►
From YouTube: Velero Community Meeting - Dec 7, 2022
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).
A
A
The
the
release
notes
can
be
listed
in
this
page
and
for
we
are
also
preparing
1.11,
and
we
have
finished
the
first
round
review
of
the
candidates
and
at
present
the
full
ecos
can
be
viewed
in
this
page
and
we
will
do
another
turn
of
review
and
currently
the
the
the
colleagues
are
doing
investigation
for
the
issues
under
on
due
to
investigation
for
the
for
the
issues
and
after
that
we
will
decide
which
issue
or
which-
and
these
things
will
go
into-
weigh
1.11
based
on
the
according
to
the
issues.
A
Imagination
result
and
efforts,
and-
and
the
other
thing
is
for
the
data
mover.
We
have
finished
the
first
round
discussion
on
the
this
mover
and
the
thing
we
want
to
create
for
the
first
phase
is
it's
called
the
Belarus
building
data
movement
for
CSI
jump
shot,
so
it
means
we
will
focus
on
move
the
data,
CSS
approach
and
and
and
additionally,
we
will.
A
We
want
to
create
Valero
speed
in
data
movement
and
also
support
the
data
movement
plugin
from
the
third
part
party,
vendors
and
I
will
give
give
some
introduction
to
this
share
the
document
to
share
the
document
later
for
the
data
moment
and
for
the
personal
update
myself
is
imagination
around
way,
1.11,
that
is
the
easels
under
my
name,
and
I'm
also
investigating
a
Valero
Beauty
in
data
movement.
Of
course,
yes,
I'm,
proud,
that's
on
my
side
and
she
lives.
C
Yeah
clearly
I'm
working
on
a
plan
of
version
1.11
discussing
with
the
team
like
what
we
need
to
do
for
the
data
mover,
is
that
one
and
also
I'm
investigating
the
1.11
kind
of
issue.
That's
undermining.
D
Yes,
so
the
finally
got
the
AC
plug-in
design
merge
thanks
for
the
reviews
there.
There
was
some
final
comments
near
the
end
about
clarifying
cancel
operation.
D
I
do
intend
to
submit
a
PR
to
clarify
that,
but
since
that's
just
the
documentation,
clarification,
not
functionality,
I
decided
to
go
ahead
and
merge
CPR
so
that
we're
not
blocking
on
that
and
I
will
submit
that
kind
of
minor
doc
change
to
that
design
in
a
separate
PR
later-
and
there
are
these
three
follow-on
PR's,
which
are
the
specifically
API
designs
in
terms
of
you
know,
showing
what
we
need
in
the
photograph
code
or
the
V2
versions
of
the
back
backup
item
action,
restore
item,
action
and
volume.
D
Snapshotter
that
add
the
new
Fields,
as
discussed
in
the
the
previously
approved
design.
So
those
also
need
to
be
reviewed.
Those
should
be
pretty
quick
to
review
because
they're,
basically,
you
know
a
small
subset
of
what
the
overall
design
was.
D
It's
just
specifying
what
the
V2
API
looks
like
for
each
of
those
plug-in
types
and
again
since
when
we
talk
about
the
the
111
timeline
shortly,
you
know
we'll
see
that
it's
a
fairly
short
release
cycle,
so
we
need
to
get
these
designs
reviewed
and
merged
pretty
quickly,
so
I
can
get
going
on
the
implementation.
D
C
Have
a
quick
question:
Scott!
Yes,
I
understand
that
you
need
a
Bia
and
Ria
to
make
sure
the
employment
workflows.
So
do
you
have
any
concrete
use
case?
You
need
the
version
2
of
volume,
snapshotter
plugin.
D
I'm
not
I
mean
the
changes
for
data
mover,
aren't
going
to
be
modifying
that
plug-in,
but
that
did
this.
This
entire
design
started
as
Dave's
update
to
the
volume
snapshotter.
You
know
the
the
idea
for
volume
snapshot,
uploads
and
we'd.
Have
the
you
know
to
Mark
progress
and
my
design
expanded
that
to
include
backup
item
actually
restore
item
action.
So
I,
don't
know
that
I
mean
for
the
data
Movement.
D
We
don't
need
to
actually
create
a
V2
version
of
those
apis,
but
we
Implement
a
plug-in,
but
we
need
the
API
to
be
consistent
so
that
an
updated
to
add
those
new
Fields.
So
you
know
if
at
some
point
we
then
want
to
take,
for
example,
the
AWS
volume
snapshotter
plug-in
and
update
it
to
use
the
asynchronous
actions,
then
the
API
definition
is
there
and
Valero
to
use.
So
so
there's
I,
guess
you
can
think
of
it
in
two
stages.
