►
From YouTube: Velero Community Meeting - Nov 16, 2021
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
Hello,
everyone
and
welcome
to
the
valerio
community
meeting,
slash
open
discussion
today
is
november
16th
2021
and
we
have
a
limited
team
here
with
us
today.
So
let's,
let's
get
started
first
up,
we
got
some
status
updates,
dave.
B
Yeah
so
still
trying
to
push
through
the
item,
snapshot
or
pr,
I
think
we've
got
all
the
comments
are
addressed.
Daniel
signed
off
and
bridgette
will
probably
sign
off
today.
I
don't
don't
think,
there's
anything
open
there.
We
still
have
room
for
making
changes
so
we'll
see
how
the
rest
of
that
stuff
merges
together.
B
I've
been
rolling
some
stuff
back
in
the
astrolabe
code
base
off
of
my
branch
and
looking
at
putting
that
behind
the
item,
snapshot
or
as
a
poc
we're
doing
some
work
with
the
postgres
snapshot
are
just
pretty
basic
and
then
looking
at
rolling
in
pvcs
and
stuff
like
that
and
I'll
answer
for
bridget,
because
she's
off
working
on
on
that
kind
of
stuff
today,
and
so
she
was
on
sport
last
week,
I
think
was
on
sport
this
week.
I
can't
remember
now,
I
think,
daniel's
on
support.
B
B
Yeah
so
scott,
I
just
wanted
to
see
if
you
were
hooked
up
with
the
oad
oadp
stuff.
So
you
know
we
have
this
proposal
for
making
a
dp
api
that
we
could
call
from
our
tens
of
mission
control,
stuff
and
tanzu
and
we're
thinking.
This
should
be
something
that
we
have
available
everywhere
and
there's
oadp
from
red
hat
right
now.
But
that
looks
to
be
pretty
much
valero.
C
Yeah
and
I
think
dylan's
on
the
call
he
might
be
better
to
answer
this
in
terms
of
the
future,
but
but
at
the
moment
basically
oedp
is
a
red
hat,
supported
bolero
install
we're
basically
using
the
same
fork
of
valera
that
we're
using
for
our
migration
tool,
but
you
know
I
think,
going
forward.
We
may
be
doing
more
with
that,
so
I
don't
know
if
dylan
has
any
thoughts
here
or
not.
D
I
can
add
to
that,
so
we're
actually
like
totally
reworking
the
api
for
odp
right
now
to
segment
it
slightly
from
valero
to
support
the
future
use
cases
that
we
envision
with
it,
and
we
I
mean
funny
enough.
We
actually
renamed
it
to
data
protection
application,
so
we're
probably
thinking
similarly
here,
one
of
like
the
big
value
adds
that
long
term
review
for
valero
is
something
that
you
find
with
a
lot
of
other
backup
applications,
and
I
think
astrolabe's
got
a
similar
concept
to
this,
which
is
people
call
it
different
things.
D
Ability
to
take
snapshots
and
either
convert
them
to
different
storage
types
or
providers,
or
also
just
simply
pick
them
up
and
move
them
elsewhere.
So
that's.
That's
really
like
the
next
big
feature,
development
effort
on
odp,
but
as
scott
mentioned
yeah
right
now,
it's
it's
a
way
to
install
valero
with
our
plugin
to
support
openshift
use
cases.
There's
also
a
couple
extra
components
to
support.
Openshift's
got
internal
images
for
an
internal
registry
where
we
host
container
images
so
backing
up
and
restoring.
That
is
like
a
whole.
D
It's
a
part
of
a
valero
plug-in.
It
also
uses
a
separate
registry
dependency.
So
for
the
most
part
it's
a
glorified
valero
with
openshift
internals
that
are
needed
to
support
backup,
restore
there
and
yeah.
It
is
funny
you
mentioned
that,
because
we
did
we
used
to
call
it
horribly
named.
The
crd
was
just
the
kind
valero
to
imply
that
you're
installing
valero-
and
we
said
this
is
super
confusing,
so
we're
going
to
rename
it
and
we
decided-
let's
just
call
it-
data
protection,
application
and
valero
is
a
data
protection
application.
