►
From YouTube: Carvel Community Meeting - May 17, 2021
Description
Carvel Community Meeting - May 17, 2021
We meet every Monday at 11:30am PT / 2:30pm ET. We'd love for you to join us live!
Details and notes from this meeting can be found here: https://hackmd.io/F7g3RT2hR3OcIh-Iznk2hw#May-17-2021-Agenda
A
A
These
meetings
are
being
recorded
and
uploaded
to
our
carvill
youtube
playlist.
So
if
you're
watching
from
home
after
this
has
already
been
recorded,
welcome,
we
hope
that
you
are
able
to
join
us
at
a
future
meeting.
We
do
meet
every
monday
at
11,
30
a.m,
pacific
time,
2,
30,
p.m.
Eastern
time
we
also
offer
office
hours
in
addition
to
those
community
meetings,
and
we
meet
every
second
and
fourth
thursday
of
the
month.
A
So
we
will
be
meeting
next
thursday
at
11
30
a.m,
pacific
time,
2,
30,
p.m,
eastern
time
and
that's
just
an
opportunity
for
anything
that
we
don't
get
to
from
the
meeting
and
requires
further
discussion
or,
if
there's
anything
that
you
wish
to
bring
to
the
team
anything
that
you
need
help
with
with
karvel
or
if
you
have
any
specific
questions,
that's
a
great
place
to
bring
that
up
as
well.
A
If
you
aren't
able
to
attend
the
these
community
meetings,
when
you
do
attend
these
meetings,
we
ask
that
you,
please
add
your
name
and
if
there's
an
organization
that
you
represent
for
using
cargo,
please
add
that
in
the
attendees
section
here
for
announcements
today
we
have
a
release
of
ytt.
A
B
Yeah
this
one
said:
there's
a
call
out
here.
We
really
want
to
thank
han
lin
she
for
his
contribution.
It
wasn't
just
adding
another
function
like
that's
the
the
net
result,
but
actually
like
help
carve
out
how
we
can
have
domain
specific
libraries.
This
is
just
one
of
them
and
help
set
a
pattern
for
be
able
to
do
that.
So
we're
super
excited
about
that
contribution.
It's
sort
of
an
ordinary
thing
and
it's
stuff
we'll
be
adding
to,
but
it
also
sort
of
sets
the
stage
for
having
other
things.
That
will
be
useful.
A
C
D
Yeah
I'll
provide
a
quick
update.
The
current
work
in
flight
is
a
story
to
annotate
a
schema
node
as
any
type
part
of
our
experimental
schema
feature
and
then
adding
to
documentation
around
it.
So
how?
How
does
this
work
and
what
is
currently
supported
in
schemas
and
then
just
there
is
an
epic
in
zenhub
right
now
and
you
can
see
sort
of
the
progress
that's
being
made
as
part
of
like
our
first
mvp
for
ytt
schemas
and
some
of
the
work
left.
There
is
some
usability
testing
and
documentation.
C
E
We're
planning
to
cut
a
release
today
that
will
include
the
open
api
schema
changes,
so
folks
can
start
consuming
that,
and
then
we
have
the
pr
to
split
the
package
cr
in
review
right
now,
so
hopefully,
a
release
with
that
change
will
be
coming
later.
This
week.
C
F
Yeah,
so
the
release
should
already
be
in
github,
but
we
haven't
officially
just
spread
out
the
word,
but
this
release
is
quite
big
because
it
solves
two
bugs
that
we
that
were
found
in
image
package
and
also
it
introduces
two
braking
changes
to
image
package,
one
that
is
related
to
the
default
tags
that
image
package
adds.
F
So
this
should
not
impact
anyone,
but
it
will
change
the
tagging
system
for
image
package
and
the
other
breaking
change
is
that
now,
when
you,
when
you
create
a
bundle
with
in
the
files
that
are
part
of
that,
bundle,
contain
permissions
to
execute
now
that
is
copied
through
the
permissions
are
copied
through,
which
means
that
if
you
use
the
prior
version
to
create
a
bundle
in
this
new
version,
you'll
get
two
different
chats,
but
that's
the
the
only
different
difference.
