►
From YouTube: Velero Community Meeting - Dec 15, 2020
Description
Dec 15, 2020
The Velero maintainer team at VMware will be on holiday between December 19, 2020 to January 4, 2021.
Status Updates
nrb
On community support this week
EOY tasks, planning for next year
Code reviews
Maintainers, please give the following PRs a look:
DownloadRequest controller to kubebuilder
Set custom certificates on a BSL
carlisia
PRs + PR reviews
wip: Adding reference to a credential to the BSL
Secrets stuff
wip: review Frankie’s PR (Restore API group version by priority): https://github.com/vmware-tanzu/velero/pull/3133
bridget
Community support first half of this week
Starting vacation Wed/Thurs
Continuing to work on multiple credentials proposal
PR reviews
About to submit a PR for delve debugging in Tilt
dsmithuchida
Submitted talk proposal for Kubecon EU on Backup/restore of operator driven apps
Started on design for upload progress and data movement tracking
Discussion Topics
carlisia
JenTing Hsiao as core maintainer [APPROVED!]: https://github.com/vmware-tanzu/velero/pull/3188
If you’d like to join us, take a look at our Governance document
What if we optimize the behavior of the sync controller in regards to instantiating a backup store: https://github.com/vmware-tanzu/velero/pull/3187
Combine the behavior of sync & BSL controller
Could we have this done in a design doc?
Reduces the amount of requests to object storage
Reduces instantiations of object storage plugins
Reconciling the server default BSL with the user configured default BSL: https://github.com/vmware-tanzu/velero/issues/3173
nrb
New GitHub Processes
Using Ginkgo
Frankie:
Uses BDD descrptions, which isn’t necessarily used
Doesn’t feel “Go-native”
Carlisia
controller-runtime uses Ginkgo’s BeforeTest entrypoints to do setup.
Dave
Used e2e tests cause we were using it elsewhere
Thought there were benefits to it that haven’t manifested
Contributor Shoutouts
@matheusjuvelino: issue: add flag to the schedule cmd to configure the useOwnerReferencesInBackup option #3176
@matheusjuvelino: Owner reference in backup when created from schedule
@imjokey: Tencent S3 Compatible Support Docs
@nemeth: doc: Use inline markdown links in tables
@yusufgungor: Preserve nodePort support with --preserve-nodeports flag
@jenting: feat: support configure BSL CR to indicate which one is the default
A
B
Hey
I
loans
community
support
this
week
to
close
out
the
week.
I
put
this
above
status
updates,
but
I
wanted
to
let
folks
know
the
team
here
at
vmware
will
be
out
on
holiday
between
december
19th.
That's
this
saturday
and
january
4th.
The
majority
of
us
will
be
back,
so
we
won't
be
having
a
community
meeting
between
those
two
weeks
for
those
of
you
who
might
be
watching
on
youtube
and
we
probably
will
not
be
answering
on
slack
or
github.
B
The
big
ticket
item
there
from
our
team
will
be
on
designed
for
multiple
secrets
per
provider,
and
I
know
scott
sego
is
looking
at
getting
the
ability
for
restore
item
actions
to
wait
on
their
additional
items.
We
may
get
that
in
1.6,
depending
on
how
things
work
out
other
than
that
been
trying
to
get
code
reviews
moved
along.
In
that
spirit,
I
want
to
highlight
a
couple
that
have
been
kind
of
old
for
the
maintainers.
B
The
download
request.
Controller
moved
to
cube
builder
has
been
open
for
a
long
time,
so
I
think
I've
got
that
approved
now
other
maintainers.
If
you
please
take
a
look
and
the
next
one
is
actually
fairly
small.
Setting
custom
certificates
on
a
backup,
storage
location
is
mostly
a
quality
of
life,
kind
of
adding
it
to
the
command
line.
Interface.
B
B
D
D
So
that
can
be
a
point
that
can
point
to
the
toy
secret
associated
with
that
location.
And
then
I
started
looking
into
that,
and
there
is
something.
Apparently
there
is
a
deprecation.
They
don't
want
us
to
use
the
kubernetes
object,
reference
type,
they
say,
create
a
different
type,
but
I
I
have
a
bunch
of
tabs
open
to
look
into
that.
