►
From YouTube: Velero Community Meeting/Open Discussion - April 7, 2020
Description
April 7, 2020
Status Updates
[nrb]
On community support this week
Reviews
Pushed up an example of readiness probes for Velero at https://github.com/vmware-tanzu/velero/pull/2382
Talked to Steve and Ashish about revisiting the Velero CSI plugin data flow/data model
Continued testing with CSI integration at https://github.com/vmware-tanzu/velero/pull/2323. Will likely require some changes given we’re revisiting our data model
Seeing some more CRD issues crop up, though not all around versioning - https://github.com/vmware-tanzu/velero/issues/2370 & https://github.com/vmware-tanzu/velero/issues/2383
Looks like some schema restriction errors that need to be investigated further
Worked with Rafael and Carlisia on a presentation for his work to the data protection working group
[steve]
released v1.3.2
discussion/reviews on CSI integration
install/config UX design review
investigating ways to report backup progress (https://github.com/vmware-tanzu/velero/issues/20)
PSA: looks like we’re hitting https://github.com/kubernetes/kubernetes/issues/86149 - manifests as a ~15-second delay at the start of every backup
[carlisia]
Addressed code reviews on CLI PR - close to wrapping that and starting to implement
Fixed the yaml/kustomize examples
[ashish]
Have a few PRs in the pipeline for CSI. Close to opening this up for wider review in the community. This is coming soon.
Survey that we announced last week is now live and available here. Call to community action!
Discussion Topics
[carlisia] DEMO of deploying Velero using Kustomize - Files: https://github.com/carlisia/velero/tree/c-cli-design/design/CLI/PoC
[brito-rafa]
Handling multiple API Group versions during restore (related to PR #2373)
Default (today)
Restore Source Cluster Preferred Version
Target-Version Mode (proposed - feature flag?)
Triggered by detection the target cluster runs a different K8s version from the source backup cluster (see PR #2346 collecting cluster k8s version during backup). Restore objects based on target cluster apiGroup preferred version.
Next step: Please look at the PR, and then Rafael will put together a design proposal
Contributor Shoutouts
@rochfordk - docs to clarify backup TTL (https://github.com/vmware-tanzu/velero/pull/2399)
@mansam - add --cacert flag to CLI commands (https://github.com/vmware-tanzu/velero/pull/2364)
@mynktl - add selected-node remapping plugin (https://github.com/vmware-tanzu/velero/pull/2377)
B
B
This
is
kind
of
a
spike
just
to
try
things
and
it
may
fit
some
things
for
I
know.
Anthony
Bette
from
Dell
had
mentioned
some
some
problems
that
this
may
take
care
of.
There's
still
some
fix
up.
I
would
have
to
do
for
this,
but
if
anybody
is
curious
about
this,
go
ahead
and
try
this
I
know
I
think
Dylan's
on
this
call.
B
This
may
be
of
interest
to
him,
but
this
would
also
make
it
possible
to
start
Valero
without
having
a
backup
storage
location
defined,
but
you
would
then
have
kubernetes
checking
this
live
in
his
endpoint
to
indicate
whether
or
not
the
pod
should
be
restarted.
So,
if
you're
curious
about
that
check
it
out,
it's
not.
B
It's
not
super
high
priority
at
the
moment,
but
something
has
to
be
together
in
an
evening
last
week,
going
back
to
the
other.
Things
have
worked
on
Steve
and
Ashish
and
I
were
talking
about
some
of
the
data
modeling
and
flow
through
the
CSI
plugins
last
week,
and
she
should
has
some
PRS
that
I
think
the
PRS
came
out
of
that
meeting.
I
mostly
merged
we've
got
a
lot
about
those
I'm
continuing
to
test
the
CSI
integration
with
core
there's,
probably
going
to
be
some
changes.
B
B
B
There
we
go
so
perversion
yeah,
so
this
error.
This
is
something
to
do
with
Sierra
DS
in
thinking
one
it
complains
about.
If
you
have
all.
If
you
have
multiple
versions
defined
in
your
schema
for
your
Sierra
D
it
once
and
they
all
have
the
same
values
it
wants
you
to
define
that
at
a
top
level
instead
of
setting
that
per
perversion.
B
So
that's
something
we
may
have
to
revisit
I've
seen
it
I
just
wanted
to
acknowledge
it,
but
I
have
not
taken
a
closer
look
at
it.