D
So
that's
thought
there,
but
then
I
guess
one
last
thing
to
add
to
it
is:
you
know
we're
heavily
focusing
on
csi
snapshots,
but
the
fact
is
like
we
have
a
lot
of
customers
that
you
know
they're
still
using
nfs
they're
still
using
traditional
file
store
stuff
and
they
need
something.
That's
a
little
bit
more
reliable
than
rustic
rustic,
especially
with
nfs
and
you've
got
root.
Squash
set
like
there's,
there's
problems
there
so
for
our
migration
application.
D
You
know
we
had
to
kind
of
start
from
the
ground
up
to
say
how
we're
going
to
handle
these
volumes
in
a
different
manner
and
we
wrote
well
again.
This
is
kind
of
getting
too
deep
in
the
wheat
for
the
migration
use
case,
but
a
glorified
rsync
tool
that
takes
pvs
rsync
to
data
out
into
a
new
pv.
So
essentially
like
a
persistent
volume
mover.
We're
trying
to
think
about
for
oedp
long
term,
like
you
know,
is
rustic
the
right
tool
to
use
for
backing
up
those
types
of
volumes.
D
Is
there
something
else
that
we
can
explore?
Maybe
some
of
the
tools
we
built
for
the
from
from
the
conveyor
community
to
do
this
too.
I
don't
know,
there's
a
lot.
We
don't
really
know,
but
all
the
work
that's
going
to
odp
right
now
is
we're
trying
to
up.
Until
this
point
it's
been
a
community
product
project
where
we
basically
took
all
of
our
valero
installation
logic
from
our
migration
tool
and
just
segmented
out
the
valero
installation
bits
that
people
could
get
valero
and
openshift
easily.
D
Now
it's
kind
of
turning
into
more
of
like
a
generic.
Yes,
you've
got
valero
and
that
does
backup
restore
volumes
and
does
backup
restore
of
kubernetes
objects
and
what,
if
you
want
to
take
snapshots
and
do
them
elsewhere?
What
if
you
want
to
integrate
with
different
backup
vendors?
So
maybe
they
can
add
a
plug
into
odp
to
to
leverage
some
of
the
use
cases
of
valero
too,
so
that
was
pretty
long-winded,
but
hopefully
that
answered
some
of
your
questions
whatever
you
need
to
know
happy
to
help
answer.
B
Well,
I
think
we
should
team
up
on
some
of
these
things,
so
our
plan
right
now
is
to
so
we
we
have
as
tanzu
this
goal
of
having
multiple
data
protection
providers
so
that
we
can
have
valero
as
the
basic
one.
But
then
we
want
to
be
able
to
swap
in
a
veritas
power
protect
cast
and
casting.
Is
you
know
off
on
their
own
path
of
how
they
do
things
and
so
be
able
to
swap
those
in
but
give
a
consistent
experience
to
the
kubernetes
admin?
So
they
get
like
a
limited.
B
You
know
basic
backup
experience,
which
is
pretty
much
going
to
be
very
similar
to
the
valero
experience,
backup
cluster
schedule,
backup
restore
cluster
those
kind
of
things,
and
then
we
can
build
things
like
tmc.
On
top
of
that
tensor
command
line
and
customer
automation
and
we'll
have
like
a
reference
implementation
on
top
of
valero.
B
But
then
the
the
goal
is
to
get
the
other
dp
vendors
to
buy
into
this
as
well,
and
I
think
if
we
can
show
them
that
this
api
gets
them
not
only
tanzu
but
also
openshift,
where
the
openshift
customers
can
take
advantage
of
the
same
api
and
get
this
consistent
experience.
D
That's
the
direction
we're
going
in
yeah,
that's
yeah!
I
would
love
to
set
up
a
separate
call
to
discuss
this
because
it's
it's
similar
online
to
what
we
were
thinking.
I'll,
be
honest,
a
lot
of
like
the
current
work
for
oedp
is
like
productization
stuff,
so
we
haven't
really
put
a
ton
of
feature
development
into
it
all
right
now,
so
these
are
conversations
that
we
definitely
want
to
have
we
just
it's
been
kind
of
on
the
backlog
to
start
shifting
all
the
engineering
effort
towards
you
know
some
more
future
use
cases.
