That's
already
merged
right.
So
so
we
can
just
include
the
link
of
the
merge
cap.
Others
are
submitted
cap,
we
can
add
a
link
there.
This
is
reviewing
progress,
I
think
that's
yeah,
and
then
we
don't
have
to
describe
what
the
cap
is,
because
that's
already
included
in
that
tab.
B
A
Yeah,
I
would
really
appreciate
like
what
she
was
saying.
Some
of
the
sections
stay
in
a
very
high
level.
We
describe
the
problem,
we
explain
why
and
roughly
describes
the
approach
and
details
can
always
be
found
in
a
separate
link
or
in
the
cap
link,
make
sense.
A
C
Question
to
this
sorry,
for.
B
B
See
in
cbt,
I
think
it's
fun.
Are
you
working
on
cap?
We
know,
or
are
you
ready
to
work
online?
We
have
a
google.
D
D
Yeah
yeah
yeah.
I
think
we
should
start
it.
I
haven't
started
yet,
but
I
think
we
should
start
it
with
this.
Sometimes
this
month,
hopefully
yeah.
B
A
B
Can
you
know
what
you
can
start
in
the
google
doc
actually
just
to
turn
your
google
doc
into
a
format
of
a
cap
first
and
then
convert
it,
because
there
are
some
other
other
things
like.
I
think
you
don't
have
to
go
to
go
through
the
details
like
for
some
production.
At
this
view,
I
think
that's
just
too
early
to
get
to
that
part.
So
maybe
just
just
the
content,
the
main
content
of
the
cap.
You
can
actually
get
already
in
the
google
doc
first
and
then
right
right
and
you
can
add
format.
A
You
don't
need
to
understand
all
this
markup
file
like
formats
etc,
unless.
H
A
Comfortable
with
it,
of
course,
a
general
question
to
this
group
over
here,
I
find
it
interesting
that
we
tend
to
not
use
a
lot
of
diagrams
in
those
write-ups.
Do
you
guys
feel
this
good
idea
to
include
some
of
our
diagrams
in
this?
I
think.
B
A
B
Actually,
we
should
at
least
have
one
diagram.
We
could
have
more
if
we
want,
but
I
think
we
talked
about
this
a
little
bit.
Remember.
We
have
that
diagram
that
shows
the
workflow,
the
you
know
the
respect
and
restore.
I
think
tom
has
some
comments
right.
I
think
you
actually
said
you're
going
to.
B
A
Find
any
examples
in
kubernetes
community
from
the
official.
A
Lot
of
like
the
documents.
E
B
I
know
that
in
cncs,
like
cncf,
like
the
storage
landscape
by
paper,
we
do
have
diagrams.
So
I
think
it's
fine,
okay.
D
A
B
A
Of
course
it
has
one.
Yes,
I
saw
it
and
the
snapshot
web
cook
has
one.
A
B
A
Yeah,
basically
I'll
just
take
a
very
quick
two
minutes
overview.
The
issue
was
right.
Now
the
one
snapshot
resource
refers
to
the
source,
pvc
by
a
namespace
name.
A
A
A
So
this
is
gonna
be
an
api
change,
but
I
have
questions
for
for
this
issue
itself
as
well.
Let
me
just
close
my
questions
out.
Firstly,
even.
A
Even
though
they
it
might,
it
ends
up
pointing
to
the
pvc
by
name
space,
the
name.
It
doesn't
mean
that
there's
no
way
we
can
figure
out
the
original
mapping
right.
So
this
work
this
statement
over
here.
I
don't
think
it
states,
because
the
pv,
the
snapshot
content,
still
carries
the
volume
id
right
and
that
can
be
mapped
against
the
pvc's
corresponding
pvs
wire
right.
That,
of
course,
needs
the
user
to
have
the
corresponding
privilege
to
see
the
worm,
snapshot,
content
and
pv
resource.
That's
one.
B
But
then
you
wouldn't
can't
justify
looking
at
those
two
objects
and
figure
that
out
right.
You
do
need
to
you.
Do
need
to
carry
your
storage
to
find
that
out.
So
I
can
as
that's
a
challenge.
B
B
B
A
A
B
A
B
A
That's
you
can
you
you
can
still
achieve
that,
whether
it's
convenient
now,
I'm
not
saying
that
it's
convenient
or
the
right
thing
to
do,
but
just
saying
that
there
are
still
ways
you
can
achieve
the
goal
here.
So
my
question
is
like
what
do
we
want
to
achieve
over
here
and
what's
without
the
I'm
going
to
go
ahead,
the
uid
over
there
yeah.