Yet
so
I
just
wanted
to.
Let
people
know
we're
aware
of
it,
and
then
the
next
thing
was
had
some
check-ins
with
Raphael
and
carlesha
on
some
of
his
work,
which
I
think
he's
going
to
discuss
later
on
some
some
things
he's
going
to
present
to
the
data
protection
working
group
on
how
he's
expanding
on
how
we
handle
versioning
on
various
API
groups
and
I.
A
A
B
Week,
I
know
that
this
it's
a
Rafael's
issue.
Work
is
not
merged,
yet
his
work
is
about
grabbing
all
versions
of
any
data
type,
whereas
this
is
particularly
to
see
our
DS
and
I
think
this
is
a
result
of
doing
some
conversion
from
V
1,
beta,
2,
V,
1
of
the
custom
resource
definition
itself
and
I.
Think
the
one
is
more
strict
about
what
it
wants
from
what
I've
read
in
these
issues.
B
D
Everyone
past
week,
so
we
released
1.3
dot
2
on
Friday.
It's
a
minor
release,
probably
not
an
essential
upgrade
for
most
folks.
It
enabled
plugin
off
there's
two
stored
data
within
the
the
main
velaro
bucket
and
fix
the
small
client-side
bug
in
the
Valero
rustic
repo
get
command
that
essentially
just
threw
a
panic
when
a
rustic
repository
had
just
been
created,
but
not
yet
initialized
so
for
most
folks,
who
are
already
out
in
1.3.1,
probably
not
necessary
to
upgrade
but
never
hurts
beyond
that.
D
A
bunch
of
discussion
and
reviews
on
the
CSI
snapshot,
work
and
and
some
config
UX
design
work,
so
I
won't
go
into
those
anymore.
I
spent
most
of
yesterday
looking
into
possible
ways
to
report
backup
progress.
So
this
is
something
we've
we've
kind
of
had
in
the
backlog
for
forever.
Maybe
our
oldest
issue,
that's
still
open,
but
we've
talked
about.
D
You
know
having
some
way
to
count
the
total
of
number
of
resources
that
were
going
to
backup
and
then
provide
updates
as
we're
going
so
yesterday,
I
was
just
looking
into
possible
ways
that
we
could
change
our
code
to
enable
this,
and
what
I
was
mainly
looking
at
yesterday
was
sort
of
separating
the
backup
process
into
two
parts.
So
you
know,
if
you
think
about
how
our
backup
code
works
it
essentially.
First
it
has
to
hit
the
kubernetes
api
to
get
the
yeah
mall
or
the
json
for
a
given
resource.
D
So
I
came
up
with
a
very
basic
prototype
of
modifying
the
code
to
work
that
way.
That
seems
to
work.
Okay,
so
I'll
probably
provide
some
more
updates
on
this
in
the
coming
week
or
two
I
didn't
really
want
to
do
it
demo
today,
because
there's
not
much
to
see
other
than
some
log
statements
indicating
that
resources
are
being
counted,
but
hopefully
we'll
come
up
with
an
approach
that
makes
sense
here
and
try
and
something
implemented.
E
D
D
And
then
yeah,
just
one
nice
thing
that
it
because
I
was
looking
at
it
this
morning,
I
noticed
that
backups
had
random
kind
of
15-second
pause
at
the
beginning
of
them.
This
might
be
occurring
with
restores
too,
but
I
know
to
sit
with
backups,
and
this
is
something
that
we've
had
issues
with
before,
but
I
discovered
that
there's
a
bug
in
the
client
go
version
that
we're
using
around
the
discovery,
client
where
all
the
requests
are
getting
throttled.
So
this
is
happening
to
us.
D
F
F
Guess
I
will
put
together
unless
Steve
or
anybody
has
any
other
idea.
I
stopped
putting
together
a
list
of
issues
out
of
that
design
and
also
sorting
by
what
to
implement
first
and
I'm
thinking.
What
I
would
like
to
implement
first
I
think
is
the
renaming
from
from
provider
to
location,
plugin
and
now
I'm
also
wondering
if
we
should
have
a
separate
branch
or
if
we
could
do
that
against
the
master
branch,
because
the
install
is
definitely
gonna
break.
F
So
that's
one
thing.
The
other
thing
oh
I,
also
excuse
me:
I
worked
on
fixing
the
yellow
files.