This
is
just
the
API
change.
D
It's
not
actually
implementing
any
plugins.
So
this
defines
the
V2
API.
It
says
the
V2
exists
for
the
plugins.
We
write
the
adapter
so
that
the
V1
plugins
still
work.
At
this
point,
we've
now
defined
what
V2
looks
like,
but
all
the
old
B1
plugins
still
work.
They
don't
make
use
of
the
new
functionality.
It
will
be
a
later
step
to
actually
Implement
V2
plugins
as
needed.
A
Yeah
for
this
question,
I
think
so
a
technical
technically
speaking,
so
we
can
also
do
the
data
moment
for
the
for
the
examples
created
by
the
volume
some
Charter.
So
we
will
mention
this
in
the
in
the
in
the
in
the
in
this
actually.
So
that
means
in
future.
We
may
also
need
to
do
the
data
movement
for
volume,
Zam
router,
so
maybe
it's
a
good
thing
to
alert
it
to
allocate
support
the
way
to
version
API
right.
D
D
D
We
don't
have
any
yet
that's
okay
and
the
idea
is
to
build
the
Tooling
in
the
Valero
to
be
ready
to
take
it,
so
that
if
someone
then
wants
to
write
that,
then
it's
everything's
right
ready,
but
because
the
design
you
know
the
the
changes,
for
example
to
back
up
to
you,
know,
call
the
plugins
and
then
collect
those
operation
IDs
and
then
go
back
when
it's
in
the
updating
stages.
D
That
will
apply
to
both
backup
item
actions
and
volume
snapshotters
for
any
plugins.
That
return
those
operation
IDs.
C
So
you're
saying
for
the
delivery.
The
reason
I
asked
this
question
is
I'm
thinking
about
the
delivery
of
1.11
I'm
thinking.
If
it's
do,
we
consider
it
a
successful
release
if
we
have
the
backup
controller
support
the
Bia,
for
example,
but
not
the
volume
snapshotter
I
think
that
would
that
would
unblock
us
from
at
least
the
CSI
data
movement
right.
D
Right
yeah
I
mean
the
the
API
update
is
separate
from
then
the
modifying
the
controller
to
use
it
so
I
mean
we
could
probably
implement
this
without
that.
But
you
know
I
think
the
API
part
is
the
easy
part,
because
that's
just
implementing
you
know.
What's
in
the
design
already,
when
we
get
to
modifying
the
controller,
it
might
make
sense
to
do
it
in
two
stages.
D
In
the
first
phase
only
adds
the
support
for
backup,
item
action
and,
for
example,
not
for
volume
snapshotter
yet
and
then
a
follow
one
which
might
not
be
a
111
could
add
it
for
volume
snapshotter.
D
E
Yeah
I'm
also
do
some
education
on
the
line
11
candidate
issue
that
assigned
to
me.
A
Okay,
thanks
and
I
mean.
F
A
If
not,
we
can
go
to
the
discussion
topics.
The
first
one
is.
It
is
added
by
Arlington
and
first
we
want
to
notify
us
that
we
will
be
on
vacation
on
the
10,
20s
and
27.
So
the
committee
meeting
on
27th
will
be
escaped
I,
think
that
is
for
the
U.S
and
Europe
time
zone
and
he's
asking
about
the
20th
community
meeting,
that
is
on
the
China
time
zone.
D
We
talked
about
this
Orlin
brought
it
up
last
week,
specifically
the
North
American
127th.
We
agreed
we're
going
to
skip
that
meeting
because
most
of
the
people
that
would
be
normally
participating
in
that
meeting
are
on
holiday.
Then
he
wasn't
sure
about
the
the
meeting
on
the
20th.
You
know
were
people
going
to
be
on
holiday,
then
for
the
end
of
the
year,
or
would
people
still
going
to
be?
You
know
working
and
basically
that
question
was
to
leave
it
up
to
this
meeting
to
decide.
C
Yeah,
so
what
do
you
think?
Will
you
guys
be
because
I'm
sure
not
everyone
in
Beijing
will
be
taking
time
off.
D
So
so
the
20th
is
not
is
is
not
on
an
official
holiday,
but
you
know
a
lot
of
people
will
be
taking
off
early
four
holidays,
so
I
expect
I
will
be
available
on
the
20th
I,
don't
know
about
others
at
Red,
Hat
I
think
some
will
not
probably
so
I'm
fine
with
having
the
meeting
I
I'm
available.
But
if,
if
it's
excited
that
we
want
to
skip
it
because
there
won't
be
raining
people
there,
that's
fine
too
I
I
can
go
either
way
on
that.
C
Yeah,
so
personally,
I
think
we
can
continue
with
that
as
long
as
you
can
be
joining,
so
we
can
quickly
go
through
some
items
if
any
yeah.
