►
From YouTube: Velero Community Meeting - Dec 3, 2019
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
So,
as
you
might
have
noticed,
we
had
a
little
break
between
the
open
discussions
here
and
the
last
one.
Because
of
cube
con
enforce
cube
con
happened,
the
entire
team
of
there
we
had
a
blast,
we're
going
to
talk
a
little
about
what
we
did
at
UConn
today,
but
also
we
want
to
make
sure
that
we
fill
you
in
on
everything.
That's
going
on
with
the
course
I
see
a
lot
of
community
members
here
today.
Super
super
happy
to
have
you
all
here.
If
you
have
any
questions,
please
just
unmute
yourself
and
speak
up.
A
A
B
Yeah
sure
I
can
hop
in
I.
Don't
really
have
much
in
the
way
status
updates.
You
know
cube
con,
followed
by
the
Thanksgiving
holiday
here
in
the
US,
so
I've
been
sort
of
away
from
the
computer
for
a
while
and
I
know
where
no
one
has
some
discussion.
Topics
on
cube
con,
so
I
won't
go
into
that
too
much
I
guess
the
one
thing
is
I
wanted
to
say
a
big
welcome
to
our
newest
team
member
Ashish
who's
on
the
call
today.
B
A
D
First
I'm
gonna
apologize,
real,
quick
I
did
see
some
chat
last
week
asking
about
the
community
meeting
last
week.
So
I
think
there
was
a
little
communications
mistake
that
a
bunch
of
us
just
took
the
week
off
for
the
Thanksgiving
myself
included
I,
but
I
also
thought
more
of
us
were
gonna,
be
here
so
sorry
to
anyone
who
showed
up
for
the
community
meeting
expecting
it
to
be
happening,
and
it
wasn't,
but
also
just
an
update
from
cube
con.
So
great
meeting
everyone
that
I
was
able
to.
D
D
So
a
piece
of
that
is
the
disaster
recovery
working
group-
that's
spinning
up
with
cig,
apps
I
know
there
was
a
discussion
of
that
yesterday
and
thus
a
gas
meeting
that
I
wasn't
able
to
attend.
But
one
of
the
disaster
recovery
engineers
here
at
VMware,
Dave,
Smith
ochita,
is
working
on
that
and
jingyang
from
six
storage
and
also
I.
Think
I
mentioned
this
before,
but
the
CSI
volume
snapshots
CRTs
and
the
code
has
moved
to
beta
the
common
controller.
That
is
in
that
that
all
CSI
drivers
kind
of
rely
on
that
interfaces
with
kubernetes.
D
That's
getting
split
up
into
a
couple
different
controllers:
it's
pretty
much
the
same
code
but
they're
kind
of
pulling
it
apart.
That's
mostly
done,
but
the
release
will
actually
happen
after
kubernetes
1.17
is
out.
The
plan
is
to
get
that
out
by
the
end
of
the
year,
and
my
plan
is
to
help
get
that
out
thanks
and
my
plan
is
to
help
get
the
tests
updated
over
the
December
and
then
I
put
a
link
to
the
discussion
it.
This
was
during
the
cig
storage,
deep
dive
at
cube,
con
jing
and
create
the
other
gentleman's
name.
D
They
talked
about
the
difference
was
in
the
CR
DS.
The
big
thing
was:
they
went
from
adding
from
basically
updating
the
spec
in
the
CR
DS
to
adding
a
status
field
that
tells
you
what
changed
and
how
the
gives
you
a
better
idea
of
what
bindings
changed
so
we'll
be
looking
at
integrating
that
into
our
CSI
support
and
we're
also
going
to
talk
a
little
bit
more
later
about
where
velaro
goes
with.
D
Our
CSI
support
and
kind
of
Steve
and
I
have
been
talking,
bouncing
some
ideas
back
and
forth
about
what
that
implementation
looks
like
and
I
think
that's
mostly
hit
I
had
some
discussions
about
contributions,
I
don't
know
if
anyone
on
the
calls
from
maya
data,
I
I
had
some
discussions
with
people
there,
but
I
knows
my
data
folks
are
working
on
velaro
and
hoping
to
get
some
of
those
changes
up
streamed.