There
were
in
the
example
because
they
were
working
and
then
they
weren't,
but
now
they're
working
again
because
I
fixed
it
was
just
the
way.
The
scheme
validation
was
laid
out
with
the
version
and
I
added
to
the
discussion
topics
I
a
demo
which
is
very
uneventful
which,
but
I
walked
through
the
files.
I'll
show
how
everything
comes
together.
F
F
G
Yeah,
so
I
don't
have
much
of
an
update.
I
have
a
few
PRS
in
the
pipeline
to
kind
of
getting
the
CSI
plugin
towards
like
becoming
more
open
or
open
for
wider
review
in
the
community.
So
this
is
coming
soon.
The
other
update
I
had
was
around.
Last
week
we
meant
mentioned
a
survey
that
we
wanted
to
run
in
the
community
to
catalog
like
blueprints
or
workflows
that
folks
may
be
using
in
the
community.
G
A
Awesome,
thank
you.
Sheesh
yeah,
the
the
survey
is
super
super
short
and
if
you
want
to
help
out
to
build
out
these
workflows,
please
make
sure
that
you
mention
who
you
are
either
on
slacker
and
github.
So
we
can
reach
out
to
you
or
tag
you
in
the
issues
this
shouldn't
take
longer
than
about
two
minutes
or
so
of
your
time.
So
if
you
have
time
and
if
you
are
interested
in
helping
us
understand
all
the
workflows
that
are
interesting
to
you,
please
yeah,
please,
please
take
the
survey.
F
F
F
A
F
So
the
on
the
branch
that
I
have
out
for
that
PR
there
so
that
we
have
the
design
folder
so
I
have
this
CLI,
EPOC,
folder
and
the
base
directory
has
what
would
be
the
things
that
would
be
installed
with
the
justifier
of
server
config
command.
That
will
just
kick
up.
Kick
out,
kick
up
the
the
server
initialization
and
so
the
customize
file
is
this
and
it
has
listed
the
initial
deployments,
which
is
not
a
complete
deployment,
because
it
doesn't
have
things
like
the
any
plug-in
and
the
CR
de
specs
and.
F
F
Command
so
potentially,
if
we
were
to
have
customized
examples,
this
would
be
in
a
different
folder.
I,
don't
know
yet,
but
at
any
rate,
I
also
included
the
mini
llamo,
so
I
could
have
access
to
a
storage
location,
a
storage
bucket,
and
so
this
so
the
out
of
all
these
files
are
here
under
this
base
directory
and
the
deployments.
F
As
it
see
is
very
short,
it's
missing,
like
I,
said
plugins
and
some
other
things.
The
overlay
is
why
you
put
on
top
of
that
deployment,
and
that
would
be
specific
to
why
you
are
the
what
you're
burning
again.
So
if
you
have,
if
you
want
to,
they
don't
know
as
plug-in.
That's
what's
going
to
go
on
your
customized
file
and
that
would
be
patched
in
so
that
is
a
deployment
resource
that
will
be
patched
into
the
Valera
resource.
So
if
you
look
here,
it
looks
very
much
like
the
Valera
deployment.
F
I
mean
we
do
need
to
have
all
of
these
initial
things
like
the
Dame.
So
it
recognizes
what
deployment
it's
going
to
patch
and
then
it
has
thing
is
very
specific
for
the
plug-in
and
potentially
you
can
have
once
we
have,
that
design
are
implemented
in
plus
the
secrets,
as
also
which
is
got
moved
to
a
different
PR.
F
Maybe
that's
going
to
be
the
right
way
to
do
this,
not
sure
yet,
but
so
these
are.
This
is
how
you
set
up
customized.
Basically,
you
put
yellow
file
somewhere
and
you
tell
customize
where
they
are,
and
then
you
run
a
customized
commands.
So
here,
for
example,
the
way
to
do
this
is
to
go
straight
to
the
overlays
that
you
want
to
deploy,
because,
as
you
see
here,
the
the
overlay
stuff
that
you're
working
with
calls
into
the
base.
F
F
F
F
If
we
got
what's
missing
like
the
plug-in
and
start
the
backup,
storage
rotations,
and
once
it
has
everything
that
that
it
needs
to
create
backups,
then
it's
going
to
mark
itself
as
ready
for
creating
backups,
but
it
will
be
ready,
as
opposed
to
like
we
see
you
know
already,
because
missing
things,
and
so
just
this
is
yeah.