It
is
a
breaking
change.
It
is
not
gonna
break
anything
that
you're
doing.
F
Currently
it
just
changes
the
shots.
So
those
are
like
two
things
that
I'd
like
to
call
out.
Then
we
have
some
some
ux
improvements
and
yeah
and
we
have
two
bugs
there
and
when
we
get
to
released
out,
we
can
talk
a
little
bit
more
about
maybe
those
bugs.
C
C
Let's
see
so,
we
got
a
lot
going
on,
but
maybe
we'll
take
a
look
at
what
stories
need
to
be
discussed
for
this
week,
so
maybe
I'll
start
by
filtering.
So
it's
a
little
easier
to
read
so
filtering
on
the
ytt
schemas
we
currently
have
schema.
Node
can
be
annotated
to
allow
any
type
in
flight,
as
well
as
the
documentation
story
up
next
is
usability.
C
C
Cool
then,
after
that
is
the
remove
the
schema's
features
flag,
but
this
is
probably
good
usability
testing,
usually
surfaces.
Some
improvements.
C
B
Yeah
we
got
some
some
feedback
that
one
of
the
stories
was
large
and
so
actually
broke
up
that
story
into
two
in
the
interim.
So
and
actually
I
had
marked
and
marked
the
story
that
was
broken
up
to
be
99
points,
because
there
is
no.
There
is
no
like
way
to
remove
a
point
in
zenhub,
so
I
just
created
something
I
thought
would
be
sort
of,
obviously
not
up,
not
the
right
points.
B
So
that's
this
one,
it's
back
to
three,
but
but
the
intent
here
is
to
sort
of
revisit
this
with
a
piece
of
it
removed
and
then
there's
a
subsequent
story
with
the
portion
that
was
extracted,
so
it's
at
least
go
at
least
worth
going
over
this,
and
maybe
we
can
throw
points
on
these
two
stories.
Is
that
cool
okay?
B
We
will
merge
data
values
when
they're
coming
from
this
plain
emo
channel.
So
remember
this
is
the
part
about
where
data
values
are
coming
in
through
this
dash
dash
data
values
file
flag.
So
when
we
give
multiple
of
these,
how
should
it
behave.
B
And
so
this
this
decomposition
falls
through
this.
This
one
dimension
here
of
whether
or
not
the
the
two
documents
that
are
passed
in,
are
either
perfectly
matching
in
terms
of
structure
or
they're
perfectly
missing
in
and
and
so
it's
just
the
union
of
the
two.
B
So
the
intent
here
is
that
there
aren't
any
weird
cases
that
will
support
in
this
version
of
the
feature
like
oh
someone,
you
provide
a
data
value
that
was
an
integer
over
here
and
now
it's
going
to
be
a
boolean
over
here
or
str,
or
a
array
over
here,
and
now
it's
going
to
suddenly
turn
into
a
map
like
that's
out
of
scope
for
this.
This
story
that
would
make
it
incongruent.
B
So
this
is
this,
brings
a
simplifying
assumption
that
the
full
set
of
data
values
that
the
user
will
provide
will
will
merge
well
in
this
way
and
and
what
they
would
get
otherwise
would
be
a
standard.
Whatever
the
error
message,
they
would
get
today
when
you
try
and
combine
these
say
if
you
were
using
overlay,
so
would
say:
hey
you're,
trying
to
overlay
a
map
on
top
of
an
array,
and
that's
not
allowed
something
like
that.
B
So
there's
some
existing
behavior
that
we
that
we
would
lean
on
potentially
for
this.
So
there's
a
couple
assumptions
baked
in
there
one.
I
guess
I
just
sort
of
realized
clearly
assuming
that
we're
using
overlays
as
the
merging
mechanism
as
opposed
to
some
custom,
one
which
I
think
is
an
okay
assumption
and
and
two
that
the
thing
that
makes
it
incongruent
is
where
the
the
the
structure
is
is
different.
B
Okay,
so
I
think
there's
two
cases
we
can
scroll
through
them.