If
we
can
that's
pretty
much
it
for
me,
I.
F
D
E
D
Yeah
I
think
I
think
you
should
just
be
able
to
join
the
Valero
mailing
lists
and
it
should
automatically
send
you
an
invite,
and
these
should
be
pretty
much
every
week.
December
will
be
strange
because
of
us
holidays
coming
up
in
three
weeks
or
so,
but
after
that
it
should
be
on
a
fairly
regular
cadence
I
can't.
E
So
much
yeah
I'm
kind
of
looking
forward
to
hearing
more
about
discussing
more
on
the
disaster
recovery
and
the
migration
use
cases
using
weather.
Oh
that's
where
we
are
trying
to
see
how
to
fit
in
into
open
ideas.
For
example,
at
my
data,
some
of
the
challenges
that
we
have
are
around
volume
groups
or,
like
consistency,
group
implementation
with
a
little
I,
see
some
Allen
back
up.
Work
has
progressed
along
yeah,
anything
to
contribute
towards
the
those
aspects
or
you
know,
help
code
contributions
or
design
30.
B
Yeah,
it
sounds
great
night.
This
is
Steve
just
jumping
in
I
would
say
you
know
if
there
are
specific
topics
that
you
want
to
cover
feel
free
to
add
something
to
the
discussion
topics
section
and
you
can
do
that
going
forward
or
you
know
we're
happy
to
to
talk
about
any
of
those
topics,
and
we
should
have
some
time
today.
If
you
want
to
jump
into
some
of
that.
E
The
cool
yeah
I'd
probably
briefly
share
like
what
I'm
grappling
with
it
says
it's
now
a
good
time
or
should
I
wait
for.
B
F
A
A
A
All
right
any
any
questions,
so
ELISA,
the
you
did
an
interview
here
with
the
new
stack:
how
how
did
it
go?
How
was
it.
F
F
Basically,
and
then
it
also
was
LA,
which
I
didn't
know,
and
so
I
was
there,
but
I
think
we
cover
a
lot
of
ground
about
Valero
use
cases
and
I
thought
was
really
good
and
it
was
good
because
we
had
three
different
people
myself
as
a
developer
Tom
as
a
product
person
and
and
that
would
be
Tom
Spoonamore
if
I
pronounced
it
correctly,
not
tell
Shay
I'm,
really
sorry
if
I
mispronounced
it
and
he
is
from
Dow.
So
we
had
all
these
different
perspectives
and
there
is
a
cute
photo
really.
A
A
F
I
totally
forgot,
yeah
I
actually
got
very
burnt
out
between
the
conference
and
doing
the
launch
of
that
project
that
week,
but
it
was
really
cool.
So
the
secret
project
is
the
public,
which
is
a
podcast
and
that
week
I
launched
there
with
five
episodes.
So
there
was
a
lot
of
content
being
sent
out,
but
the
cadence
of
the
show
is
one
Episode
a
week
and
Peppa's
are
going
to
drop
on
the
Monday.
So
you
can
start
your
day.
F
Listening
about
listening
to
people
talk
about
cloud
native
and
yes,
so
Jonas
is
putting
up
on
the
screen.
The
hosts
of
the
podcast
we
rotate.
So
usually
we
couldn't
have
at
least
three
of
us
and
we
might
be
interviewing
people
who
are
not
the
host
or
not
really
interviewing
about
talking
to
our
last.
F
The
last
episode
that
we
put
out
was
with
Joe
Beda,
and
it
was
a
really
cool
conversations
conversation
with
we
try
to
ask
questions
that
weren't
being
asked
in
interviews
that
he
has
done
out
there
in
the
next
recording
we
are
going
to
have
it
with
Craig
and
I
could
not
pronounce
his
name
buddy
help
me.
You.
G
F
I
crack
is
awesome
if
anybody
has
so
amazing,
so
I'm
really
excited
to
to
be
talking
to
him
and
oh,
and
if
anyone
has
questions
for
Craig,
there
is
also
hacking
tea.
There.
People
can
have
questions
too
anyway.
That
is
the
secret
that
was
the
secret
project
secret.