C
D
C
C
D
C
D
B
C
Yeah,
I
I'll
send
you
a
link
afterwards
carlesia,
but
I
also
came
across
a
secret
key
selector,
which
is
like
a
combination
of
like
a
local
reference
for
the
secret
name
and
then
also
like
a
thing
to
select
which
key
and
the
secret
you
want
to
use.
So
that
might
be
yeah.
Definitely
that's
it.
I
haven't
tried
it
out.
I've
like
I,
don't
know
how
it
works,
but
I
was
looking
at
a
few
different
fields
or
a
few
different
things
in
the
api,
docs,
so
cool.
You
said
that
your
way.
D
D
C
Hi
everyone
so
nolan,
and
I
are
kind
of
splitting
the
community
support
this
week
because
it's
my
first
week
doing
it
officially
so
as
of
new
year.
I
think
that'll
be
like
fully
on
the
on
the
schedule,
so
I'm
doing
it
for
the
first
half
of
this
week
until
I
start
my
vacation,
which
will
be
at
some
point
tomorrow
or
from
thursday,
so
I
still
haven't
quite
figured
that
out
yet
which
days
I'll
be
working
as
crazy
mentioned.
C
C
Apart
from
that,
I
also
been
doing
lots
of
pr
views
and
but
haven't
gotten
to
the
ones
which
are
most
important,
so
I
will
focus
on
that
this
afternoon
and
then
I
also
owe
a
pr
for
the
delve
debugging
support,
which
I
mentioned
last
week.
I
have
it
all
in
place.
I
just
need
I'm
finishing
off
the
documentation
for
it,
so
I'll
get
that
up.
This
afternoon,
as
well.
E
So
we've
had
some
experience
with
like
postgres
operator
what
istio
cappy
so
we'll
try
and
pull
that
together
get
some
best
practices
and
also
look
at
some
future
roadmaps
for
how
to
fix
this
problem
permanently.
And
if
anybody
out
there
has
had
their
own
experience
with
that.
You
know,
and
you
have
something
to
share
that.
You
know,
send
it
in.
F
E
You
know
we'll
put
it
into
the
talk
of
you,
know,
use
cases
and
stuff.
So
there's
that
and
I
started
on
a
design
for
doing
upload,
progress
and
handling
managing
data
movement.
There
was
a
previous
one
that
was
there
that
got
abandoned,
I'm
taking
some
of
that
plus
adding
some
newer
stuff
in
that
we've
run
into
and
we'll
have
something
for
everybody
to
review
after
the
new
year.
So
that's
it
for
me.
A
B
G
No
problem
well,
carlesia
was
kind
enough
to
mention
my
pr
and
I
don't
know
do
you
think
this
is
an
appropriate
place
to
mention,
like
maybe
discuss
the
pros
and
cons
of
ginkgo.
You
know
I
I'm
probably
missing
a
lot
like.
I
don't
really
even
know
why
kubernetes
uses
it
and
I
don't-
and
actually
I
don't
know
if
this
is
the
right
form
to
mention
it.
If
that's
what
you
were
thinking,
nolan.
B
G
Oh
right,
yes,
the
I'm
struggling
to
run
the
end-to-end
test
on
my
machine
and
it
has
nothing
to
do
with
ginkgo.
It's
just
I'm
not
quite
sure
how
to
do
that.
So
I
posted
some.
I
can
post
the
error
messages,
I'm
getting
on
the
kubernetes
valero
slack
and
maybe
get
some
help
there,
but
it
the
code
part,
is
done.
All
the
unit
tests
are
done.
G
D
G
G
D
We
have
a
new
contributor
in
the
double
excited,
is
not
a
vmware,
you
know
contributor,
so
I
I
think
that's
good,
because
potentially
yeah
give
us
a
chance
to
get
different
perspectives,
so
his
name
is
jen
ting
and
it's
from
souza.
Is
that
how
you
pronounce
that
company's
name?
Okay,
so
he
works
with
souza
and
he's
been
contributing
pretty
frequently
to
the
valero
project.
He
is
already
a
maintainer
on
our
helm.
Chart
repo
and
he's
super
helpful,
even
on
slack,
so
I'm
really
excited
he's.
G
D
A
D
B
Awesome,
I
also
dropped
a
link
to
our
governance
document
for
any
anybody
else
interested
in
becoming
a
maintainer.
