►
From YouTube: Carvel Community Meeting - January 11, 2023
Description
Carvel Community Meeting - January 11, 2023
This week we had several discussion topics and issues raised by community members covering easing pain points with kapp-controller, making kapp friendlier towards custom resources with stable APIs from other projects, and integration of ytt as a flux controller. Plus, loads more!
Full notes + agenda here: https://hackmd.io/G8dN30WvQl-8Sirnp8AgRA?view#January-11-2023-Agenda
A
A
It's
a
great
way
for
you
to
come
and
listen
in
on
what
the
maintainers
are
working
on
meet
other
members
of
the
community
bring
up
anything
you
wish
to
discuss
about
any
of
the
cardboard
tools
if
you're
having
trouble
getting
started
or
if
you
have
any
feature,
requests
anything
like
that.
You
can
bring
that
up
with
the
team
in
this
meeting
or
you
can
just
simply
listen
in
and
participate.
A
A
If
you
aren't
able
to
join
us
for
any
of
these
meetings,
but
you
still
want
to
get
a
hold
of
us
there's
a
few
ways
you
can
do
that
we're
in
the
Carville
Channel
and
the
kubernetes
slack
workspace.
You
can
email
us,
you
can
find
us
in
GitHub
and
any
of
our
project
repos
and
you
can
find
us
on
Twitter
at
Carvel
underscore
Dev.
A
For
those
that
use
cardboard
tools,
we
would
love
to
know
more
details
about
how
you're,
using
our
tools,
so
we've
created
this
pinned
issue
here,
in
which
you
can
simply
add
a
comment
to
and
add
details
about
how
you're,
using
the
cargo
tools
and
information
about
yourself
in
the
organization
that
you
work,
for
that
is
using
the
cardboard
tools
and
optionally,
add
the
logo
for
that
organization,
and
we
would
get
it
added
to
the
adopters
file.
So
this
is
just
a
way
for
us
to
know
more
about
how
folks
are
using
the
tools.
A
A
A
On
to
announcements,
so
you
know
we're
still
working
on
onboarding
into
cncf.
It's
a
Sandbox
project.
For
the
most
part,
we've
got
pretty
much
everything
done,
except
for
the
big
thing
is
migrating,
the
all
of
the
repos
into
the
new
organization
to
Carvel
Dev
and
as
a
reminder.
What
we
mentioned
in
the
last
meeting
before
we
went
on
break
was
the
migration
will
have
an
impact
on
any
projects
that
are
using
the
cargo
tool
directly
via
go
libraries.
A
So
we
put
a
little
blurb
in
here
and
just
to
keep
that
everybody
aware
of
that,
and
you
know
once
we
have
done
all
of
this
we'll
make
sure
to
post
in
the
kubernetes
slack
Channel
as
an
early
warning,
so
just
something
to
keep
in
mind
as
we
get
all
this
taken
care
of.
We
hope
to
get
it
all
done
by
the
end
of
the
month,
but
we
will
see
based
on
any
blockers,
we
come
across,
etc,
etc.
So
any
questions
about
that.
A
Okay,
moving
on
to
content,
so
since
the
last
meeting
on
January
5th,
we
had
this
super
cool,
webinar,
video
kind
of
a
interview,
demo,
all
sorts
of
stuff-
it's
called
enlightening.
It's
a
VMware
show
and
Lee
Capelli
was
so
awesome
enough
to
to
do
that.
It
was
a
really
fun
demo
and
all
that
so
I
encourage
you
to
go
watch
that
he
goes
over
all
the
Carville
Tools
in
length
and
Whitney.
Who
is
the
host?
She
does
notes
the
same
time.
A
A
That
I
think
would
be
a
good
project
for
me
to
take
on
is
helping
those
outside
of
VMware,
and
even
people
who
work
at
VMware,
who
aren't
familiar
or
aren't
working
on,
Carvel
learn
how
to
become
a
contributor
and
what
that
path
to
maintainership
might
look
like.
So
you
know,
as
we
start
the
new
year
and
you
know
we
we're
working
on
all
sorts
of
stuff.