No
more
now
get
I'll
grab
the
link
for
the
heck
MC.
H
So
that
was
interesting
and
then
I
was
also
at
the
helm,
boof
a
little
bit
and
got
into
interesting
conversations
about
Valera
there
as
well,
where
people
were
kind
of
talking
about
how
they
can
backup
helmets
and
I
pointed
out
that
there
was
like
a
plugin
related
to
that,
but
not
completely
up-to-date.
But
it's
something
that
we
may
want
to
look
back
into
other
than
that
there
was
this
known.
As
briefly
mentioned,
the
data
protection
working
group
is
starting
to
be
spun
off.
H
That's
going
to
be
a
cross
sig
collaboration
between
sig
Apps
and
six
storage,
so
Shing
and
Dave
Smith
Ishida
from
VMware
who,
who
are
kind
of
leading
this,
had
discussed
this
cig
apps
yesterday
morning
and
then
the
next
step
is
to
start
a
mailing
list.
Take
it
to
six
storage.
On
the
first
day,
at
9:00
a.m.
Pacific
time,
so
if
you're
interested
in
kind
of
learning
more
about
that
group
and
getting
involved
there,
I
would
suggest
going
along
to
the
six
storage
meeting
on
first
day.
That's
a
good
place
to
go
and
get
involved.
H
F
H
A
All
right
so
at
cube
con,
we
had
a
bunch
of
amazing
community
members
actually
join
us
for
a
community
lunch
at
cube
con,
which
was
super
super
fun.
So
we
had
had
a
bunch
of
people
join
us
for
on
to
tables.
Here
you
can
see
we
had.
We
had
a
nice
little
bolero
flag
as
well.
That
I
believe
Noland
now
has
in
his
office.
A
B
Can
jump
in
for
our
table?
We
talked
for
a
while,
with
with
Carl
from
cruise
automation
and
with
Dylan
from
Red
Hat
about
kind
of
the
multi-tenancy
issue,
or
feature
that
some
folks
are
interested.
So
the
idea
is
that
you
know,
rather
than
having
velaro
backups
and
restores
essentially
be
scoped
to
like
a
cluster
admin
role,
kind
of
enabling
namespace
owners
to
be
able
to
backup
and
restore
their
own
names
faces.
B
There
may
be
some
other
levels
of
granularity
for
the
permissions
era,
but
namespaces,
certainly
the
first
one
so
Carl
from
Cruise
is
definitely
interested
in
this.
For,
for
his
team
and
Dylan
talked
a
little
bit
about
it
and
interest
that
they
might
have
in
it
as
well.
It's
something
we've
had
kind
of
in
our
backlog
for
a
long
time
and
something
we'd
like
to
do.
B
But
you
know
there
are
definitely
some
technical
challenges
that
are
not
I
would
say
not
dissimilar
from
the
home
problem
and
and
kind
of
the
issues
around
tiller
running
as
a
cluster
admin
inside
the
cluster.
So
we're
hoping
to
continue
discussions
there
and
do
some
work
on
trying
to
come
up
with
a
technical
design
that
that
might
enable
that
kind
of
multi-tenancy
solution.
So
I'm
looking
forward
to
continuing
to
discuss
that
and
work
on
that
in
the
new
year
with
community
members.
D
So,
if
you're
trying
to
back
up
through
a
managed
service
such
as
eks
or
aks
or
gke,
you
don't
get
access
to
a
TD.
You
just
can't
use
that
tool
and
also,
if
you're,
if
you
do
have
access
to
that,
it's
a
big
hammer.
So
you
don't
you
don't
get
any
granularity
whatsoever
with
restoring
or
backing
up
you
just
get
the
whole
STD.
D
D
That
was
kind
of
another
theme
that
I
heard
was
Valero
came
from
a
cloud
centric
world,
where
a
snapshot
is
something
you
take
and
then
Amazon
and
Google
and
Microsoft
all
like
replicated
to
their
object.
Storage
for
you
and
people
working
in
a
more
traditional
world.
A
snapshot
is
like
a
point
in
time,
pointer
to
use
a
box,
and
it's
a
temporary
thing
that
you
leave
around
to
do.