This
kind
of
explains
what
we're
looking
for
we're
essentially
looking
for
people
who
are
active
consistently
and
genting
certainly
has
been
he's,
been
been
active
in
implementing
features
and
responding
to
questions,
and
things
like
that.
So
we're
definitely
looking
for
more
external
maintainers.
So
if
you're
interested
take
a
look
at
this
also,
you
can
definitely
get
in
touch
with
us
and.
D
So
you
know
I
keep
running
the
logs
and
then,
if,
when
a
bsl
is
unavailable,
there
is
an
error
message,
but
then,
when
we,
when
the
the
the
sync
controller
runs
which
also
runs
continuously,
then
we
get
another
error
message.
It
wasn't
able
to
sync
because
you
know
couldn't
get
to
the
backup
storage
location.
So
I'm
thinking
I
was
thinking.
D
Why
are
we
throwing
this
error
twice?
We
should
only
do
it.
Do
it
once
like
pick
one
place,
do
it
once
and
be
done
with
so
then
I
was
thinking
this
is
a
small
change
but
could
have
a
big
impact.
So
it's
a
proposal
to
consider.
I
I
shift.
First
of
all,
I
shift
the
code
around
so
that
this
this
particular
change.
I
think
we
should
definitely
implementing
it,
wouldn't
have
an
impact
other
than
optimize.
D
If
there
are
no
items
to
sync,
then
don't
even
initiate
the
object
storage
because,
right
now
we
instantiate
it
first
and
then
we
get
to
the
point
where
oh
is
there
anything
to
sync?
No
continue,
so
we
always
instantiating
it
if
it
passes
the
some
some
prior
checks,
so
that
change
is
definitely
worth
putting
it
in.
I
think
the
other.
The
second
part
of
this
proposal
is
what,
if,
when
we,
when
it's
time
to
sync,
we
instead
of
trying
to
fetch
the
bsl
fetch
the
backups,
then
see
if
there
is
any
difference.
D
What
if
we
just
check,
is
this
backup
location
available?
And
if
it's
not
just
move
on
to
the
next
one,
either
throw
a
warning
or
an
error
but
then
move
on
to
the
next
one.
I
mean
it's
really
not
an
error,
because
the
bsl
controller
is
already
through
in
the
air,
so
just
warn
or
do
a
debug
message
move
on
to
the
next
one.
Now
for
this
to
to
work,
because
we
wouldn't
be
fetching
from
the
store
a
priority,
we
would
then
also
get
a
list
of
backups.
This
is
my
proposal.
I'm
not
you
know.
D
D
Work
when
the
bsl
checks
in
tries
to
validate
the
the
location,
the
bsl
controller,
since
they
already
got
the
the
store
object,
it
would
go
and
fetch
I
mean
we
only
need
the
names,
but
I
guess
we
have
to
fetch
all
the
backups
and
store
the
names
of
those
backups
in
the
list,
which
would
be
a
new
field
in
the
backup
storage
location
and
when
it's
time
to
sync
we
would
fac.
We
would
fetch
that
list
of
backup
names
from
the
bsl.
D
B
D
E
E
D
Yeah,
the
bsl
controller,
already
periodically
checks
the
back
the
the
the
location
for
validating.
So
at
that
time
it
would
fetch
the
list
of
backups
and
then
just
make
that
avail
when
the
sync
controller
is
doing
their
its
own
loop.
Instead
of
doing
the
same
thing,
we
just
you
know
check
based
on
the
so.
D
D
D
So
when
the
sync
controller
runs
the
then
it
doesn't
when
the
sync
controller
runs.
If
it
needs
to
do
a
sync,
then
it
has
to
fetch
the
backup,
storage
location,
the
back
the
backup
storage
again,
yes,
but
you
won't
have
to
do
that
unless
he
knew
he
had
to
do
a
sync
and
the
way
you
would
know
is
back
by
taking,
instead
of
fetching
the
list
of
backup
to
then
figure
out.
D
Bsl
already
has
a
periodic
sync:
no,
no,
not
a
sink.
I'm
sorry,
the
bsl
yeah
naming
is
important.
Yes,
exactly
you
wouldn't
do
the
same.
No,
the
bs,
just
to
be
precise,
the
bsl
won't
be
syncing.
Anything
you
just
be
fetching
the
names
of
the
list
of
backup
names.
That's
it.
The
syncing
is
still
happening
in
the
sync
controller.