F
A
And
the
Indiana,
the
other
item
from
Orleans
like
I'm,
not
sure
what
had
been
discussed
for
this
PR
and
on
the
last
community
meeting,
but
only
asking
us
to
discuss
its
PR.
We'll
have
any
yeah.
D
Yeah,
so
so
the
reason,
basically,
he
was
on
the
call
last
week
he's
in
European
time
zone.
So
this
meeting
is
not
at
all
convenient
for
him.
Unfortunately,
last
week
none
of
the
people
that
had
participated
in
the
pr
were
here
because
it
was
also
because
shubham
was
also
still
out,
he'll
be
back
next
week,
but
and
basically
there
was
some
back
and
forth
on
this.
That
I
think
the
way
that,
right
now,
the
restore
PV
is
only
active
relating
to
the
Native
snapshots
and
csis.
D
The
CSI
plug-in
does
not
I
use
it
so
that,
if,
if
you
say,
restore
PV
is
false,
for
example,
and
you're
using
CSI,
it
still
attempts
to
use
them
and
I.
Think
his
point
in
here
in
the
pr
was
that
the
the
backup
volumes
field
is
used
to
I.
Think
it's
called
the
snapshot.
It's
mentioned
in
the
pr
I
forget
the
name.
There's
there's
a
backup
field
that
specifies
you
know
whether
we're
taking
snapshots
I
think.
Maybe
it's
just
called
snapshot
volumes,
but
anyways.
D
That
field
is
checked
by
both
CSI
and
by
the
the
volume
snapshotter
and
so
the
concern.
The
thing
you
have
with
this
PR
was
the
the
backup
with
the
restore
Behavior
weren't
consistent
here,
and
so
it
made
sense
from
his
point
of
view
for
his
use
case
to
to
use
this
flag
to
say,
restore
preview
is
false,
which
means
we
won't
restore
either
CSI
snapshots
or
volume
snapshot
or
snapshots
and
I.
D
Think
there
was
some
comments
in
the
review
from
people
saying
that
they
weren't
sure
this
was
needed
and
I
guess
the
the
question
I
had
for
him
and
I
guess
for
people
here
as
well
or
basically
anyone
that's
reviewed
this.
If
we
make
this
this
change,
it
sounds
like
that
helps
his
use
case
where
he
wants
to
have
this
flag
to
be
able
to,
for
because
he's
using
CSI
for
snapshots
to
say
no
I
don't
want
to
restore
the
volume.
D
So
I
said
you
know,
restore
PV
is
false
and
then
the
CSI
plug-in
won't
act
on
that.
Is
there
any
downside
to
doing
that
is?
Is
there
a
use
case
that
that's
better
if
the
CSI
plug-in
doesn't
follow
that
it
seems
to
me
that
probably
having
the
CSI
plugin
use
that
flag
to
not,
then
restore
solves
at
least
one
use
case
and
I?
Don't
think
it
breaks
any
of
these
cases.
But
that's
the
question.
We
need
to
I
guess
figure
out.
C
Yeah
there
may
be
some
break
changes
like
there
may
be
people
using
the
flag
and
expecting
the
CSI
snapshot
still
restored,
but
I
think
his
proposal
here
in
this
PR
is
more
consistent
to
me,
because
this
field-
I,
don't
know
the
history
about
Valero.
But
when
these
fields
are
were
added,
I,
don't
think
there's
a
CSI
plug-in.
D
Right
we
didn't
have
CSI,
so
when
CSI
plugin
was
added,
it
sounds
like
on
the
backup
side
we
took,
you
know
we
took
into
account
the
restore.
There
was
another
flag,
that's
mentioned
in
one
of
the
comments
here,
but
on
the
restore
side,
the
plug-in
didn't
so
I.
From
my
point
of
view,
I
think
what
he's
doing
in
this
PR
is
probably
reasonable.
D
Know
shubham
was
one
of
the
ones
that
did
look
at
this
before
so
I
can
talk
to
shubham
about
it
as
well
when
he
gets
back
next
week.
He's
he's
still
on
holiday
one
more
week
this
week,
but.
D
He
was
off
the
end
of
last
month
and
the
beginning
of
this
month.
I
believe
he's
back
next
week
on
Monday,
so,
okay,
the
the
12th
I
think.
C
Okay,
I
think
shun
also
has
some
common
here.
So
what
do
you
think
should.
B
I
think
I
I
think
I
I'm
honestly
paid
with
you
and
I
agree.
This
change
will
make
it
more
consistent
with
the
previous
Behavior,
but
I
I
found
to
have
some
different
thought.
So
yeah
yeah.