D
B
Yeah
and
I'll
just
jump
in
there
and
I
know
definitely
yesterday
at
the
see
gaps
meeting
where
they
were
where
Dave
was
talking
about
the
data
protection
working
group.
That
was
an
issue
that
he
and
I
guess
he
primarily
brought
up
in
terms
of
kind
of
the
limitations
of
the
existing
CSI
snapshot
API
and
that
there
is
no
facility
for
doing
that
kind
of
data
move
so
I
think
that's!
You
know,
that's
an
example
of
an
issue
that
they
want
to
discuss
and
try
and
tackle
in
that
working
group
and
I.
B
D
Kubernetes
itself
is
just
facing
this
more
and
more
now
that
it's
it's
kind
of
grown
it's
facing
now.
How
do
we
move
it
into
this
space
and
so
we're
starting
to
tackle
these
kinds
of
issues
and
translate
it
back
from
this
cloud
native,
bringing
this
cloud
native
approach
to
kind
of
more
traditional
data,
centers.
E
We
have
a
similar
situation
where,
though
we
run
in
cloud
providers,
opening
ears
takes
a
point
in
time
snapshots
and
then
for
the
restore,
unlike
cloud
providers,
we
we
don't
have
those
volumes
on
the
other
side,
so
we
hoped
in
whether
whether
a
plug-in
to
transfer
the
data
after
taking
the
point-in-time
snapshot
to
ns3
something
similar
to
ristic
plugin
that
you
already
have
and
then
on
the
restore
side,
we
reconstruct
that
Von
Doom
from
the
backups
that
we
have
taken
yeah,
it
would
be
yeah.
I
was
not
aware
of
the
data
protection
working
group.
E
D
A
Awesome
I
had
one
really
interesting
discussion
during
these
meetings,
as
well
with
a
gentleman
from
Library
of
Congress.
They
are
looking
to
move
into
running
kubernetes.
They
are
there
still
and
they
still
have
a
few
development
clusters
and
so
on,
but
they
one
of
the
things
that
they
said
was
that
they're
actually
using
Valero
as
one
of
the
to
showcase
that
kubernetes
can
be
safely
deployed
and
backed
up
and
recovered
and
restored,
because
that
is
of
course
very
very
important
for
Library
of
Congress
to
be
able
to
restore
data
in
case
something
happens
so
yeah.
A
A
A
E
So
I'll
just
present
one
use
case
and
just
want
to
hear
what
people
think
here
or,
like
you
know,
has
it
been
already
taught
before
right
so
with
the
Valero
right
now
we
have
this
capability
of
taking
it
back
up
at
a
namespace
level
by
or
or
by
applying
the
labels
and
that
kind
of
stuff
right.
So
one
challenge
that
we
have
is
typically
an
application
actually
comprises
of
more
than
one
state
full
workload.
E
Now,
if,
for
example,
it
could
be
like
a
elasticsearch
with
some
of
the
database
that
has
like
some
information
that
ties
into
the
information
that's
provided
in
the
elasticsearch
right
now.
So
when
we
do
a
que
backup,
we
tend
to
snapshot
one
volume
at
a
time.
So
if
kind
of
becomes
inconsistent,
if
we
want
to
restore
it
on
the
other
side,
so
is
there
a
way
for
us
to
define
at
a
higher
level
like
it
on?
This
is
an
application
snapshot.
A
B
Yeah
I'm
I'm
not
positive
at
the
top
of
my
head,
if
they're
dealing
with
a
sort
of
consistency
groups
or
if
it's
more
of
an
inconsistent
back
up
at
the
namespace
level,
but
this
I
mean
the
idea
of
consistency.
Groups
for
volume
snapshots
at
least
is
definitely
something
that's
been
brought
up
before
within
the
Valero
community.
I
know
port
works.
If
you
look
at
the
issue
that
I
linked
there
was
interested
in
this
and
I
believe.
D
D
We
looked
I
think
this
was
beginning
of
2018
and
we
looked
and
rook
the
rook
team
was
saying:
hey
staff
offers
this,
but
like
we
couldn't
find
that
it
was
an
actual
implemented
thing.