This
first
one
is
where
you
see
the
bar
and
bar
here
and
values
and
values:
two
are
both
arrays
and
it's
just
that
the
whatever
the
value
that
was
for
value
two
that
replaces
it
so
that
that
is
interesting
and
different
than
like
sort
of
your
default
overlay
behavior.
So
that's
something
in
there
and
here
in
the
second
example,
you
can
see
where
it's
just
we're
completely
disjoint
set.
So
it's
like
perhaps
a
use
case
to
motivate
this
would
be.
B
Someone
has
a
very
large
set
of
data
values
and
they
want
to
break
it
up
into
files
to
make
it
easier
for
the
consumer
to
provide
them.
It's
like.
Oh,
this
is
all
your
system
configuration
your
network
configuration.
These
are
your
image
configuration
be
able
to
like
provide
those
in
separate
files,
and
they
would
just
like
combine
easily.
B
Like
in
the
bigger
picture,
yes,
for
the
sake
of
this
story,
it's
out
of
scope,
so
we're
not
trying
to
support
that
if
something
quote-unquote
bad
happens,
that
that
that's
a
degraded
or
or
degraded
experience
that
you
know
you're
using
you're
using
that
feature
in
an
undocumented
way.
At
the
moment,
you
can
think
of
it
that
way,
but
in
the
bigger
picture,
absolutely
so,
the
these
data
values,
if
you
think
about
how
data
value
pre-processing
occurs.
B
After
all,
the
other
kinds
of
overlays
are
processed
the
data
values
that
get
set
from
the
command
line
or
from
the
environment.
Those
are
sort
of
last
in
line
to
get
applied
in
the
data
values
pre-processing
algorithm.
This
is
one
of
those.
This
is
just
extending
that
set
of
command
line
parameters.
G
F
Considerations
about
schemas
what
should
happen,
should
it
just
fail
for
now?
Should
it
just
whatever
happens
happens
when
you
provide
them
minus
minus
the
enable
schema's
flight?
Oh,
if
you
turn.
F
Okay,
is
there
like
any
subsequent
story
that
talks
about
the
order
that
these
should
be
applied,
so
we
do
have
environment
variables
and
yep,
and
also
you
can
provide
the
values
there.
There
are
stories
afterwards
that
say
where
this
sits
right.
B
Yes,
we
don't
have
one
specked
out
yet,
but
but
there
there
is
one
that
will
do
that
when
it
talks
about
the
ordering
with
respect
to
kerry's
question,
which
was
data,
values
overlays.
So
yes,
we
will
cover
that.
B
B
F
F
D
D
So
I
think
it
makes
sense
to
considered
data
values
flags
last
since
they
function
as
sort
of
like
overlays.
Now.
B
This
is
actually.
This
is
a
really
important
conversation,
but
I'm
thinking
like
the
scope
of
what
we're
talking
about
right
now,
is
like
what
we
need
to
know
in
order
to
point
the
story,
so
there's
overlap
in
the
sense
of
like
okay.
Well,
you
get
to
the
point
where
you're
combining
things,
what
would
it
be
and
we're
not
trying
to
necessarily
in
this
story
to
find
that
behavior
per
se,
but
I
get
it
that
like
that,
could
be
an
easy
win.
B
So
would
you,
but
just
since
you
kind
of
cracked
that
can
just
a
little
bit
carrie
really
really
quickly.
On
that
note,
there's
there's
a
undocumented
feature
where,
if
you
add
a
plus
in
front
of
the
key
name
for
a
dash
data
values
invocation
like
for
the
key
name,
if
you
added
a
plus
in
front
of
it,
it
actually
will
do
it'll
do
a
insert
also
a
missing
okay,
equal
true,
so
that
exists
yeah
exactly
and
then,
and
in
fact
I
think,
steady
state.
B
Just
sort
of
blindly
add
and
the
intent
would
be
that
by
the
time
that
happens,
schemas
are
online
in
in
the
support
that
you
used
to
get
with
data
values
overlays
the
kind
of
checking
that
you
used
to
get
from
that
you'll
get
a
better
version
of
that
from
schema,
because
you'll
get
structure
stuff
with
nicer
messages
and
the
typing
stuff
with
nice
images,
so
yeah,
okay,
yeah
you're,
all
poking
at
the
right
stuff,
though
like
this
is
relevant.