D
B
D
Mean
obviously,
since
specifically
had
some
concerns,
we'll
need
to
wait
till
he
gets
back
to
sort
of
see
if
you
know
he's
okay
with
this
now
or,
if
not,
you
know,
propose
an
alternative,
because
I
think
either
way.
I
think
you
know
there's
a
use
case
here
that
needs
to
be
resolved
and
I
think
the
easiest
way
to
resolve
it
is
this
way
using
this
existing
flag.
D
But
if
that's
not
going
to
resolve
it,
then
we
may
need
a
new
flag,
but
that's
going
to
be
more
confusing,
so
I
think
that's
the
question
we
think
I
could
talk
through
with
Shabbat
when
it
gets
back
as
well.
Yes,
I
mean
you
know
one
way
or
another.
Here's
a
real
real
world
use
case
that
a
user
needs
a
way
to
do
this
and
I
I
think
doing
it
using
the
existing
flag
rather
than
adding
to
the
API.
It's
probably
the
most
straightforward
way
of
doing
it
right.
C
A
And
they
have
so
the
other
one
is
from
myself
for
the
introduction
of
the
doc
and
Daniel.
Would
you
like
to
go
first
because
the
main
lines
may
take
some
time.
A
No
I
think
we
can.
We
can
go
through
the
goals
and
and
to
see
if
everyone
have
some
questions
or
anything
like.
C
That
sure
yeah
I
can
I
can
talk
about
the
timeline
first.
C
Finalizing
the
1.11
Roma
Wiki
I
first
wanna
check
the
timeline
with
Scott.
We
didn't
have
a
chance
to
discuss
with
you,
so
I
I
think
we
will
discussed
internally
and
be
aware
about
the
timeline
for
everyone,
11,
with
the
consensus
that
we
want.
This
build
shorter
release,
making
sure
the
support,
progress,
monitoring,
be
a
part
of
the
major
features.
Another
area
we
want
to
make
improve.
C
Is
there
used
to
be
several
requirements
to
add
more
flexibility
in
terms
of
the
filters
in
backup
CR
like
the
classical
resource,
include
excluding
the
conditional
backup
of
PVS.
There
has
been
some
in
discussion
internals,
so
so
I
think
we
are
ready
to
deliver
some
of
the
features
to
improve
that
flexibility,
so
sky
on
the
track
with
you
about
the
timeline.
I
know
that
note
that
the
feature
freeze
is.
C
D
Yeah
and
and
just
to
clarify,
then
feature
complete
is
when
we
actually
need
all
the
PRS
that
implement
the
feature,
reviewed
merged
and
everything
and
feature
freezes
just
being
certain.
We
know
which
features
are
in
and
out,
and
this
is
it
obviously
designs
approved.
D
Got
it
yeah,
yeah
and
I
was
talking
to
to
Wes
and
Dylan
about
this
earlier
today,
I
I
think
these
dates
a
reasonable
I,
just
wanna.
What
the
main
caveat
here
is
it's.
It
is
a
fairly
quick
cycle
which
again
we
both
agreed,
you
know
was,
is
always
what
we
want.
It
means
we
do
need
to
be
fairly
diligent
about.
You
know
for
111
feature,
PR's
getting
reviews,
and
you
know
response
to
reviews
and
all
that
in
a
fairly
efficiently,
because
you
know.
C
If
you
have
absolutely
yeah
yeah
I
I
I
have
to
admit
that
that's
not
something
we
from
where
we
did
very
well
in
last
release,
because
people
are
pulled
out
to
in
the
middle
of
the
design.
People
are
pulled
out
to
focus
on
other
product
features,
and
we
we
pause.
The
design
and
the
communication
didn't
went,
go
very
well,
so
I
would
suggest
here.
We
we
work
more
closely
like
there.
C
There
is
quite
Euro
that
you
make
some
update
to
the
pr
and
there's
a
notification
on
GitHub,
but
that
email
was
somehow
ignored.
So
it
would
be
easier.
If
you
can,
you
know
just
pin
us
via
slack.
Both
you
or
shubong
I,
think
that
that.
D
So,
and-
and
the
other
thing
of
course,
is
that
at
this
point
at
least
on
the
on
the
features
that
that
we're
working
on
with
the
item
action
stuff,
the
the
design
part,
which
is
kind
of
that's
the
big
thing
that
we
have
to
agree
on,
you
know
all
ahead-
that's
approved
now
so
I
think
going
forward.
You
know
where
features
for
individual
parts
of
that
design.
So
hopefully
those
will
be
a
little
bit.
You
know
quicker
to
get
reviews,
because
we
don't
really
need
everybody
to
look
at
them.
We
just
need
to
get.