D
D
D
E
B
B
E
So
it's
a
reasonable
question
there,
but
why
do
we
have
then
a
health
check
and
a
sink?
I
think
that's
the
basic
questions
right.
Why
do
we
have
a
health
check
and
a
sink?
Why
don't
we
just
do
them
in
one
together,
so
we
should
have
a
good
reason
for
why
we
split
out
the
backup,
storage
location
controller.
To
do
the
health
check
all
right.
D
I
mean,
logically,
having
all
the
backup
storage
location
operations
in
the
corresponding
controller
is
a
lot
easier
to
grasp
and
it's
it's
more,
it's
better
organized
instead
of
like
oh
now,
let's
make
the
sync
controller.
Do
all
of
these
other
things
as
well,
but
you
know
it
would
be
more
efficient
than
because
there's
also
a
rate
limit
issue
with
instantiating
a
object
store
right.
D
So
we
like
we
when
we
when
we
designed
the
when
we
implemented
the
bsl
controller,
we
were
you
know
giving
of
configuration
option
to
just
for
the
user
to
specify
how
the
frequency
of
that
validation,
because
we're
concerned
about
the
rate
limiting
limits
and
but
and
then
we
have
you
know,
then
we
have
the
sync
just
doing
that
all
the
time.
No
matter,
I
guess
there
is
a
sync
frequency
that
can
be
configured
too
so
sort
of
equivalence.
D
B
Would
like
to
see
a
design
dock
on
it
before
before
we
go
with
it.
I
think,
because
I
think
it
could
it
could
have
upsides
they're.
Just
not
super
clear
to
me
at
the
moment
like
if
it's
just
getting
rid
of
log
statements,
it
seems
like
an
awful
heavy
lift,
but
I
don't.
I
don't
think
it's
just
log
statements.
No.
D
E
E
E
E
B
B
B
Right
so
I
would
say,
like
I,
I
guess,
I'm
not
opposed
to
making
it
more
efficient
like
and
also
I
see,
jonathan
muting
time
boxing
the
discussion
yeah.
So
I
I
think
I
think
I
would
like
to
see
it
personally,
I'd
like
to
see
a
proposal
for
maybe
combining
all
of
this
into
something
a
little
bit
more
holistic.
F
F
F
F
D
A
D
And
not
that
sorry,
not
that
I
was
making
that
drinking
made
in
I
was
reviewing
and
nolan
raised
some
questions,
so
I
just
I
laid
over
everything.
I
don't
want
to
take
the
time,
I'm
not
definitely
going
to
read
the
whole
thing,
I'm
just
saying:
there's
a
decision
that
needs
to
be
made
and
the
decision
might
be
the
way
it
is
it's
handled
or
like
no
less
less
use,
genting's
suggestion
that
he
made
a
suggestion
for
addressing
this
in
a
different
way
or
something
else.
So
I
just
wanted
to
address
this.
D
D
I
mean
for
the
user
to
say
that
on
the
cli,
as
opposed
to
the
owners
on
the
server
side,
that
sort
of
applies
that
gets
configured
as
server
startup
and
it
rates
there's
some
considerations.
So
I
just
wanted
to
address
this
just
nailed
this
before
1.6.
So
I
just
I'm
just
asking
people
to
take
a
look
at
it,
see
if
there's
any
gacha,
you
can
see
any
war,
your
preferred
way
of
handling
it.
B
B
D
D
B
B
B
So
things
like
milestones
labels.
I
know
there's
ways
to
automate
that
I
think
terraform
can
do
that,
but
we
decided
to
pull
everything
together
into
one
spot,
additionally,
we're
using
github
projects
to
automate
things
and
there's
github
actions
that
support
github
projects,
but
they
don't
support
github
github
organization
project
boards
very
well.
So
that
is
the
motivation
for
moving
most
everything
into
the
valero
repo
for
issues.
B
B
Also,
we
made
a
support
board
at
that
link.
This
is
mostly
what
the
support
team
will
be.
Looking
at
brand
new
issues
will
get
added
to
this
board
and
we
have
a
grooming
meeting
every
week.
Take
a
look
at
them,
make
sure
to
get
answered,
and
then
we
double
check
to
see
if
people
have
gotten
back
to
us.