B
M
B
M
B
Yeah,
but
I
think
that
the
problem
is
actually
that
they
are
not
for
for
a
lot
of
storage
systems,
so
I
think
I
think
they
were
trying
to
show
that
I
think
humboldt
was
looks
like
he
can't
talk.
B
It's
just
a
listen.
He
wants
to
listen,
oh
okay,
so
basically
it's
like
a
ui
in
the
ui.
They
want
to
list
all
the
snapshots
that
is
created
for
one
pvc,
so
humble.
Does
it
work?
Did
you
hear
what
we're
talking
about?
If
you
just
look
at
the
volume,
if
you
just
look
at
the
wedding
handle
from
the
pv
okay,
so
he
says
yes,.
A
Yeah,
let
me
let
me
try
to
explain
why
I'm
a
little
concerned
about
putting
uid
there.
Now,
if
you
think
about
a
case,
let's
say
a
pvc
being
deleted,
a
snapshot
created.
Oh
sorry,
snapshot
created
on
the
pvc
did
then
recreate
if
the
pvc
happened
to
point
to
the
original
volume
original
pv
that
the
previous
pvc
points
to
with
the
same
name
of
course,
then,
basically
now
the
snapshot
points
to
some
uid
that
doesn't
match
the
current
pvc's
uid.
However,
it
is
actually
created
from
the
underlying
way.
B
A
You
won't
screw
up
by
still
misleading
information
right.
That's
true.
K
I
have
a
question
yeah:
could
we
store
the
resource
version
of
the
pvc
in
addition
to
namespace
and
name.
A
K
A
M
The
resource
version
goes
up
because
it's
based
on
the
xcd
version,
so
it
would
go
up,
but
the
fact
that
it
goes
up
or
not
it
changes.
That's
what
matters
and.
M
M
Like
the
use
case,
is
this
the
use
case
is
that
we're
trying
to
build
this
natural
catalog
right
based
on
pvc
and
snapshot
objects?
I
guess
this
is
a
use
case
right
you're,
trying
to
using
the
in
a
way
we're
building
we're
trying
to
look
by
quoting
like
the
pvc
objects
and
snapshot
objects,
we're
trying
to
find
the
catalog
of
all
the
snapshots
for
a
given
volume
right
and
this
catalog
is
not
something
you
know
external
to
kubernetes,
we're
trying
to
leverage
existing
currencies.
M
Api
objects
for
that
purpose,
and
just
by
going
by
the
name
of
the
pvc
and
the
snapshot
mapping
is
not
sufficient.
So
that's
the
problem
we're
trying
to
solve
yeah.
I
guess
see.
Humbleness.
B
It's
so
if
you
just
look
at
the
well,
so
it's
still
not
sufficient
right.
Basically,
we
do
have
this
problem.
You
can
actually
use
that
static
binding.
So
that's
a
problem.
So
if
you
just
look
at
the
name
and
look
at
the
volume
handle,
that's
still
not
enough.
B
L
What
why
would
it
not
be
okay,.
L
B
I
think
I
so
the
uid.
B
Actually,
I
think
it
does
make
it
more
unique.
I
think
the
steady
case
well
so
because
uid
will
change
will
be
different.
Every
time
you
create
a
pvc.
So
what
you
are
saying
is
it's
still.
It
is
still
a
snapshot
for
that
particular
for
the
for
the
particular
warning
handle,
but
it
is
actually
no
longer
the
pvc
but
the,
but
that
is
fine
right.
We
were
trying
to
find
the
perfect
match
between
the
original
pvc
and
volume
snapshot.
B
Am
I
missing
something
so
right?
So
it
is
true
that
if
you
do
static
binding
your
snapshot
is
it's
your
snapshot
of
that
original
volume
handle?
However,
it's
no,
not
the.
B
Not
the
volume
snapshot
of
the
origin
of
the
new
pvc
anymore.
That's
what
I'm
just
trying
to
say.
So
I
think
the
id
still
still
adds
this
uniqueness
to
it,
which.
A
My
argument
here
is
that,
even
though
we
put
the
uid
that
cannot
be
reliably
used,
excuse
me
to
associate
a
snapshot
with
a
particular
volume.
B
B
B
Right,
so
if
you
have
the
uid
filter,
then
it
will
always
point
to
the
original
pvc.
So
now,
even
if
you
deleted
original
pvc,
you
create
a
new
pvc
to
find
the
the
existing
pv.
That's
a
new
pvc
which
has
a
different
uuid
which
should
not
be
connected
to
this,
only
snapshot
that
was
created
earlier.
I
guess
that's
the!
I
think
that's
what
he
was
trying
to
do.