E
D
For
each
PR
and
that's
a
little
more
isolated,
so
as
long
as
you
know,
the
the
people
assigned
to
review
the
PRS
are
available,
and
then
you
know
we
can
communicate
on
slack
to
make
sure
that
that's
happening
and
and
obviously
you
know
for
the
stuff
that
you're
working
on
there.
If
you
need
to
or
me
to
review
again,
you
know
ping
me
as
well.
D
If
I'm
not
responding,
you
know
in
slack
or
whatever,
so
that
we
can
just
kind
of
I'm
just
just
wanting
to
make
sure
that
we're
all
kind
of
in
agreement
here
that
to
make
these
dates
we're
going
to
have
to
be.
You
know,
on
top
of
the
reviews,
as
well
as
the
implementation.
G
D
Especially
because
I
know,
on
the
red
hat
side,
we
will
have
that
week
and
a
half
run
the
end
of
December,
where
we're
not
at
all
in
the
office
and
so.
D
And
and
I'll
I'll
mention
that
I'll
the
next
meeting
as
well
just
to
kind
of
bring
up
those
days
because
there's
that
period
of
time
from
around
the
23rd
of
December
until
the
end
of
the
year,
where
again
I'm
not
going
to
be
in
the
office.
So
you
know
don't
expect
me
to
be
able
to
respond.
You
know
within
that
time
period,
but
other
than
that
I
should
be
pretty
responsive
to
reviews
and
all
that.
C
Yeah
yeah,
so
if
you
need
any
immediate
response,
just
hit
us
on
slash
and
by
the
way
around
the
end
of
January,
it
will
be
the
Chinese
New
Year
in
here.
So
we
will
also
be
you
know,
out
of
office
for
around
one
or
two
weeks
around.
D
D
As
we
get
those
updated
into
those
dates,
in
both
cases
on
the
the
community
meeting
notes,
so
we
know
kind
of
who's
around
when
that'll
help
as
well.
C
So
you
you
think
this
these
days
are
okay,
for
you
guys
to
meet
your
target
day
right,
because
you
you
you
you
you
tell
you
said
that
you
will
follow
up
with
didn't
about
the
the
right.
D
Right
and
then
yeah
I
did
talk
to
Dylan
and
Wes
as
well
today
and
I.
Think
from
our
point
of
view,
these
dates
are
pretty
much.
You
know
what
we
would
have
been
looking
for
so
I
I
say
this.
Is
this?
This
looks
reasonable
to
me
and
Dylan
and
Wes
also
saw
these
states
as
well.
We
talked
about
this
in
a
meeting
earlier
today,
so.
C
Okay,
so
let's
work
towards
these
days,
but
please
also
let
us
know
how
much
flexibility
will
be
allowed.
I
mean,
for
example,
if
we
delay
one
or
two
weeks,
would
that
also
be
okay
or
you
have
a
hard
target
date?
If
you
have
that,
you
can
add
that
to
this
yeah
I.
G
E
C
Yeah
yeah
I
just
want
to
make
sure
that
it
doesn't
impact
your
Target
on
your
you
know,
product
line
or
the
other
schedule
on
the
right
hand
side.
Otherwise
we
can,
you
know,
be
more
flexible
in
terms
of
the
date
I.
C
So
so,
as
for
the
support,
progress,
monitoring
or
shall
I
share,
my
screen.
C
So
the
way
we
organize
the
future
is
the
issues
is
like
before
we
use
the
Milestone
to
track,
what's
really
in
One
release,
so
theoretically,
for
those
issues
that
already
put
in
the
Milestone
we're
gonna
remove
the
1.11
candidates
because
we
already
commit
we're
gonna
handle
that
in
one
level
release.
C
So
currently
we
are
doing
the
evaluation,
as
you
mentioned,
and
we
are
adding
more
issues
into
this
Milestone,
but
for
these
ones,
I
believe
these
are
are
all
related
to
the
on
the
the
progress
monitoring
change,
so
Scotty
they're,
okay,
assign
these
ones
to
you
and
shubham.
That's.
D
Fine,
you
just
you,
can
sign
up
or
just
all
to
me
and
then
I
can
move
them
over
to
ship.
You
know
if
there's
something
he's
working
on,
but
but
yeah.
Basically,
all
of
those
are
ones
that
I
had
created
or
subama
created
to
track
the
individual.
You
know
implementation
tasks
associated
with
that
overall
design.
D
You
know
one
you
know
so
so
there's
one
each
for
the
API
design,
the
API
updates,
rather
for
v2,
and
then
there's
one
for
the
backup,
controller
changes
and
one
for
the
restore
controller
changes
so
and
then
the
item
operations
file.
So
so
those
are
all
the
kind
of
lower
level
tasks
to
implement.