B
So
this
is
what
the
team
will
be
using
to
get
get
on
top
of
support
stuff
make
sure
things
are
answered
for
milestone.
Stuff
work
is
actually
scheduled,
we're
going
to
use
github
milestones,
and
I
made
a
pr
for
a
github
action
that
needs
to
be
fixed,
there's
a
typo
in
there,
but
the
goal
is
to
when
you
milestone
an
issue.
B
There
are
there's
an
currently
there's,
a
limitation
that
you
can't
add
it
when
you
just
make
the
issue
you
have
to
make
the
issue
then
add
it,
but
that's
something
that
can
be
fixed
later
and
there
was
one
more
thing:
oh
yeah
we're
not
tracking
prs
on
the
boards,
we're
just
going
to
track
the
issues,
because
we
found
tracking
prs
and
issues
made
a
ton
of
noise.
You
ended
up
moving
two
or
three
things
across
the
board,
which
just
wasn't
great.
B
G
G
Behavior
driver
development,
where
you
can
use
very
clear
to
easy
to
understand
to
non-developer
language
such
as
like,
given
gosh,
I'm
like,
I,
actually
I'm
not
very
good,
with
bdd
I'm
very
new
to
bdd,
but,
as
I
understand
it,
they
developed
it
so
that
maybe
the
scrum
leader
or
people
who
are
not
developers
can
understand
what
these
tests
do
and
they
they
write
these
tests
beforehand.
They
want
in
this
very
user-centric.
G
Other
benefits
of
ginkgo
are
these:
they
have
gold
mega
matcher,
so
you
don't
have
to
use
reflect
deep,
equal
kind
of
things
which,
by
the
way
we
have
we
use
currently
use,
testify
to
do
that.
Has
a
matcher
and
like
assertions
and
things
like
that.
Other
reasons
to
use
ginkgo-
or
it
has
you
know-
has
neat
little
keywords.
It
has
really
pretty
highlighting
and
colors
you
can
do
instead
of
using
regular
expressions
with
the
dash
run,
which
go
test
use
you
can
use
focus.
G
G
The
kubernetes
valero
slack
channel
I'll
bring
it
up,
and
I
welcome
more
feedback
from
the
community
is
what
what
would
be
the
benefit
for
the
valero
team
to
use
ginkgo
I
feel
like,
and
this
is
I'm
a
true
to
the
spirit
go
developer
and
that
I
really
hesitate
when
bringing
in
frameworks
to
do
something
that
requires
just
a
few
more
lines
of
code
with
the
standard
library
so
and
gingko
and
gomega
are
pretty
big
frameworks.
We
end
up
having
to
depend
on
them
having
to
read
the
docs
to
figure
out.
G
What's
going
on,
it
has,
I
think,
annoying
quirks
like,
for
example,
anytime.
You
use
it
before
each,
depending
on
the
scope,
you're
going
to
have
to
run
it
every
before.
Every
single
it
assertion
is:
that's
not
intuitive
nor
obvious.
You
have
to
read
the
docs
to
know
it
and
I
I
feel
like
if
it
was
written
just
with
standard
with
a
standard
go
library
or
even
a
simpler
framework.
G
Like
testify,
a
new
go
developer
or
someone
new
to
community
who
doesn't
know
bdd
would
be
able
to
come
and
read
it
easily
without
having
to
refer
to
the
docs.
So
these
are
the
points
I
bring
up,
maybe
before
we
go
a
little
bit
too
far
into
depending
on
ginkgo,
and
I-
and
I
noticed
this
because
I'm
starting
to
write
some
tests,
some
intent
tests
and
by
the
way,
we're
bypassing
most
of
ginkgo
right
now
as
it
is
so
it
doesn't,
it
doesn't
feel
like
we
really
need
it
so
I'll
bring
up.
Oh.
D
Yeah
sorry
go
ahead.
No!
I
no
thanks
for
bringing
this
up.
I
want
to
just
us
to
focus
the
conversation
a
bit
more,
so
ginkgo
is
now
used
everywhere
in
valero
ginkgo's
game
ginkgo
is
used
in
two
places.
Only
one
is
the
controllers
in
two.
The
second
place
is
the
e2e
tests,
because
I
was
the
one
who
introduced
ginkgo
in
the
controllers.
I
will
explain.