It
was
like
more
like
a
road
map
feature
so
I,
don't
know
that
that's
actually
manifested
I
also
linked.
This
is
something
we've
talked
about
in
the
past
that
Google
has
proposed.
D
D
Definition
for
backing
up
an
entire
application,
so
this
goes
beyond
just
a
storage,
but
like
this
would
say,
this
is
everything
in
the
application,
so
it
would
I
think
if
I'm
remembering
the
details,
it
would
say
this
is
the
storage.
These
are
the
manifests.
This
is
how
it
would
get
redeployed
all
of
that.
D
It
I,
don't
know
if
it's
directly
related
but
I'll
have
to
go,
find
the
cap
there's
a
kept
for
applications
like
elastic
search
and
Cassandra,
and
things
like
that
that
someone
at
VMware
was
proposing
to
six
storage
that
was
kind
of
it
kind
of
pokes
holes
in
the
kubernetes
storage,
abstractions,
but
I,
don't
know
if
that's
I,
don't
know
how
relevant
that
is
to
this.
It
was
more
about
managing
storage,
topology.
E
I
think
that
the
one
day
dimension
about
the
Google
and
Luke
sounds
interesting,
so
also
like
any
issues
that
are
there
and
have
a
little
community
itself,
probably
reach
out
to
one
of
you
to
get
some
pointers
on
it.
To
take
that
forward.
I
I
see
that
the
notes
are
already
being
updated
with
those
issues.
So
that's
great.
A
E
All
right,
so
the
elastic
thing
was
so
they.
This
is
an
application
that
came
up
with
their
own
backup
operator
to
help
with
migrating
the
data
or
like
even
maintaining
a
disaster
copy
can
velaro
make
use
of
those
operators
to
our
caches.
Just
like
the
way
we
have
today,
plug-ins
for
storage
providers.
Can
we
have
something
at
an
application
level
that
says
ok,
you
know
this
is
a
application
that
you
need
to
backup.
Don't
worry
about
the
previa
cups,
but
you
actually
make
use
of
some
other
commands.
B
Yeah
it's
a
great
example:
I'm,
not
intimately
familiar
with
the
details
of
kind
of
how
the
elastic
operator
does
backups
but
I'd.
You
know
I'd
definitely
be
interested
to
kind
of
look
into
either
the
hooks
or
or
the
plug-in
functionality
that
bolero
offers
and
see
if
there
is
a
way
to
to
kind
of
fit
it
into
that
framework.
I
mean.
If
you
you
know.
If
you
were
to
write
a
custom,
backup
item
action,
you
could
essentially
have
it.
B
You
know,
execute
arbitrary
logic,
and
so
it
could
make
potentially
make
calls
out
to
the
operator
signal
the
operator
to
trigger
a
backup.
But
my
guess
is
that
you
know,
even
if
you
could
make
it
work
right
now.
There
probably
are
some
rough
edges
around
it,
and
so
it
might
be
an
area
where
we
would
want
to
make
some
enhancements
but
definitely
have
to
explore
some
more
and
kind
of
understand
it.
A
little
bit
better.
E
E
B
Yeah
so
I
think
the
I
mean
the
approach
you
talked
about
where
you
have
a
mini
o
in
the
target
cluster
that
you're
essentially
sending
backups
directly
to
I.
Don't
see
any
reason
why
that
wouldn't
work
I
mean
it's.
You
know
as
long
as
as
long
as
the
source
cluster
can
access
that
cluster,
so
there's
some
kind
of
ingress
or
load
balancer.
B
That
should
work.
You
know,
certainly
always
you
always
need
to
think
about.
If
you
have
your
backup
stored
in
the
cluster
that
you're
backing
up,
then
you
know
it
may
not
be
the
most
resilient
approach,
but
I
think
for
a
migration
solution
that
that
makes
sense,
and
we
you
know
we
have
we've
tried
or
or
talked
to
users
who
have
done
sort
of
similar
things
to
that.
B
Got
it
yeah
I,
don't
I,
don't
know
that
we
have
any
issues
specifically
around
this
in
the
backlog.
You
know
we
certainly
have
some
more
general
issues
around
kind
of
sharing
backups
between
clusters,
but
I
think
yeah.