A
And
for
those
that
do
have
other
stuff
they
want
to
share
for
Carvel.
We
did
create
this
sign
up
sheet
of
sorts.
For
you
to
add
your
name
and
what
topic
you
wish
to
discuss
about
your
use
of
carbol?
If
it's
in
you
know
an
integrating
cargo
with
other
tools
and
like
how
they
work
together,
any
any
sort
of
interesting
thing
you
wish
to
share
with
the
community.
We
would
love
for
you
to
do
that.
B
C
B
D
D
So
what
happens
is
when
our
cab
doesn't
deploy
it
sort
of
leaves
a
config
not
behind
saying
that?
Oh,
these
are
the
changes
that
were
made
and
we
sort
of
limit
that
to
200
conflict
maps
and
I,
believe
cap
controller
override
thats
limit
and
sets
it
to
an
even
lower
limit,
while
deleting
when
a
delete
starts.
Cap
controller.
Sorry
cap
does
sleep
behind
a
conflict
map
like
that,
but
we
did
not
bother
setting
a
Max
limit
on
it
because
to
when
a
delete
exits
successfully,
we
garbage
collect
everything.
D
So
what
happened
was
we
ran
into
a
scenario
where
a
user
had
a
packet
installed
on
the
cluster,
which
was
stuck
enough
fail
in
delete
state
and
it
it's?
It
had
been
in
that
state
for
a
while
and
it
actually
sort
of
populated
the
cluster
with
a
lot
of
conflict
maps.
So
that's
something
we
plan
on
fixing
up
just
to
restrict
the
maximum
number
of
config
matters
left
behind.
D
So
reducing
unnecessary
list
calls.
So
this
is
something
tanufi
has
been
working
on
it's
so
before
cap
does
the
diff.
It
generally
lists
everything
on
the
cluster,
and
it
does
two
calls
one
to
see
if
want
to
list
everything
and
see
if
that
resource
exists
on
the
cluster
but
is
not
owned
by
Gap.
It
also
does
a
call
with
caps
label,
like
it
checks
if
there
are
any
Resources
with
that
the
cap
app
label
on
the
cluster,
but
it's
sort
of
not
needed
for
a
new
app.
D
So
it's
like
a
slight
enhancement
to
how
many
calls
you're
making
the
rejects
matching
is
something
we've
been
working
on,
keep
plan
on
closing
it
out
soon.
We
have
often
seen
news
so,
let's
say
you're,
defining
a
template
rule
or
a
rebase
rule.
I
think
rebase
rule
is
what
the
use
case
was
for
up,
which
says
how
you
treat
different
fields
in
the
config.
We
we
allowed
users
to
like
Supply
a
path
which
is
basically
like
an
array
of
strings,
but
what
folks
said
is
that
they
might
want
to
check
that
copy.
D
So
when
characters
are
diff
what
it
used
to
do
so
far,
is
it
used
to
copy
over
the
diff
from
the
existing
resource
on
the
cluster?
However,
there's
this
one
pretty
not
so
usual
case
that
over
over
different
versions
of
a
crd,
how
the
status
is
structured,
might
change,
and
what
is
happening
is
when
cap
was
trying
to
apply
an
older
version
of
the
status
it
was
erroring
out,
so
we're
walking
towards
completely
ignoring
status
while
doing
diffs,
because
that's
not
something
a
user
would
want
to
change
in
the
first
place.
D
Yep,
so
not
a
lot
of
new
stuff
going
on
in
key
control
right
now,
but
this
is
something
like
this
is
actually
something
Thomas
had
pointed
out,
I
think
a
week
a
couple
of
weeks
ago,
and
we
are
planning
on
sort
of
discussing
this
problem.
The
issue
itself,
so
I
I
think
Thomas
raised
two
issues.
One
was
regarding
how
K
control
should
be
creating
game
spaces
if
it
doesn't
exist
and
the
other
one
was
about
having
K
control,
be
like
the
First
Choice
when
it
comes
to
installing
cap
controller
and
I
think
yeah.
D
D
So
it
does
make
sense
to
sort
of
tell
the
users
that
okay,
if
your
namespace
is
not
created
already
right
now.