You
know
a
PR
for
one
and
another
and
portions
of
that
design.
A
Then
let
me
start
on
the
data
moment
part.
So
here
I,
let's
just
have
a
view,
have
a
look
at
the
the
goals
and
the
background
and
and
first
and
so
we
have
several
goals
after
the
discussion.
So,
first
of
all,
we
want
to
build
the
Belarus
building
data
moment
modules
and
use
that
module
to
backup
the
CSS
and
charts.
A
So
so
this
means
that
first,
it
means
that
we
will
first
focus
on
the
CSS
SoundCloud,
but
it
doesn't
mean
that
we
we
will
leave
everything
else
behind.
Instead,
we
want
to
create
a
generic
some
generic
modules
data
mode
modules
that
could
be
could
serve
all
biker,
backup
kinds.
For
example.
A
Currently
we
focus
on
CSS
and
plot
backup
and
then
in
future
we
may
be
able
to,
because,
technically
speaking,
when
the
same
product
is
created
by
the
volume
some
shorter,
we
still
be
able
to
move
the
same
chart
and
there's
still
requirement
to
move
that
same
shot
and
and
also
some
winners
or
some
storage
render
or
some
and
some
Cloud
platform.
So
the
supports
the
direct
access
to
the
same
chart,
so
that
will
be
different
in
the
workflow.
A
So
we
want
this
General
State
movement
modules
supports
this
direct
access,
backup
as
well
and
and.
A
And
backup
can
most
of
most
of
the
property?
Yes,
but
I'm,
not
sure
if,
if
we
have
some
other
possibilities
that
we,
for
example,
from
my
perspective,
I
think
the
the
pot
volume
backup
is
either
either
another
backup
kind
too,
because.
A
It
is
doing
the
data
moment
also
so,
but
I
think
yes
for
for
most
of
the
cases
we
need
to
back
up
based
on
Snapchat,
for
a
big.
By
that
we
will
get
better
data
consistency.
C
A
Understand
it
and
the
mix
under
to
be
on
the
same
page
right
now,
most
most
of
in
most
of
cases,
the
Barca
kinda.
It's
it's
a
it's
talking
about
the
different
to
backup
from
the
same
chart
or
from
the
source
yeah.
E
A
For
the
better
world,
let's
try
to
violate
later
yeah.
The
next
one
is
we
want
to
reuse
is
facing
data
path.
That
is
the
file
system
uploader
that
is
based
on
copia
and
the
unified
repository,
and
for
that
we
have
under
the
Banning
phase,
and
the
reason
have
already
been
mentioned
in
the
new
file
repository
design,
yeah
and
the
next
one
is.
We
want
to
reuse
the
existing
without
modules
as
much
as
more
as
possible.
A
This
is
quite
easy
to
to
understand
because
we
will
reuse
the
using
module
well,
while
significantly
reduce
the
efforts
of
the
development
and
reduce
the
complexity
of
the
reliable
architecture
and
modules
and
workflows
right
and
yeah,
and
the
next
one
is.
We
want
to
leave
the
design
space
for
other
types.
A
Besides
the
current
file
system
loader,
for
example,
we
I
think
the
infiltrable
to
the
Block
Level
uploader
right,
because
because
the
comparison
without
file
system,
uploader,
Block
Level
uploader
have
some
some
some
unique
user
cases
and
some
unique
benefits.
A
And
finally,
we
yeah.
A
Yeah,
yes,
something
like
here:
it
is
inside
the
data
path
that
is
to
touch
the
data,
so
so
uploader
is
to
touch
the
data
that
is
to
array
data
from
one
place
to
the
other,
all
or
back
something
like
that
sure
and
yeah.
And
finally,
we
want
to
support
the
plugins
there's
a
moment,
plugins
from
other
vendors
and
in
particular
we
know
that
our
IDP
has
already
has
has
some
existing
at
the
moment
modules.
We
want
to
simulate
integrate
with
that
I.
C
Think
all
this
yes,
but
but
I
think
when
we
talk
about
plugins,
they
may
be
confusing,
because
when
we
talk
about
plugin
in
Valero,
it's
a
narrower
definition
about
plugin.
When
we
talk
about
plugging
in
Valero,
we
think
about
the
grpc
stuff.
So
do
we
want
to
support
the
gipc
stuff
or
we
just
make
the
data
mover
module
pluggable
I
think
that
may
imply
different
things
in
Valero.
A
Yeah
yeah
correct
so
here
I
think
the
the
mini
is
lighter,
so
we
want
to
make
the
vendors
that
are
more
pluggable.
Okay,
let
me
modify
this
like
plug.