This
is
never
going
to
be
ready
now,
if
I
do
que,
apply
and
go
to
the
plugins
I'm,
it's
going
to
try
to
deploy
the
things
from
the
base,
but
it's
just
going
to
ignore.
F
F
F
F
I'm
Chucky
now
yeah
it
is
here
it
is
on
the
other
part.
Well,
I
felt
I
wanted
to
deploy
the
bait
the
stuff
with
the
bass
and
show
that
he
wasn't
here
and
then
deployed
the
overlays
and
so
that
he
was
here,
but
just
trust
me
so
this
got
patched
in,
but
there
is
I
haven't
I
didn't
have
time.
I
wanted
to
show
this
today
and
I
didn't
have
time
to
do
to
the
gear,
dig
in
and
see
why.
F
F
Is
is
an
open
source
project?
You
got
cut
off
a
little,
so
if
I'm
not
answering
completely
what
you're
asking
it's
a
open
source
project
sponsored
by
VMware
as
well,
and
it
just
it
gives
you
this
visual
dashboard,
so
you
select
whatever
namespace
you
want
to
work
with
I,
just
just
alliterative,
a
little
deployments
of
others,
not
here
anymore.
Oh
you.
G
F
It
lets
you
browse
through
things.
You
can
see
your
customer
sources,
you
can
see
like
if
you
have
a
deployment
running,
you
click
on
there
and
it
will.
Let
you
see
the
log
for
that,
for
the
pods
I
mean
we'll
take
it
to
the
respective
pods
and
then
you
can
see
the
log.
For
that
part,
that's
pretty
cool.
D
At
least
yeah
I
have
a
quick
question
on
the
ya
know:
files
so
I
know
for
the
POC
you
have.
Those
files
are
static,
files
that
are
checked
into
your
branch.
It
so
is
the
idea
that
these
will
be
dynamically
generated
by
the
CLI
commands,
or
are
we
going
to
be
maintaining
these
statically
in
the
repos?
Yes,.
F
Great
question,
so
actually
the
the
way
I
put
this
together
was
from
with
our
style
command.
We
have
the
dry
run,
so
I
got
the
mo
output
and
I
just
broke
it
into
these
separate
files.
So
the
idea
is
very
much
that
each
commands
will
have
llamo
outputs
that
you
then
can
dump
the
entire
llamo
file
into
whatever
directory
makes
sense.
D
F
F
D
B
The
thing
I
liked
with
this
customized
approach
is
that
if,
if
you're,
applying
a
basically
a
patch
to
the
deployment,
we
don't
necessarily
have
to
you
specify
a
whole
lot
of
stuff,
we
don't
have
to
specify
custom
API
or
anything.
You
say
well,
your
plug-in
carries
some
sort
of
customized
overlay
it
patches
to
deployment
with
whatever
you
need
environment
variables,
while
it
mounts
whatever
and
your
your
plug-in
knows
how
to
get
to
those.
The
only
thing
we
expect
isn't
a
knit
container
that
puts
it
in
the
right
place.
B
G
B
B
E
B
H
A
A
I
Yeah,
this
is
a
continued
one,
with
conversation
about
supporting
multiple
API
group
versions
and
if
you
see
on
the
screen,
that's
the
PR
in
the
pipeline
for
the
change
that
now
backing
up
all
versions
of
the
API
group.
You
see
on
the
top
a
screenshot
of
the
standard,
bolero
backup.
Now
you
have
the
directory
name
spaces
under
each
resource,
the
PR
in
route.
I
Basically,
it
takes
the
same
backup
for
backward
compatibility
and
then
now
we
have,
for
each
supported
version
of
the
same
resource,
have
a
subdirectory
for
the
version,
with
a
string
preferred
version
on
telling,
which
is
the
preferred
version
of
the
backup,
and
this
directory
here
matches
the
content
over
here.
Does
it
make
sense
now.
I
Looking
forward
how
we're
gonna
take
advantage
of
those
multiple
versions,
doing
the
restore
and
that's
I
wanna
discuss
and
bring
to
attention
is
I
have
a
second
slide
here
today,
the
default.
We
just
restore
the
preferred
version
of
the
source
cluster
right,
because
we
only
have
one
version.
Now
we
going
forward
we're
gonna
have
three
versions
in
that
example.