B
Yeah-
and
this
is
a
good
time
to
have
the
discussion
and
we're
going
to
do
a
lot
of
the
work
for
valero,
so
you
know
if
you
guys,
are
giving
input
and
you're
like
hey
this
works
for
us
too,
then
we
can.
You
know
it's
a
stronger
case
to
like
the
veritas,
and
you
know,
castings
of
the
world
to
say:
hey,
look,
you
got
these
two
already
and
then
we
can
take
that
back
to
the
kubernetes
community
as
a
whole
and
say:
look
you
know
we're
doing
this.
B
D
Absolutely
yeah
that
sounds
great
okay,
while
we're
talking
about
this
too,
I
want
to
mention
shoe
bomb
is
on
the
call.
I
don't
know
if
you
guys
have
talked
to
him
too
much
he's
also
on
the
odp
project.
D
E
E
He
wrote
in
the
chat,
hello,
everyone
so
dave.
I
had
a
question
for
you.
Sorry.
If
this
question
sounds
like
I,
I
am
new
to
storage,
so
I'm
not
really
familiar
with
all
the
words
you
said
like
in
and
stuff,
but
are
you
thinking
that
astrolabe
will
eventually
replace
rustic
in
valero,
so
instead
of
using
rustic,
you
know
with
a
damon
set
installing
all
the
pots.
Are
you
thinking
astrolabel
to
the
data
moving
eventually.
B
Yes,
so
a
couple
of
different
things,
so
one
is
astrolabe,
is
a
set
of
apis
and
we
have
things
that
back
those
apis
and
ashley
basically
gives
you
things
like
snapshot
plus.
B
Let
me
access
the
data
in
volumes
and
snapshots
and
it
gives
out
a
variety
of
transports
you
can
extract
via
so
that's
the
astrolabe
api
in
general
and
then
right
now
we
are.
We
have
like
a
really
stupid,
s3
repo
that
just
puts
the
whole
object
into
s3
objects,
we'd
like
to
have
something:
that's
smarter
and
we're
actually
looking
at
using
copia
behind
the
astrolabe
apis
as
the
destination
for
things
and
copia
gets
us
things
like
d-do
versioning,
and
it's
a
it.
B
So
that's
something
we're
evaluating
right
now.
We've
got
a
poc
working
with
this
that
we
can
demo
when
we
do
our
astrolabe
demo
in
a
couple
weeks.
So
the
idea
is
that
it's
broken
into
multiple
things,
so
the
current
rustic
pods
and
data
mover
is
something
that's
on
the
chart
for
valero.
So
this
is
currently
something
that's
inside
the
vsphere
plug-in
where
we
move
between
vsphere
snapshots
and
s3,
but
we'd
like
to
include
other
things
right
now.
It
uses
it.
B
B
So
the
existing
rustic
daemon
set
will
probably
augment
that
with
a
generic
data
movement
capability
because
we
can
use
the
same
run
context
out
there
and
make
that
into
our
scale
out
data
mover
and
that
will
handle
both
block
volumes
using
the
afterlab
apis
and
also
file
systems
via
the
astrolabe
apis
and
we'll
be
able
to
hook
blocked
volumes
and
file
systems
and
other
things
into
copia
as
our
back
end.
But
the
repo
is
is
swappable
so
that
somebody
like
a
power
protect,
can
say
yeah.
B
I'm
going
to
use
this
astrolabe
front
end
to
snapshot
things
and
extract
data
from
the
different
things
that
I've
snapshotted,
but
their
back
end
becomes
their
own
proprietary
like
data
domain,
and
they
can,
you
know,
hook
the
two
together.
So
that's
that's
kind
of
the
goal
there
does
that
does
that
help.
E
B
Yeah
and
so
and
then
casting
is
a
another
data
protection
vendor.
They
were
fairly
early
on
in
the
kubernetes
space
and
they're.
Now
part
of
the
theme
which.