What
we
do
is
if
they
run
into
the
case
where
someone's
trying
to
install
in
a
public
namespace.
We
tell
them
that
you
either
change
whether
Cube
config
is
looking
or
you
create
the
namespace,
but
it
does
make
sense
to
just
be
like
specify
a
namespace
and
use
this
like
we'll
make
sure
it's
created
so
yeah
I,
guess
those
are
the
two
sides
I
was.
D
We
were
just
hoping
to
iron
it
out
on
the
issue
itself:
the
create
namespace
one
being
the
most
straightforward
one,
whereas
I
guess
there's
some
thought
process
doesn't
thought
some
thought
needs
to
go
into
the
cases
involved
when
we
are
installing
cap
controller
through
key
control.
Vegas
Thomas.
Do
you
want
to
add
something.
F
D
F
D
The
main
space
where
the
Gap
config
map
stays.
Actually,
this
conversation
is
in
context
of
package
installs,
so
let's
say
a
packet
doing
a
packet
install
in
a
particular
namespace
and
it
doesn't
exist
so
a
flag
which
just
says
created
for
me.
So
yes,
the
cap,
config
map
will
also
reside
in
that
particular
namespace
got
it
okay,.
B
Yeah
I
just
threw
this
one
up
here:
I
didn't
have
anything
to
do
with
it,
but
just
drawing
some
attention
to
joao
who's,
not
here
and
Alex
Barbados,
so
Alex
actually
submitted
the
issue
and
a
PR
as
a
way
to
resolve
this.
So
the
issue
being,
as
you
can
see
here,
if
you
run
the
image
package
list
command,
you
can
get
a
very
large
response
of
tags,
and
many
of
these
tags
are
these
ones
that
are
suffixed
that
end
in
the
dot
image
package.
B
So
they're
not
often
that
helpful
it's
for
humans
to
read
so
when
they
return
on
a
CLI,
it's
generally
not
helpful,
so
the
Alex
did
submit
a
PR
that
was
merged
and
I'm
sure
we'll
talk
about
this.
Some
more
once
we
actually
cut
the
image
package
release,
but
just
wanted
to
give
a
shout
out
to
Alex.
Thank
you
for
the
contribution.
A
D
So
to
give
some
context,
so,
let's
say
using
a
version
secret,
the
deployment
and
deploying
them
using
Gap.
The
expectation
is
that
the
versioning
affects
the
deployment
right
out
of
the
box.
However,
let's
say
using
a
k
need
of
custom
resource,
which
was
the
issue
that
came
up
and
it's
referencing
a
version
secret.
D
So
this
user
did
point
out
that,
since
K
native
has
like
stable
apis
now,
it
would
probably
make
sense
to
have
caps
about
K
native
out
of
the
box
and,
like
the
viewers,
are
also
thinking
about
that.
We
should
in
fact
start
integrating
with
tools
which
sort
of
a
lot
of
our
users
use
are
and
are
familiar
with
so
yeah.
That's
it
really
so
any
thoughts
around
what
the
barrier
should
be
or
how
we
should
go
about
these.
F
Yeah
I
couldn't
agree
more
I
think
if
there
are
the
fact
that
a
resource
is
defined
in
a
crd
and
not
by
the
project
itself,
doesn't
make
it
any
less
valuable,
especially
if
they're
just
really
pervasive,
then
they're
effectively.
That
way,
I
think
we
we'd
want
to
be
careful
to
like
invest
in
what
are
obviously
stable
apis.
F
So
that's
an
example
of
a
resource
that
we'd
kind
of
want
to
wait
for
the
dust
to
settle
there,
but
yeah
I
mean
I'm
sure,
there's
a
Litany
of
things
that
people
use
pretty
typically
thinking.
Looking
at
user
manager
like
that
kind
of
stuff
is
probably
pretty
common,
yeah
I,
don't
know
how
we
would
go
about
doing
that,
but
I
love.
This
idea
of
making
the
right
thing
easy.
F
D
A
A
A
A
G
So
yes,
that's
that's.