M
M
Again,
two
scenarios:
the
scenarios
where
the
snapshot
and
the
volume
are
decoupled
from
each
other-
and
I
think
it's
perfectly
okay
for
a
snapshot-
object
to
point
to
a
pvc
that
no
longer
exists
in
the
case
where
you
know
some
snapshots,
part
of
the
volume
you
cannot
delete
the
volume
without
or
yeah.
You
cannot
delete
the
volume
so
yeah.
If
you
do
the
volume,
basically,
you
lose
all
the
snapshots.
M
So
for
that
case
again,
we
are
not
preventing
anything
on
the
look
on
the
kubernetes
side,
we're
not
doing
any
any
preventions
there.
It's
just
a
storage
driver
that
enforces
the
right
behavior
right
for
that
scenario,.
M
We're
covering,
as
far
as
like
the
pvc,
two
snapshot,
mappings
and
and
and
they
can
so
so
like
basically,
I
mean,
shall
we
walk
through
all
the
scenarios.
I
think
this
maybe.
B
Yeah,
so
I
think
he
I
think
he
was
just
trying
to
find
the
proper
pvc
to
run
a
snapshot
mapping
whatever
it's
original
one,
that's
what
he
was
looking
for.
It's
not
really
looking
at
the
world
into
snapchat
and
he's
looking
at
the
warden's
snapshot
object
to
pvc
object.
Mapping.
I
think
that's,
I
guess
that's
the
thing.
L
But
yeah,
if
that's
just
that's
the
goal
right
like
what
shenzhen
said,
you
can
go
to
the
volume
sensor,
content
and
figure
out.
If
this
snapshot
is
the
point,
but.
B
I
B
M
B
B
He's
just
typing
the
chat,
he
has
a
problem.
He
said
not
use
that
moment
for
charging
it's
just
for
even
the
program
having.
Basically
it
is
the
wrong
map
in
that
I
agree
yeah,
but
I
think
that
we
actually,
we
actually
do
have
other
problems
with.
Currently,
this
is
a
model.
This
independent
pvc
anymore
snapshot.
G
B
I
think
he
was
trying
to
show
he
was
trying
he
was
trying
to
he
want.
They
won't
show
that
in
the
ui.
I
think
his
use
case.
I
remember
he
said
it's
a
ui.
You
want
to
show
display
that
in
ui
this
is
pvc.
This
is
your
one
snapshot
so
that
he
wants.
They
want
the
original
one.
That's
that's
what
they
want.
I
guess,
but
I
guess
yeah.
Okay,.
G
B
I
think
for
from
the
from
the
back
of
that
vendor
side,
I
think
normally
you
don't
need
that,
because
you
know
you
probably
just
delete
it
anyway
right,
but
I
think
this
is
for
I
think
he
says
openshift,
ui,
yeah
and
and
because
I.
G
J
B
B
A
M
Now
because,
because
you
know,
customer
can
delete
the
pvc
and
the
corresponding
pv
and
recreate
one
with
a
similar
name.
I
guess
the
names
would
change.
The
pv
names
will
change
because
of
the
way
csi
assigns
random
names
to
the
pvs.
But
I
don't
know
if
you
can
rely
on
that.
Behavior
yeah.
B
Pv
also
yeah
pv
can
also
change.
That's
true
pv.
You
can
delete
that
and
you
can
still
keep
that
you
can
keep
that
warning.
That's
still
possible
as
well,
so
it's
not
okay,
so
he's
a
real
mapping,
but
so
humble,
but
what's
the
purpose,
I
guess
the
people
here
are
asking
like.
What's
the
purpose
of
showing
the
relationship
between
the
original
pvc
and
the
only
snapshot.
B
B
J
I
M
B
M
G
G
If
you're
creating
this
now
right,
we
we
do
have
a
product
that,
like
indexes
even
just
like
regular
aws
snapshots,
and
we
have
some
of
the
same
issues
when
we
encounter
these
sorts
of
sort
of
disruptive
renames
and
basically
what
we
tell
customers
is.
You
know
when
we're
indexing,
pre-existing
snapshots,
best
effort,
and
you
know,
oh
okay,
so
you
do
that.
Okay,
yeah,
that's
outside
of
kubernetes,
but
inside
kubernetes,
because
we
create
them.
It's
a
lot
easier.
But
but
so
I
empathize
with
the
problem
of
trying
to
look
at
stuff
that
you
didn't
create.
G
I
think
the
vmware
guys,
I
don't
know
if
dave
is
dave
is
on,
but
I
know
vmware
had
similar
issues
when
people
tried
to
sort
of
index
vm
snapshots
for
the
for
the
same
thing.
It's
it's
tricky
coming
in
afterwards
for
sure.