Here:
okay,.
I
One
of
the
ideas
is
whatever
name:
we're
gonna
call
target
version
mode
is
now.
What
is
the
preferred
version
of
the
target
cluster?
We
are
restoring
because
not
necessarily
gonna
match
the
source
cluster,
so
the
first
idea
is
like
trigger
the
detection.
If
the
target
server
is
running
a
different
version
of
kubernetes,
then
the
backup
was
taken
off
in
how
to
do
this.
There
is
another
PR
which
put
a
label
on
the
backup
CRT.
What
the
version
of
kubernetes
can
do
a
quick
check.
I
If
it's
different,
then
we
can
look
at
the
backup
and
pick
the
version
on
the
backup
that
matches
the
preferred
version
of
the
cluster.
That's
that's
got
someone
on
bring
to
discussion
here.
If
doesn't
make
sense,
if
does
can
we?
You
know
it's
gonna,
be
a
feature
flag
and
so
on
so
I
can
start
wrapping
and
putting
some
code
around
the
logic
for
restore.
These
are
the
mode
I
rather
shell
for
discussion,
which
is
a
little
bit
more
aggressive
like
when,
doesn't
anything
match.
I
Then
we
can
like
try
to
do
some
conversion
on
the
fly.
I
would
say
this
is
coming
after
this,
because
on
the
target
version
mode,
we
are
covering
situations
where
eventually,
multiple
escaping
kubernetes
releases,
we're
gonna
have
different
preferred
versions,
and
then
we
want
to
address
that.
It's
a
little
it's
much
safer
is
a
kind
of
no-brainer
here,
then
you
know
whatever
magic.
We're
gonna
do
down
the
road.
To
do.
Conversion
so
here
doesn't
include
any
conversion
at
all
of
the
objects
is
just
matching
the
versions
that
we
have
on
the
back
up.
I
G
G
D
Just
gonna
say
you
know
before
you
jump
into
coding,
it
all
might
make
sense
to
just
circulate
a
you
know
can
just
be
on
an
on
the
issue,
just
what
the
kind
of
what
the
UX,
what
your
proposed
UX
is.
Gonna.
Look
like
for
this.
As
far
as
you
know,
is
there
a
feature
flag,
dot
button
than
those
motors
in
a
new
CLI
flag
on
the
restore
command,
along
with
the
new
API
field,
on
the
restore
object.
G
Yeah,
so
what
I'm
saying
is,
maybe
folks
may
not
want
multiple
versions
of
api's
on
every
backup
that
they
take,
maybe
honor
self
when
they
want
to
upgrade
or
when
they
want
to
move
to
another
cluster
or
to
a
different
version.
It
is
only
then
they
may
want
to
take
backups
with
multiple
versions
of
their
resources.
G
F
D
J
We're
gonna
add
the
fly
I'm
assuming
the
default
would
be
just
restore
everything,
and
this
would
be
like
a
you
know.
Legacy
motor
single
a-sittin
of
the
flight
would
be
kind
of
single
version
mode
where
you
would
eliminate
this
extra
stuff,
but
with
no
flag
the
default
would
presumably
be
the
new
version
with
all
this.
J
I,
just
if
we
were
talking
about
adding
a
flag,
so
you
could
do
a
single
version,
backup
versus
the
multiple
version
backups
and
guessing
you
would
want
to
default.
If
you
had
it's
best,
buy
anything
to
be
the
full
future
format,
and
this
would
be
an
option,
would
not
be
on
by
default,
which
would
allow
a
user
to
only
backup
the
single
old
format.
One
version.
J
J
I
F
I
B
The
person
who
would
care
is
the
person
like
the
application
developer,
who
made
a
CRD
and
they
were
like
I,
know
specific
they're
they're,
making
like
a
backup
blueprint,
and
they
know
they
want
to
do
some
sort
of
conversion.
That's
the
type
of
person
who
would
want
to
do
that
in
specific
I.
Think
for
this
PR
we
don't
want
to
get
that
that
specific
I
think
this
is
a
foundation
and
we
can
introduce
if
people
want
that,
we
can
address
it.
B
I
Yeah,
oh
I,
agree
me.
This
specific
version
could
be
a
restore
action
right,
not
our
backup,
because
if
you
cherry
pick
on
bring
the
backup,
then
you
the
codes
gonna
be
super
messy.