This
would
this
would
definitely
be
an
interesting
area
to
explore.
Some
more
can.
D
I
ask
for
a
clarification
there
when,
when
you're
asking
for
this
direct
point
point,
are
you
asking
for
doing
doing
the
point
point
without
say
doing
the
Mineo
in
the
target
cluster
so
you're
asking
can
I
can
I
send
everything
directly
over?
Hopefully,
okay
just
want
to
make
sure
I
clarified,
because
I
think,
as
as
Steve
said,
I
think
using
Mineo
is
totally
possible.
It
would
be
a
matter
of
documentation
to
say,
hey,
here's
how
you
might
set
that
up,
but
in
terms
of
straight
across
that
would
be
a
that
would
be
more
engineering
work.
E
I
think
yeah
the
solution
where
the
Mineo
is
pretty
straightforward.
Just
thinking
of
cases
where,
like
maybe
did
maybe
some
constraints
on
the
storage
on
the
destination
side,
for
just
maintaining
extra
backups,
and
all
that.
So
this
is
mostly
like
an
use
case
for
yeah.
You
have
like
multiple
clusters.
You
just
want
to
move
an
application
around.
E
D
E
E
Yeah
and
I
I
heard
the
new
stock
video
I
think
that
also
the
car
Lisa
brought
up
similar
use
case
where
kubernetes
upgrades
could
be
like
a
good
potential
candidate
to
use
this
as
well.
So
you
don't
want
to
touch
the
existing
cluster,
but
you
bring
up
a
new
kubernetes
cluster
with
a
newer
version
and
start
migrating
one
application
at
a
time
to
the
other
side,
I.
B
I
believe
I
can't
remember
the
details
I'm
a
little
bit
fuzzy
there,
but
I
believe
that
there
was
a
maybe
a
kept
that
was
proposed
upstream
around
application
migration.
I,
don't
recall
if
it
was.
If
it
dealt
with
cross
cluster
migrations,
it
may
have
been
kind
of
namespace
to
namespace
migrations,
I'll
dig
around
and
see
if
I
can
find
a
link
to
it
or
I.
Don't
know
if
you've
seen
that
before
hi.
B
B
B
Yeah
so
yeah
anyway,
we're
basically
starting
to
work
on
planning
what
we
want
to
do
for
1.3,
and
you
know,
we
definitely
know
that
CSI
snapshot.
Integration
is
going
to
be
a
headline
feature
for
1.3,
given
that
the
upstream
api's
are
moving
to
beta
with
the
the
next
release
of
kubernetes,
but
we
we
also
want
to
fit
some
additional
items
in
there.
So
we're
we're
continuing
to
work
on
doing
the
planning
in
the
scoping
year
over
the
next
couple
of
weeks.
B
I
think
you
know,
given
that
we
have
essentially
three
work
weeks
before
everyone
sort
of
takes
off
for
end
of
your
holidays.
We'll
spend
a
decent
part
of
that
time,
just
working
on
planning
and
and
starting
to
work
on
the
items
we
can,
but
so
beyond.
Csi
snapshot,
integration
I
put
a
link
here
to
our
Zen
hub
board,
which
is
kind
of
how
we
go
about
doing
planning
and
I've.
There
are
some
items
that
have
been
moved
into
the
release
backlog
pipeline
there.
B
So
if
you're
interested
to
see
kind
of
what
we're
currently
thinking
as
far
as
1.3
scope,
you
can
go
in
there
and
look
at
the
release,
backlog
and
there.
These
items
are
all
tagged
with
the
1.3
milestone.
The
top
few
that
are
in
here
now
are
all
related,
CSI
snapshot
integration,
and
then
we
have
a
handful
of
other
bugs
and
small
enhancements
that
we'd
like
to
tackle.
B
D
So
it
might
like
be
a
weird,
no
op
thing,
so
I
think
the
the
big
question
I
have
right
now
is:
will
the
CSI
Valero
plugin
move
from
a
the
item,
actions
to
the
snapshot
or
interface,
or
will
it
stay
on
the
item
actions?
That's
the
kind
of
thing
I
want
to
decide
before
moving
into
beta
and
also
updating
for
the
beta
CRTs.