A
E
B
You
know
it
it
depends
on
which
direction
you
look
at
from
the
valero
point
of
view,
their
competitor,
but
they're,
not
open
source.
From
the
vmware
point
of
view.
The
tanzania
point
of
view
they're
another
data
protection
vendor
that
we
really
want
to
be
able
to
support.
So
there's
an
important
part
of
both
not
just
competition.
B
There's
code
sharing
we're
actually
doing
some
work
with
them
to
integrate
their
canister
technology,
which
is
a
blueprint
for
snapshotting
and
backing
up
applications
or
currently
have
a
poc
underway
to
integrate
that
with
astrolabe,
so
that
we
can
trigger
canister
blueprints
from
astrolabe,
so
that
you
could
define
an
application
as
a
canister
blueprint,
make
that
a
naturally
protected
entity
and
then
go
ahead
and
trigger
that
and
that's
something
we
announced
at
kubecon,
and
you
may
recall
that
casting
was
one
of
the
diamond
sponsors
at
cubecon,
so
decided
to
throw
some
money
at
the
kubernetes
space.
E
A
Cool
awesome,
any
any
other
discussion
topics
here
or
comments
on
the
topic
at
hand.
A
So
you're
going
to
set
up
a
separate
meeting
for
more
discussion
around
this
or.
B
Yeah,
I
think
we'll
we'll
we'll
get
it
well
eleanor
and
I
will
catch
up
with
dylan
and
we'll
see
how
we
bring
this
out
right
now
we're
kind
of
doing
requirements
internally.
I
want
to
get
outside
requirements,
and
you
know
getting
a
strawman
api
together
and
we'll
start
tossing
it
around
and
then
we'll
bring
that
back
into
the
community
meeting
when
we
have
a
little
something
to
discuss.
D
That
sounds
great,
and
I've
got
a
couple
folks.
On
my
end,
that
would
probably
be
interesting
that
call
as
well
so
yeah,
if
you
just
reach
out
I'd,
be
happy
to
chat
on
this
further.
A
B
A
From
bruce,
who
is
part
of
the
the
new
valero
team
here
over
in
china,.
B
B
Looks
like
something
in
the
group
version.
So
yes,
so
it
gives
us
a
little
more
flexibility
on
group
versions.
It
looks
like
so
that's
good.
So
thank
you.
Bruce.
B
A
B
B
Api
pagination
code,
so
if
we
had
a
large
number
of
objects
at
some
point
it
fails
we
had.
Somebody
was
up
to
like
four
gigabytes
of
stuff
that
came
back
in
one
big
chunk
and
it
blows
out
of
grpc.
So
pagination
has
been
an
important
feature.
The
first
round
of
that
looks
like
it
had
a
bug,
so
this
is
a
fix
for
that.
So
thank
you,
alay
and
then
what's
this
last
one
here
from.
A
B
So
this
is
just
a
a
note
that
when
we
use
the
s3
path
style,
so
that's
so
typically
that's
something
that
we
use
with
like
either
an
on-prem
object
store
like
a
mineo
or
sometimes
if
people
have
like
proxied
s3,
they
might
want
to
go
through
the
url
style,
and
when
you
do
that,
it's
actually
required
that
you
set
the
region,
even
though
it
seems
like
it
should
automatically
figure
out
the
region,
because,
but
you
don't
need
a
region,
because
there
is
no
region,
but
you
actually
have
to
set
that
because
of
the
sdk.
A
Awesome
and
that's
it
for
today.
Thank
you.
Everyone
for
joining
next
week
is
thanksgiving
week
here
in
the
u.s,
so
many
of
the
u.s
based
employees
here
at
vmware
will
be
offline,
but
yeah.
Well,
we'll
still
have
the
community
meeting
next
week,
which
is
in
the
beijing
time
zone.
So
seven
a
no,
not
seven
a.m,
7
p.m!
Eastern,
where
I
am
next
week,
if
you're,
if
you
want
to
join
we'll,
still
have
that
and
have
the
conversation
continuing
so
with
that.
Thank
you.
A
Everyone
have
a
fantastic
rest
of
the
week
and
see
you
all
soon.