B
G
B
B
I
don't
know
if,
but
if
this
is
all
just
handled
by
the
backup
software
itself,
then
I
think
it
probably
knows
how
to
what
to
do
with
it
right.
I
think
it's
getting
more
challenging
if,
like,
if
user,
also
go,
try
to
delete
them
manually
or
something
then
then
things
will
get
messed
up
so
so
maybe
this
is
actually
not
really
a
problem
for
the
for
backup
vendors.
Since
it's,
if
it's
all
managed
by
the
same
software,
then
you
should
probably
know
what
it.
G
Is
yeah
this
this
does
seem
like,
but
I
do
get
humble's
point
that
so
so
you
have
a
backup
admin
that
knows
stuff,
but
but
you
could
have
an
again,
maybe
maybe
you're
more
of
your
kubernetes
admin,
your
cloud
admin,
someone
like
that
says
what
the
heck
is
all
this
stuff
right
and
and
to
this
filtering
point
you
can
kind
of
look
and
say:
okay.
Well,
all
these
snapshots
seem
to
be
tied
to
this
pvc.
That
seems
to
be
tied
to
this
app.
G
Let
me
go
call
this
app
admin
and
ask
him
what
in
god's
name,
he
thinks
he
or
she
thinks
they're
doing
so
so
I
mean
I
don't
know
if
that's
exactly
the
use
case,
but
but
I
could.
I
could
certainly
see
that
so
it's
separate
from
what
we
in
the
backup
space
do,
but
it
definitely
makes
sense
in
terms
of
a
snapshot,
management,
space
and
a
resource
management.
B
A
Cool:
let's
continue
the
discussion
in
the
github
link.
A
If
it,
let
me
move
to
the
next
agenda,
we
are
like
10
minutes
away
from
today's
meetings,
so
somebody
passed
the
link
over
there
got
the
storage
day
right
shouldn't.
Do
you
know,
is
that
you.
C
Quick
plug
if
you're
attending
kubecon
feel
free
to
submit
some
talks
or
if
you
want
to
speak
at
cmdm
today,
it's
a
you
know
a
collocated
event.
I
think
it's
on
tuesday
before
kubecon
and
so
feel
free
to
submit
or
attend.
If
you're
curious
as
well.
I
think
it's
a
good
is.
C
That's
true
yeah,
so
let
me
let
me
post
an
update.
I'll
just
add
it
to
the
dock.
Welcome
to
it,
but.
B
B
C
Yeah,
so
this
is
a
good
group
of
community
members.
I
think
it's
really
relevant
for
this
group.
Definitely.
C
F
I
may
be
going,
but
you
know
just
like
tom
said:
100,
not
20
sure
so
we'll
play
it
by
the
year.
Yeah.
A
Sinai
seriously,
considering
if
we
go
there,
it
will
definitely
cause
some,
maybe
gathering
up
within
this
working
group.
I
can't
wait
to
see
everybody's
face.
H
G
One
of
the
one
of
the
people
at
aws
asked
if
if,
if
they
wanted
to
hear
us,
if
they,
if
we
were
interested
in
hearing
about
their
ack
stuff-
and
I
don't
I
don't
know
if
people
here
care
or
want
to.
G
Aws
crap.
F
G
So
so
it's
the
intent.
The
interest
on
the
backup
side
would
be.
You
know
this
would
be
a
way
to
say
if
a
customer
is
using,
let's
say
an
rds
database
as
part
of
their
cube
application
that
we
that
you
could
discover
that
they're
using
the
rds
and
then
potentially
even
sort
of
trigger
an
rds
snapshot
if
you
wanted
to
so
so
so,
there's
there's
at
least
some
interest
yeah
again,
I
I
I
didn't
know
if
it
was
a
general
interest
to
people
in
the
group.
It's
interesting.
A
Stephen,
this
is
actually
interesting.
Do
you
mind
at
the
terminology
into
the
dock?
Because
back
in
the
days
one
thing
we
were
discussing
is
how
do
we
do
backups
for
cloud
providers
or
service
provider
provided
services?
A
They
are
likely
the
they
likely
will
leave
outside
of
the
kubernetes
kubernetes
cluster,
but
barely
just
expose
the
service
in
the
cluster.
I
believe
back
then
or
or
discussion
ended,
because
this
is
too
complicated
at
that
moment,.
A
Maybe
potentially
you
know
using
some
hooks
to
trigger
those
external
snapshots
or
backups,
but
putting
the
reference
of
those
external
snapshots
and
backups
within
of
within
the
kubernetes
backup
so
that
it
can
later
on
be
referred.
Yes,.
A
All
right,
that's
all
for
today,
then
I
went
to
between
20
minutes
only
returned
four
thanks.
All
thank.