D
So
that's
kind
of
where
my
thinking
is
right
now,
I
think
I
want
to
review
your
prototype
of
the
volumes
snapshot
or
interface
a
little
bit
more
closely
and
then
kind
of
go
from
there.
I
know.
You've
also
have
a
prototype
on
the
rustic
volume
snapshot
or
plugin
from
a
from
a
maintenance
standpoint
for
us,
it
would
be
great
to
have
everything
go
in
one
code
path,
but
I'm,
not
100%
sure.
If
that's
the
answer
right
now,
so
that's
kind
of
where
my
my
thought
process
is
right.
Now,
yeah.
B
That
makes
sense,
I
mean
you
know
my.
My
overarching
goal
is
to
try
and
have
similar
or
identical
user
experiences
and
also
code
paths
for
all
of
these
different
ways
of
backing
up
following
volumes
right,
so
we
started
with
the
volume
snapshot.
Our
plugins
rustic
operates
in
a
in
a
very
different
way
than
those
do
and
has
a
different
user
experience
and
different
code
paths
and
with
the
introduction
of
CSI
integration,
you
know
that's
potentially
yet
another
user
experience
and
code
path
and
so
I.
You
know
I'm
I'm.
B
My
goal
is
to
just
try
and
kind
of
keep
those
things
as
simple
as
they
can
be,
both
for
the
user
and
for
developers
who
are
maintaining
the
code
and
so
exploring
whether
CSI
could
fit
into
a
volume.
Stop
snapshot
or
was
you
know,
was
just
me
trying
to
wrap
my
head
around
whether
that's
even
a
possible
approach
to
keep
things
simpler,
yeah,
given
the
the
issues
that
I
saw
around.
How
restores
would
work
I'm
honestly,
not
sure
that
that
makes
sense
as
an
approach
and
I
think
you
know
your
point.
B
Around
kind
of
the
longevity
of
each
of
these
things
is
a
good
one.
You
know
I
think
I,
think
we're
kind
of
saying
and
the
community
is
saying
CSI
is
the
is
the
strategic
direction
that
kubernetes
storage
is
going,
and
so
over
the
long
run
you
know
once
once
the
your
providers
move
to
CSI
and
that's
the
standard
way
for
getting
storage
access
for
for
any
of
your
providers.
There's
not
much
of
a
reason
for
Valero
to
keep
around
its
own,
its
own
plugins,
at
least
for
block
storage.
B
B
So
we
don't
end
up
with
with
just
kind
of
three
alternate
new
X's
and
and
code
paths.
For
each
of
these
things,
yeah
the
you
know
we
can
talk
about
the
rest
ik
bit
separately,
I
mean
a
lot
of
the
same
issues
apply,
although
you
know
the
path
that
we
want
to
take
with.
That
may
be
different,
but
yeah
anyway.
Those
are
those
are
kind
of
my
thoughts
on
the
topic.
Yeah.
D
D
That's
out
there
right
now,
wouldn't
recommend
for
production,
but
I
want
to
start
moving
to
the
point
where
we're
really
we're
not
just
saying
oh
hey:
can
this
work
I
want
to
say
like?
Let's
start
thinking
about
like
UX
and
like
my
my
goal,
I
think
I've
expressed
this
most
of
the
people
on
the
team,
but
I
would
like
to
see
the
CSI
plugins
get
included
in
velaro
by
default.
D
D
Don't
know
when
that
will
happen,
not
saying
it
will,
but
that's
where
I
would
like
to
see
it
go
eventually,
but
that's
future
for
the
immediate
time
being.
I
think
we
need
to
explore,
make
sure
we
integrate
correctly
and
also
make
sure
CSI
is
where
we
need
it
to
be.
So,
like,
like
I,
said,
I'm
gonna
help
work
on
getting
the
tests
ready.
It
sounds
like
their
test
Suites
currently
broken
because
of
their
code
split
and
also
they're,
getting
ready
to
release
the
images
for
the
new
controllers.
D
And
and
also
I
just
want
to
make
sure,
since
we
talked
about
like
the
valar,
the
Valero
vol
snapchatter
plugins,
like
eventually
going
away
like
I,
don't
I
want
to
make
sure
I
soothe
any
fears
like
we're,
not
gonna
turn
those
off
tomorrow.