My
first
use
case
and
I've
been
very
happy
with
with
ytt
and
cap
controller
for
this
and
with
the
learning
curve,
of
course,
which
was
a
one
of
the
the
D
breaker
for
my
team
because
of
oit
imperative
syntax.
The
learning
curve
is
was
too
high
for
most
of
my
team,
so
we
agreed
that
I'm
I'm
encapsulating
most
of
ytt
into
libraries
into
star
libraries,
so
that
the
rest
of
the
team
could
eventually
use
the
same
product
dim
but
not
be
exposed
to
to
White
complexity.
G
I
continued
using
ytt
and
cap
for
running
acceptance
tests
using
a
curtor
because
we
don't
have
a
CI
we're
only
using
CD,
so
I
had
the
trouble
of
how
to
to
run
the
cutter.
Yeah
I,
don't
know
if,
if
you
know
about
cattle
is
a
decorative
tool
for
kubernetes
and
you
need
to
provide
a
decorative
yaml
files
to
to
deploy
kubernetes
resources
as
well
assets
and
so
I've
been
using.
G
With
this
report
for
immutable
configurations,
I've
been
using
ytt
to
deploy
the
cattle,
the
cattle
yaml
files
as
conf
within
config
Maps
and
and
to
mount
this
contain
bags
into
current
job
and
to
schedule
the
current
Japan
job
and
to
schedule
the
job
automatically.
Whenever
a
test
file
change,
the
config
map
is
versioned,
and
so
a
new
job
is
is,
is
launched
automatically.
So,
basically
my
test
whenever
I
modify
in
git.
One
of
my
test
file,
my
new
job,
is
triggered
and
launched
my
test,
so
this
has
been
really
useful
for
me.
G
So
it's
been
yeah
I'm,
still,
learning
I'm
still
new
to
to
cover
all
tools,
but
it's
been
very
nice
Journey
for
now
and
I
search
in
the
issues
and
in
slack
and
I
didn't
see
yet
mention
of
of
flux
users,
so
I
would
use
cover
and
I
I
have
the
information
that
there
is
quite
a
bit
of
opponent.
Is
there
to
expand
use
of
Carvel
beyond
the
cover
Community.
G
The
flux
team,
as
as
similar
aspect
to
Carvel
in
terms
of
git
sources,
which
kind
of
duplicated
in
similar
to
the
fetch
aspect
in
the
in
the
in
the
cap
controller
aspect.
So
it
looks
like
the
sources
are
very
similar
to
flag
sources
as
a
fetch
part,
and
the
template
seems
that
it
could
be
really
fitting
well,
a
flux
controller
that
would
do
yttd
templating,
just
like
the
flux,
handshot
or
the
flux.
Customized
controller
is
doing
customized,
templating
and
so
I'm
wondering
whether
there
wasn't
any
any
thoughts
or
similar
discussion.
G
F
F
G
And
I
wonder
whether
also
it
might
benefit
to
the
Carvel
Community
to
Leverage,
The,
modular
flux,
approach
of
adding
into
additional
input
sources
and
maybe,
and
maybe
as
well
notifications
so
to
I-
see,
for
example,
I'm
not
sure.
Yet.
If
cap
controller
supports
buckets
or
CI
images,
as
inputs,
for
example,
which
flux
does
and
so
maybe
there'll
be
one
way
to
expand
the
use
cases
of
carbo
by
leveraging
the
building
blocks,
that
the
flux
Community
is
contributing
as
well.
Yeah.
F
F
G
Yeah
I've
been
digged
so
much
into
a
flux
controller
to
to
have
an
idea
whether
that's
compatible
and
how
the
the
artifacts
from
The
Source
controller
are
obtained
and
the
communications
events
between
controller
or
how
how
all
of
this
fits,
but
yeah
that
would
be
idea
if
the
same
component
could
have
two
personalities.
That
would
be
awesome.
Thank
you
very
much.
I
really
appreciate
this
conversation.
Thanks.
F
Yeah
thanks
for
bringing
this
up,
this
is
really
valuable.
We're
also
looking
at
informally
around
ways
that
we
can
plug
into
the
Argo
experience
as
well.