B
Yeah,
good
chatter,
okay,
let's
try
see
where
our
let's
try
and
throw
points
on
this
and
see
what
kind
of
spread
we
have
if
any
had
a
chance
to
think
about
this.
Looking
like
everybody's
ready
to
go
one,
two,
three:
okay,
rough
consensus:
it's
a
two.
E
B
B
B
B
G
G
B
B
B
This
is
just
saying
this
is
a
this:
is
a
continuation
or
related
to
that
previous
story.
That's
all.
That's
meant
to
indicate.
F
Oh
okay,
because
in
these
two
scenarios
you
we
always
convert
to
race
two
maps.
So
I
was
just
thinking
that
that
sentence
meant
there
were
other
like
that.
We
would
do
like
maps
to
like
we
do
maps
to
arrays
or
maps
to
scalars
like
in
a
different
story
or
not.
But
we
assume
that
all
types
will
go
to
all
types
in
in
this.
D
B
If
we,
if
we
replaced
it
right
so
today,
try
that,
with
the
current
current,
the
way
overlay
works
today
and
it
would
fail
because
it
would
say,
expected
a
map
but
actually
got
an
array,
and
so
that
was
backward.
That
was
that
would
fail
that
way.
Then
you're
like
oh.
I
want
to
replace
that.
So,
if
you
throw
in
an
overlay
replace
over
re,
then
you
get
the
result
that
you're
looking
for
you
get
a
merge
of
olive
bar
and
a
replace
of
re,
and
so
that's
not
necessarily.
B
I
don't
mean
just
to
share
that
in
order
to
say
well.
This
is
how
we
should
implement
it
that
there's
a
possibility
of,
like
maybe
that's
part
of,
maybe
that's
what
makes
sense
like
there's
there's
some
design
work
here.
B
Is
this
something
where
there's
possibly
two
modes
in
which
overlay
the
overlay
up
feature
itself
works,
because
we
think
we
want
to
keep
the
existing
behavior.
We
definitely
want
to
keep
the
existing
behavior
for
now.
B
So
we
don't
want
to
just
suddenly
change
the
way
that
overlays
work
so
either
it's
a
mode
that
we
say:
hey
overlays
when
you're
merging
use,
simple
merging
and
it's
like
okay,
I
won't
complain
if
the
structures
are
different,
I'll,
just
like
replace
or
you
may
decide.
I
want
the
complexity
to
be
in
the
whatever's
trying
to
calculate
these
overlays
from
these
inputs
and
say:
okay,
we'll
have
to
go
through
and
if
there,
if
there's
structural
types,
they
replace
somehow
or
something.
B
G
B
C
F
F
We
on
image
package
try
to
see
if
the
images
that
the
images
that
are
part
of
the
bundle
are
co-located,
with
the
bundle
to
make
sure
that
we
can
update
your
images
lock
to
point
to
the
new
locations
of
the
of
those
of
those
images,
so
that
you
can
have
a
more
up-to-date
images
lock.
F
If
the
images
are
behind
a
behind
a
registry
that
has
authentication
the
pool
will
fail.
C
F
Don't
have
authentication
to
find
the
images.
Nevertheless,
image
package
doesn't
really
need
to
have
access
to
these
to
these
images
in
order
for
it
to
be
able
to
pull,
because
when
you
do
a
pool,
the
only
thing
that
we
do
is
we
pull
the
the
the
contents
that
are
part
of
the
oci
image
of
the
bundle
right.
So
this
is
the
scenario
that
that
we
encountered
ourselves
with
so
this.
F
This
brings
some
issues,
as
we
can
see
here
right
and
the
solution
that
we're
trying
to
propose
is
a
solution
that
is
based
on
our
renaming
copy,
with
renaming
of
images
where
every
time
that
you
copy
an
image
every
time
that
you
copy
the
geocopier
bundle
image
package
would
save
the
location
where
all
the
images
are
being
copied
to.