Right,
it's
this
is
not
to
say
hey.
This
is
going
away
in
the
next
six
months
or
whatever
it's
long
term
like
we
want
to.
We
want
to
be
good
citizens
in
the
community
and
not
be
making
our
own
weird
thing
for
storage.
D
You
know,
storage,
vendors,
they're
gonna,
be
writing
CSI
plugins
and
we
don't
want
them
to
be
writing
a
CSI
plug-in
with
volumes
snapshotting
and
the
valerio
plugin.
That's
doing
volumes
snapshotting
for
too
long
right,
like
we're,
not
gonna,
be
great
community
partners.
If
that's
the
case,
you're
gonna
be
cursing
our
name.
If
that's,
if
that
happens
for
too
long,
so
that's
kind
of
the
logic,
but
we
also
get
in
today
in
2019
in
early
2020.
That's
not
we're
not
gonna,
be
in
a
place
where
that's
really
realistic.
I.
E
D
Right
now
it
does
doesn't
have
the
separate
backup
one.
The
volume
snapshot.
Api
is:
there's,
there's
a
few
different
components
to
the
CSI
spec.
There's
like
a
G
RPC
API
that
all
the
the
drivers
and
everything
will
talk
to
you
and
then
there's
the
kubernetes
api
that
Valero
will
use.
It
doesn't
talk
the
GRP
C
API,
the
GRP
C
API
as
I
understand
it.
That's
all
pretty
much
GA
for
the
snapshotting,
the
backup
for
like
data
copying,
and
things
like
that
is
a
road
map
item.
D
E
D
Yeah
this
again,
this
was
a
so
odd.
None
and
I
actually
talked
about
how
the
current
CSI
implementation
works.
That
sorry,
not
the
current,
the
previous,
the
the
alpha
one,
although
the
beta
ones
very
similar
how
the
kubernetes
CSI
implementation
works
right
now,
and
it
was
very
much
designed
by
cloud
providers
and
internally
on
VMware,
we
had
this
design
discussion
to
the
CSI.
D
Spec
is
going
to
need
to
grow,
to
have
that
data
movement
API
and
the
GRP
C
API
and
that's
gonna
require
more
on-prem
providers
to
come
to
the
six
storage
meetings
and
say
hey:
these
are
our
use
cases.
Can
we
agree
on
on
like
this
common
denominator,
API
to
move
data
and
again
those
are
every
other
Thursday
at
12
p.m.
Eastern,
9
a.m.
Pacific
just
so
next,
two
days.
B
D
B
Just
I'll
call
out
the
new
data
protection
working
group,
that's
in
the
process
of
being
formed
like
that's.
This
is
a
perfect
example
of
you
know
what
they're
trying
to
go
after
so
I
would
definitely
encourage.
You
know
anyone
who
has
a
vested
interest
in
seeing
this
to
to
get
involved
with
that,
and
they
will
be
they'll,
be
presenting
the
idea
for
that
working
group
that
sig
storage
on
Thursday
this
week.
D
One
more
quick
plug
I
added
I'll
put
it
in
the
chat
and
put
it
in
the
notes.
There's
a
github
topic
for
Valero
plugins.
If
you
are
making
a
Valero
plug-in-
and
you
add
this
topic
to
your
github
repo,
then
it
will
show
up
in
that
getup
search
automatically
get
up.
Oh
go
index
it
for
you,
and
that
is
also
now
linked
in
our
plugin
Docs.
So
just
want
to
call
that
out.
If
anybody
is
catching,
this
recording
and
I
forgot
to
put
it
in
my
status,
update.
A
Okay,
so
thank
you,
everyone
for
for
joining
the
call
today
lots
of
really
good
conversations.
Thank
you
all
for
all
the
feedback
and
yeah
interesting
discussions
here
today.
This
recording
will
be
up
on
YouTube
most
likely
later
today
or
early
tomorrow
morning,
and
with
that
hope
to
see
you
all
next
week
and
have
a
wonderful
week,
everyone
bye-bye
bro.