A
Okay,
so
so
here
for
the
first,
for
the
first
points
Daniel
you
mentioned,
so
it's
something
about
the
protocol
right.
So
we
have.
We
have
provided
different
ways
for
the
replacement
that
for
the
for
the
for
the
for
the
plug,
but
at
the
moment
to
replace
Belarus
Beauty
in
that
moment,
so
that
will
come
down
to
which
way
we
use
and
that
that
will
use
different
protocol.
A
A
We're
just
to
clarify
Cisco
is
tells
you
to
clarify
how
we're
gonna
going
to
do
that
for
Belarus
building
data
movement.
That
was
the
goals
and
at
the
readout
the
first
discussion.
We
have
discussed
the
modules
or
component
compounds
and
workflows
or
actions
for
the
backup
and
result,
and
so
I
think
these
two
diagrams
that
show
the
backup
and
the
result
and
and
some
I'm
I'm
not
doing
to
go
through
every
details
of
this
talk,
but
I
want
to
emphasize
some
points
of
of.
A
First
of
all,
it's
like,
as
you
can
see,
we
are
trying,
as
just
mentioned
here
for
the
goal.
We
are
trying
to
reuse
everything
we
can
reuse
in
the
in
the
current
in
a
Valero,
the
node
agent,
the
data
box
and
the
CSI
plugin
right
and
for
the,
and
that
is
the
first
point.
A
The
second
point
is
that
we
have
provided
three
different
ways
for
the
replacement
that
is
from
this
from
this
kind
of
specific
module
and
from
this
building
data
more
and
from
end
of
all
the
entire
entire
CSI
or
kind
specific
data
movement,
blockable
modules,
so
different
ways
are
for
different
purpose.
So
we
can
read
a
rhythm
here,
yeah
and
the
the
the
the.
The
third
point
is
like
we
make
from
from
this
building
data
more.
A
We
make
everything
generic,
so
this
means
that
this
Valero
building
didn't
move
a
could
solve
every
kind
of
the
backup
CSI.
Now
we
serve,
we
serve
the
CSI
Central
backup
and
in
future
we
can
solve
any
other
Central
backup
because
the
interface
because
it
is
a
generic
and
because
it
is
generic,
the
interface
is
also
generic.
A
We
don't
want
to
involve
any
kind
of
specific
information
in
this
CR,
because
this
is
the
input
right
and
and
the
and
the
date
part
of
the
data
path
is
the
unique
one
all
across
the
Valero
world.
A
So
this
one
is
even
generally
generic
that
it
doesn't
understand
any
any
any
combinated
information
or
cumulative
thing,
so
the
input
is
even
generic,
for
example,
for
the
file
system,
backup
the
input
is
a
path
to
path
to
the
to
the
root
directory
we
are
going
to
backup
and
for
for
Block
Cloud
backup.
The
input
is,
it's
also
a
path
that
represents
the
the
name
of
the
block
device,
something
like
that
and.
C
I
think
we
may
need
more
discussion
offline,
okay
about
the
placement,
so
I
Believe
by
placement.
You
mean
some
modules
that
can
be
a
replacement
yeah,
sorry,
so
so
so
by
replacement,
you
mean
some
of
the
modules
that
we
provide
as
battery
included
as
modules,
but
these
modules
can
be
replaced
by
third
party
module
right.
That's
what
you
mean
by
replacement
yeah.
C
C
G
C
Probably
enough,
because
in
real
high
use
case,
there's
a
different
part
somehow
by
using
the
whole
sync,
but
we
mean
we
may
need
more
discussion.
Offline
I
understand
the
point
that
there
may
be
a
case
that
the
third
party
wanna
reuse,
the
data
path
in
Valero.
We
use
the
copia
to
move
the
data,
but
I
don't
think
that's
a
very
high
priority.
A
Think
for
the
for
the
for
the
oidp,
you
mentioned
it's
right
to
the
second
point,
to.
E
A
The
the
building
more-
that
is
exactly
the
same
as
what
you
mentioned,
and
the
the
other
two
ones
right
now
we
don't
have,
if
not
on
the
on
the
top
priority,
because
of
runaway
for
the
for
the
way
1.11.
We
don't
need
to
implement
that,
but
the
things
I
want
to
mention
here
is
like
when,
because
we
want
to
keep
this
possibility,
so
we
need
to
make
the
interface
or
protocol
to
make
it
clear
or
make
it
as
a
public
one
right.
A
For
example,
here
we
we've
already
make
it
a
generic
kind
of
public.
We
are
using
the
biav2
or
irv2
for
other
interface.
So,
okay,
we
don't
need
to
worry
about,
and
for
this
one
the
protocol
is
also
this
crd,
so
this
one
has
already
been
implemented
after
we
Implement
I'm,
no
I
mean
the
interface
part
has
already
been
implemented
after
we
have
this
one.