D
Why
is
because
so
the
controllers
are
moving
to
the
runtime
controller
on
time
framework
and
there
has
to
be
some
environment
configuration
that
happens
before
you
can
run
the
controllers
and
the
the
runtime
controller.
The
the
cube
builder
framework
uses
ginkgo
to
accomplish
that
that
in
I
you
know,
so
I
use
that
as
well
one
to
be
consistent
with
what
the
cube
builder
controllers,
how
they
are
implemented
and
tested,
but
two
also
because
that
problem
was
figured
out,
so
I
just
had
to
follow
the
instructions
so
for
the
e2e.
E
G
E
Don't
know
so
there
you
go
and
I
think
it's
a
japanese
plant
anyway,
the
yeah
we
wound
up
using
ginkgo
in
the
ede
test,
mainly
because
we
had
to
use
something,
and
I
I'd
heard
that
we
were
using
ginkgo
for
edd
testing
in
other
places.
So
I
thought
there
was
a
whole
bunch
of
benefits
to
edd
testing
that
would
come
out
of
this
and
there
aren't
so.
E
I
think
it's
going
to
depend
like
how
many
I
think,
the
the
value
that
I
see
it
is
if
you
have
a
very
large
test,
suite,
there's
some
ways
to
describe
and
manage
things,
and
I
don't
know
if
we'll
you
know
right
now,
our
cost
in
using
ginkgo
is
relatively
small
and
taking
it
out
is
not
a
big
deal.
Putting
it
in
is
not
a
big
deal,
so
you
know
if
you're
planning,
you
know,
10
15
tests
do
what
you
want.
D
D
Well,
the
the
e3
tests
are
different
kind
of
tasks,
so
you
might
have
different
kind
of
needs
from
unit
tests.
I
don't
know,
but
if
we
are
going
to
use
ginkgo,
there
is
a
way
to
do
table
tests
with
ginkgo.
Then
we
that
should
be
that's
what
should
be
used.
I'm
not
I'm
not
making
a
proposal
here
for
to
use
one
or
the
other.
I'm
just
saying
that
if
the,
if
that's
what's
decided,
then
that's
what
should
be
used
and
otherwise
that's
just
yank
it
out
and
not
use
it
at
all.
Yeah.
E
Yeah,
it's
become
one
of
these
things
where
so,
like
the
cappy
team
was
using
gitgo.
So
it's
like
okay.
Well,
let's
look
at
what
they
did
and
model
it
after
them,
but
then
the
cappy
team
shows
up
and
says:
well
we
don't
really
like
it
like
okay,
but
the
test
group
says
hey.
You
should
use
this
and,
oh
and
honestly
I
mean,
after
going
through
this
thing,
trying
to
wrap
my
head
around
it.
It's
like
it's
like
oakland,
there's!
No
there
there.
B
D
A
B
B
G
A
B
Yeah
so
the
first
two
were
actually
kind
of
related.
The
top
one
was
a
follow-up
from
the
second
one,
so
it's
from
they're,
both
from
I
I
apologize
matthus
juvelino.
This
is
for
introducing
owner
references
in
the
backup
object
which
he
needed
for
getting
proper
tracking
in
argo.
So
the
backups
were
basically
created
by
the
schedule
and
argo
needed
to
see
that
the
backups
were
owned
by
the
schedule.
B
So
this
introduces
an
option
to
use
valero
in
such
a
way
that
the
schedules
are
owned
by
I'm
sorry,
the
backups
are
owned
by
schedule.
This
is
not
the
default,
it
may
become
the
default,
but
no
promise
is
there
yet.
So.
Thank
you
very
much
methodist
for
both
of
those
the
next
one
is
from
I'm
joki.
It
is
documentation,
I'm
not
sure
what
they're
joking
I
I
think,
they're
based
in
china,
the
the
change
here
was
contributing
some
documentation
on
how
to
use
valero
with
the
tencent.
B
B
B
So
thank
you
for
your
efforts
in
coding
this
and
testing
this
yusef
and
then.
Finally,
this
is
one
from
genting
the
last
one
before
it
will
probably
get
him
added
to
the
maintainers
list.
Genting's
been
taking
on
a
lot
of
our
cli
enhancement,
work
that
we
had
scheduled
for
1.6,
and
this
one
is
a
big
change
to
add
a
field
to
bsls
or
backup,
storage
location,
custom
resources
to
configure
the
default.