F
So
this
information
will
be
present
in
a
place.
So
now,
when
you
do
a
pool,
the
only
thing
that
image
package
needs
to
do
is
did
I:
it
finds
a
bundle
it
finds
this
other
place
that
contains
all
the
locations
of
the
images
and
that's
it,
and
it
can
just.
He
knows
that
if
he
needs
to
pull
things
or
if
he
doesn't
need
to
pull
anything.
G
B
F
F
So
I
don't
know
if
we,
if
you
want
to
go
and
point
this
story
right
now
or
not,
but
this
is
the
general
idea
of
what
the
story
it
talks
about
and
for
us
to
know
that
there
will
be
another
place
that
is
going
to
be
also
an
oci
image
that
is
going
to
be
in
the
same
place
that
the
bundle
is
in
the
same
repository
that
the
bundle
is
so
when
you're
doing
a
pool.
F
You
will
you'll
have
access
to
that
image,
and
if
the
image
does
not
exist,
it
means
that
the
bundle
was
not
copied.
So
we
can
just
say
like
okay,
fine,
we're
good,
we
don't
need
to
go
and
try
to
access
all
the
images,
so
this
also
solves
another
problem
that
is
linked
further
a
little
bit
further
down
and
that
the
pool
was
a
o
n
operation,
because
every
time
that
we
wanted
to
get
the
the
pulling
the
the
image
pulling
a
bundle.
F
You
have
to
go
to
the
bundle
and
then
we'd
have
to
go
to
all
the
images
so
by
having
all
this
information
concentrated
in
another
another
image.
You
just
need
to
do,
go
to
the
bundle
go
to
this
image
and
if
the
image
is
present,
then
it
means
it
was
copied.
So
we
moved
from
a
on
scenario
to
a
to
a
scenario
where
we
just
need
to
go
two
times
to
the
registry,
so
this
is
more
or
less
the
story
that
we
have
here.
F
There
are
some
use
cases
like
some
scenarios
that
need
to
be
added
to
it.
So
yeah
there's
one
scenario,
and
that
was
it
10
days
ago,.
B
C
F
So
I
don't
I
don't
know
if,
if
y'all
want
to,
maybe
maybe
we
need
more
a
little
bit
more
detail
on
what
we
expect,
like
the
use
cases
that
we
have
here
in
order
to
point
a
story,
I
think,
but
at
least
now
you
have
like
a
general
idea
of
this.
What
is
the
story
is
about
I'll
try
to
get
it
ready
for
next
week
for
us
to
be
able
to
point
it
out.
C
Let's
see,
is
there
a
next
story
that
would
be
helpful
to
discuss?
I
know
ytt
is
after
that,
as
for
non-epic
stories,
because
I
could
do
not
that
or
was
that
not
in
an
epic.
C
Ytt
uht
controller
cap,
you're,
probably
not
gonna,
get
the
cap.
Did
anyone
add
any
topics.
C
C
A
Sure
thing
so
yeah,
I
everyone!
Thank
you
for
joining
this
week's
edition
of
the
carville
community
meeting.
I
did
add
something
to
the
announcement
section
that
occurred
last
friday.
Kubecon
europe
released
their
videos
on
youtube,
so
I
added
the
link
to
carville
session
up
at
the
top,
so
please
be
sure
to
check
that
out
and
I'll
also
link
it
into
the
youtube
description
so
that
you'll
be
able
you're
able
to
go
and
check
out
that
session.
A
If
you
feel
you
want
to
join
us
live,
we
have
a
few
opportunities
for
you
to
join
us.
We
meet
every
monday
at
11,
11,
30
a.m,
pacific
time,
2,
30,
p.m,
eastern
time
for
our
community
meetings
and
then
we
we
meet
for
our
office
hours
every
second
and
fourth
thursday
of
the
month
and
so
we'll
be
meeting
next
thursday
at
11
30
a.m,
pacific
time,
2,
30,
p.m,
eastern
time
and
that's
an
opportunity
for
more
in-depth
discussions
and
anything
you
may
need
help
with.