A
So
here
I
just
want
to
say
to
mention
that
we
have
this
these
ways,
but
this
will
not
definitely
not
increase
any
effort
to
weigh
1.11,
because
all
this
and
and
we
have
these
ideas,
we
know
which
one
must
be
a
public
interface
and
then
that
will
guide
our
design.
That
is
for
the
purpose,
but
we
don't
need
to
implement
all
these
three
ways
in
with
1.11.
E
C
B
C
Here
in
this
meeting,
I
think
we
we
may
discuss
offline,
but
I
I
feel
it
will
be
either
that
we
just
provide
one
level
of
a
plugable
module
or
tell
the
third
party
like
this
is
the
module
you
can
replace.
We
will
not.
We
should
probably
avoid
telling
them
you
can
replace
this
level
this
level
and
this
level
at
least
in
one
level,
I,
don't
think
that's
something
we
can
really
agree
on.
Otherwise,
we
will
spend
too
much
time
discussing.
We
will
just
see
okay,
okay,
I.
A
I
basically
agree:
I
agree
with
you
and
let's
see
how
to
how
to
how
to
make
this
in
the
design.
I,
yes,
I,
think
the.
Currently.
We
only
need
to
mentioned
and
Design
This
one
that
is,
but
but
the
other
thing
I
want
to
say
that
indeed,
people
don't
have
any
any
or
we
don't
have.
A
Even
even
we
don't
mention
that
this
is
a
possibility,
so
people
will
use
that
even
don't
mention
that
in
our
dog,
but
it's
okay,
this
one
is
this
is
a
thing
or
other
point,
or
we
want
to
have
the
same
way:
1.11
design,
yeah,
that's
for
the
for
the
for
the
three
points
and
everything's
is
reused
and
which
one
should
be
a
public
interface
and
and
yeah
starting
for
the
come
for
the
point-
and
the
last
point
is
like
here
in
this
diagram.
A
We
we
see
some
I'll,
say:
isolation
like
we
call
these
parts
like
a
CSI
at
the
moment
and
because
it
Parts
like
as
a
building
that
more
and
we
call
this
part
as
a
database.
That
is,
that
is
something
functionality
as
isolation.
A
It
doesn't
mean
that
everything
is
a
real
component,
a
real
components
or
modules,
and
we
also
could
regard
them
as
a
compounds
like
these
ones
and
this
ones
these
ones
are
as
a
compound
compounds
of
the
CSI
data
movement,
and
this
part
are
compounds
of
the
the
the
the
building
did
more.
A
But
yes
again,
it's
come
to
come
down
to
the
the
name
of
the
term,
let's
discuss
and
to
say
to
see
how
we
can
make
it
better
to
understand
in
the
dark,
but
the
functionality
is
is
clear
clarified
here
we
can
check
this
offline
and
the
actions
is
also
clarified
here
for
every
action
I
mentioned
in
this
diagram
and
and
finally,
we
based
on
the
current
result.
A
We
have
listed
some
sub
these
items
because
for
each
of
the
model
we
have,
we
have
to
carry
quite
some
details,
but
this
is
not
a
fan
light
one
or
it
is
even
not
enough.
We
may
add
a
delete
on
modify
in
this
list
and
we
we
now
have
some
open
questions,
but
I
believe
that
we'll
have
more,
we
will
have
more
later.
So,
let's
start
I'll
continue.
A
The
discussion
based
the
the
current
Doc
and
if
we
have
any
questions
we
can
list
it
here
or
we
have,
we
can
add
comments
or
or
we
can
discuss
directly
through
the
slack
yeah,
that
from
my
side
do
we
have
any
instant
questions
right
now.
C
So
Scott,
is
it
you
or
should
be
part
of
all
these
data
mover
design
I
know.
D
Sure
so
Shivam
has
been
more
involved
in
the
Daily
Mirror
side
than
I
have
so
I
was
hoping
when
he
gets
back
next
week.
He
can
take
a
look
at
this
and
provide
some
feedback
there
as
well,
and
he
also
is
the
one
that
submitted
the
design
PR
for
data
movement
in
internal
data
movement
as
well
so
yeah.
C
We're
going
to
combine
the
effort
like
maybe
we
continue
the
discussion
here
and
reflected
in
his
PR
or
one
way
or
another,
or
maybe
merge
some
high
level
PR
and
add
more
detail.
Pr,
we'll
figure
out
which
one.
E
A
A
If
okay,
then
we
can
continue
the
discussion
offline
for
this
and
I
think
we
have
cover
all
the
items
right
today
and
do
we
have
any
further
questions
or
comments
about
today's
meeting.