The
API
they're
worth
change
was
to
get
all
the
versions
was
basically
one
method
start
up
the
preferred
group
version.
We
got
the
whole
universe,
so
code-wise
was
not
complicated
right.
I
I
B
D
G
I
think
the
only
people
who
would
probably
care
people
who
have
their
own
object
storage
and
they
pay
for
like
the
fights
that
that
they
have
to
store.
So
if
they're
running,
like
scheduled,
backups
like
once
a
backup
once
a
day
and
the
backups
stick
around
for
30
days,
so
they
have
like
30
days
worth
of
backup
sitting
in
their
storage.
G
On
the
other
hand,
they
may
say
I'm
going
to
run
scheduled
backups
without
this
flag,
but
I'm
going
to
run
using
this
flag
only,
but
I
want
to
boast
for
my
great.
So
that's
that's
where
I
was
coming
from,
but
I
don't
think
most.
Anybody
is
like
paying
for
pipes
on
their
storage,
so
I
think
it's
fine.
I
I
J
G
B
Because
I
think
I
had
suggested
on
on
write
that
in
the
name
in
that
namespaces
directory
and
that
you
drop
a
file
that
was
named
like
the
like,
we
have
the
v1
preferred
version.
I
was
trying
to
suggest
dropping
a
file
named
the
version
or
preferred,
or
something
like
that,
yeah
that
it
just
wouldn't
be
referenced
in
the
normal.
In
the
current
paths,
it
would
just
be
ignored,
but
any
ones
could
use
it,
but
yeah
it
restore,
might
be
a
little
bit.
Messier
yeah.
I
Actually,
like
the
the
fact
that
we
have
you
know,
it's
subdirectories
is
kind
of
like
cleaner,
at
least
for
you
know,
I
I
think
would
be
very
polluted.
We
have
like
how
old
of
all
the
objects
on
the
same
subdirectory
point
versions.
I,
don't
know
visually,
it
would
be
complicated.
I
think
the
brain
works
better
in
subdirectories.
I
B
I
B
Yeah
so
like
this,
this
format
is
going.
It's
not
necessarily
useful
if
you're
doing,
if
you're,
if
you're
staying
up
to
date
with
your
kubernetes
and
you're,
always
doing
like
point
release
upgrades
you're,
probably
fine,
but
if
you're
doing
a
big
jump
from
110
111
to
116
you're,
certainly
gonna
notice,
hey
my
deployments
aren't
supported
anymore,
so
this
is
in
order
to
get
ahead
of.
That
problem
is
the
whole
reason
we're
doing
we're
we're.
Looking
at
this
work
to
make
ups,
you
make
your
backups
more
more
long-lived
and
more
useful
for
doing
bigger
jumps.
I
J
I
Yeah,
that's
the
whole
point
of
it
right.
I
gotta
tell
the
end.
User
does
not
know
top
of
their
head.
The
API
groups
preferred
version
the
support
version.
Speaking
out
of
my
experience,
you
usually
only
one
particular
API
groups,
but
by
the
end
of
the
day,
you
really
memorize
the
versions
that
you
run
in
terms
of
kubernetes
version.
I
I
really
want
to
see
someone
take
a
backup
on
111
and
restore
and
125
and
transparently
happens
right
and
that's
the
whole
point
of
this
exercise.
I
A
D
To
be
able
to
use
a
custom,
see
a
bundle
on
the
client,
so
I
think
that
essentially
wraps
up
the
that
work
around
custom,
CA,
bundles
I
know
we
have
one
one
issue
open
for
for
adding
some
Docs,
but
but
the
codes
all
in.
So
thanks
a
lot
Sam
for
all
of
those
PRS
and
then
my
UNK
I
got
the
PRN
that
we
talked
about
over
the
last
couple
of
weeks
around
adding
a
plugin
to
remap
the
selected
node
for
local
PDS.
D
A
Awesome
yeah.
Thank
you
all
contributors
for
making
bolero
even
better.
Thank
you
Steve
for
writing
these
down
and
making
sure
that
they
are
always
communicated
here
so
thankful
for
that
and
with
that
we're
ending
the
April
7
episode
of
the
Balearic
community
meeting.
We
hope
you
all
have
a
fantastic
rest
of
your
week
and
yeah
check
with
you
on
slack
and
see
you
all
on
the
same
channel
here
next
week
have
a
good.