So
this
is
very
relevant,
even
if
you
don't
hear
us
talking
a
lot
about
it.
Part
of
that
just
has
to
do
with
we're
trying
to
be
responsible
by
staying
focused
on
what
we've
committed
to
up
to
this
point,
but
yeah
I
would
love
to
talk
further
about
it.
I
don't
know:
I've
been
hogging
the
mic.
Anybody
else
have
thoughts
about
this.
F
G
Maybe
I
I
have
a
second
question.
If,
if
we
still
have
a
bit
of
time,
I'm
not
sure
yeah
is
the
cover
Community.
Does
the
cover
committee
has
experience
into
the
kind
of
CI
CI
tasks?
I
was
mentioning
like
running
tests
along
with
deployments.
Other
members
are
doing
that
and
there
are
ways
I
could
get
inspiration
from
their
work.
F
G
C
G
And,
and
specifically
one
one
trouble
that
that
I'm
currently
having
is
that
I'm
using
the
the
data
for
the
data
module
to
push
my
yaml
test
files
into
config,
Maps
and
I
would
like
to
find
a
ways
as
those
data
files
could
be
YT
templates,
while
being
not
individual
files.
Currently
I'm
using
this
as
a
folder
and
I
push
I
iterate
all
those
data
files
and
I
push
them
as
a
config
map
and
like
to
do
the
same
thing,
but
using
my
TT
syntax
within
them.
G
G
G
C
F
C
F
See
yes,
I
see,
okay
and
the
okay
interesting
yeah
I
mean
the
first
thing
that
comes
to
mind.
Is
we?
We
have
an
outstanding
idea
that
we've
been
kicking
around
about
being
able
to
emit
comments
out
of
ytt
templates,
which,
since
our
templating,
is
in
the
commenting
space,
it's
you
can
get
it
like
one
level
of
indirection
where
you
can
describe
what
the
template
would
look
like
in
comments.
F
G
C
G
F
Yeah
so
today
there
isn't
a
function:
you're,
not
missing
some
kind
of
function
where
we
say
here's
some
templating
evaluate
this.
For
me,
what
we
would
typically
do
is
somehow
provide
that
in
a
library
itself
and
then
evaluate
that
Library,
that's
how
we
are
exposing
the
templating.
The
overall
templating
experience
so
be
interesting
and
you're.
Talking
about
loading
up
those
files.
F
No
I
mean
I
mean
something
that
sits
underneath
the
ytt
underscore
lib
each
each
one
of
these
directories
is
considered
the
home
for
a
library
itself,
okay,
so
yeah,
maybe
maybe
kind
of
poking
at
that,
together
in
in
more
long
form,
format
of
a
conversation
be
happy
to
get
on
a
zoom
call
with
you
and
looking
for
that
'd
be
awesome.
Thank
you
very
much.
Okay,
cool
yeah,
let's,
let's
sync
up
afterward
and
all
try
and
find
some
time
for
us.
F
C
A
A
H
A
A
So
he's
been
very
helpful
for
the
development
of
Carvel
and
we
appreciate
him
so
much
so,
but
Scott
feel
free
to
look
at
the
video
afterwards
and
maybe
see
what's
going
on
so
guillermay
I'm
I'm
hopeful
he
he
is
given
a
lot
of
stuff
to
talk
about
today
and
super
appreciative
of
that
and
then
Thomas
provided
some
insight
as
well
and
we're
going
to
be
creating
some
issues
there.
So
super
fun
stuff
to
start
the
new
year.
I
really
appreciate
everybody's
contributions.
Here
it's
it's
good
stuff.
A
H
A
A
Okay,
well,
thank
you,
everyone
for
joining
today.
Thank
you
for
all
the
discussions.
As
I
mentioned,
it's
it's
super
appreciated.
We
we
love,
we
love
the
engagement,
it's
super
helpful
and
it
keeps
us
motivated
to
continue
doing
the
work
that
we
do
and
for
those
of
you
watching
from
home,
please
come
and
join
us
live.
We
would
love
to
to
hear
from
you
and
have
you
join
us?