►
From YouTube: Kubernetes SIG CLI 20200506
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
A
Our
agenda
is
pretty
packed
with
item,
so
let's
get
going
right
away
a
couple
announcements,
the
119th
release
dates
settle
down,
so
the
official
dates
are
enhancements.
Freeze
is
in
less
than
two
weeks,
May
19th
and
the
coach
briefs
is
planned
on
4
June
25th.
If
you're
interested
in
in
looking
through
the
dates
more
carefully,
there's
a
link
to
the
release
there
to
the
overall
release
date,
along
with
the
people
that
are
in
charge
of
the
release,
I.
B
A
C
Good
morning,
everyone,
okay,
so
so
outside
of,
why
doesn't
have
a
that?
Let
me
restart
we
saw
outside
apply,
we
don't
know
if
the
oysters
has
changed
or
not
when
it's
being
applied,
which
means
that
we
have
a
single
string
of
text
when
somebody
applies
from
cooker,
all
we
say
apply
or
something
which
is
different
from
client-side
apply
where
we
have
the
year
configured
versus
unchanged,
and
we
don't
know
how
important
these
theses-
and
we
want
to
know
if
we
need
to
do
something
about
it.
C
B
C
B
C
C
We've
had
one
person
create
a
ticket
saying
that
they
would
want
to
see
some
sort
of
status,
HTTP
status
code
item
by
the
API
server
to
know
if
the
host
has
changed
owner,
so
we
have
yeah.
At
least
we
have
suspicions
that
people
would
want
this
feature,
but
we
don't
see
why
they
would
want
to
use
this
feature.
B
You
know
even
just
from
a
human
perspective.
The
visibility
into
like
did
something
change.
His
plank
in
the
logs
is
gonna,
be
helpful
at
some
point
and
you're
trying
to
you
know,
track
down
an
issue
and
see
what
happened.
It
may
be
a
resource
change
that
you
didn't
expect
or
resource
you
expected
to
change,
didn't
like
all
that
stuff
is
useful.
B
C
A
B
C
What
we
have
in
mind?
Yes,
but
that's
that's,
a
change
in
API
machinery.
We
need
to
agree
on
that.
It's
fairly,
it's
not
something.
We've
done
before
I
see
how
it
will
have
value.
Even
client-side
apply
I
in
this
blog.
Specifically,
we
have.
We
have
situations
where
we
send
a
patch
and
the
patch
does
nothing.
So
we
show
us
configure
because
we
sent
a
patch
even
if
the
patch
does
nothing
and
people
are.
C
C
B
C
They're
also
from
the
API
machinery,
we
don't
want
any
people,
we
don't
want
people
to
use
these
to
trigger
specific
behavior,
because
we
think
it's
wrong.
It's
and
it's
an
entire
pattern
for
the
declarative
world
and
it's
it's
not
something
we've
done
in
the
past,
so
we
have
to
think
about
the
consequences
of
doing
that.
B
B
B
Yeah
I
guess
another
way.
Another
thing
is
like
whether
it's
correct
or
not
all
the
time
like
whether
or
not
a
resource,
whether
or
not
resources
were
updated,
is
returned
today,
like
I,
think
coop
control
exits
one.
If
it
things
had
tried
to
update
stuff
recoupment
roll
apply,
I
believe
exits
nons
into.
A
Another
thing
so
really
know
the
egg.
The
exit
codes
are
not
there.
That
was
discussed
last
time
during
the
Box
crop
that
we
would
lost
a
half
eggs
of
codes,
but
we
do
I.
Remember
that
insert
several
of
the
commands
and
proj
in
coded
mature,
both
apply.
We
have
a
condition
which
verifies
whether
the
patch
that
we
were
sending
over
to
the
server
empathy.
A
C
A
E
B
Then
now
you
want
to
build
a
system,
that's
composed
of
maybe
a
couple
of
these
things
right,
so
mixing
the
set
of
patches
and
in
config
map
and
secret
generators
for
a
live
database
right
or
mix
in
the
set
for
a
in
memory
database,
which
may
be
a
different
ones.
And
so
the
ask
here
is
to
effectively
be
able
to
define
a
customization
file
that
gets
composed
with.
B
B
The
ability,
some
components
I
think
the
kind
was
the
main
change,
you're
right
and
so
I
think.
There's,
certainly
a
discussion
to
be
had
of
is
the
right
way
of
expressing
a
component
through
kind
and
keep
it
in
resources,
or
is
it
better
to
make
it
the
customization
and
then
have
it
be
in
a
separate
field
or
combination
of
the
two,
both
a
separate
field
and
the
separate
kind
right
like
all
that
sort
of
stuff?
What
should
the
kinds
name
be?
B
Is
it
patch
or
is
it
component
I
think
I
think
all
that
stuff
is
stuff.
You'd
want
to
dig
in
before
certainly
needful
emerging
any
PR
or
accepting
any
feature,
but
it
probably
should
be
deferred
until
after
conceptually
that
the
mechanics
of
the
thing
are
agreed
upon
and
that
B.
This
is
a
feature
that
we
want.
I
mean.
G
Hey
force
off,
thank
you
so
much
for
taking
the
time
to
write
out
a
detailed
example
and
a
proposal.
So
no
I
have
not
had
the
time
to
go
through
this
thing
before
this
meeting,
but
I'm
going
to
go
through
it
this
week.
The
the
wrong
issue
that
spawned
this
this
the
there's
been
a
couple
of
issues
that
were
titled
long
lines
of
support,
composition,
customized
and
I've
engaged
on
those
threads
and
tried
to
boil
things
down
to
what
the
top
you're
trying
to
solve
and
then
I
would
pose.
G
G
G
G
I
got
like
this
thing:
I,
don't
know
if
you
can
see
this
thing,
but
my
USB
port
I
only
have
one
USB
port
on
this
ancient
laptop
I'm
using
and
that's
going
to
my
mouth
because
I
like
mice,
okay,
so
I
just
rambled
a
long
time,
but
my
general
approach
has
been
try
to
fix
it
using
what
we
have
and
then
prove
that
it's
awful
and
can't
work.
And
then
we
do
the
new
thing.
G
B
G
If
it
turns
out
it's
possible,
but
it's
horrible
right,
then
that's
that's
exactly
what
I
want
to
look,
but
I
want
to
look
for
a
demonstration
that
so
there's
two
things.
Is
it
horrible
right
and
also
is
this
a
is
what
we're
the
problem
we're
trying
to
solve
an
example
of
a
best
practice
we
want
people
to
who
we
think
is
good.
Then
we
believe
in.
B
Thinking
about
what
you
just
said,
I'd
probably
say
that
if
it's
possible
to
do
with
the
framework
and
painful-
that's
probably
the
ideal
case,
we're
looking
for
because
it
means
yes,
it's
in
scope,
it's
not
something.
That's
just
like
totally
out
of
scope
like
it
can't
be
done
with
the
frame
like
some
random
thing.
It
can
be
done
with
the
framework,
but
the
mechanics
of
doing
so
are
can
be
improved
right.
That's
right!
Right.
G
Well,
I
mean
it
again
would
boil
down
to
a
specific
case,
because
because
there's
many
many
examples
of
multiple
overlays
that
modify
the
same
base
in
our
high-level
unit
tests,
high
level
feature
tests
and
an
example.
So
it
is
a
special
I
mean
it
is
a
particular
case,
I'm
sure,
there's
a
particular
case
where
you
have
trouble.
So
that's
what
I
want
to
dig
into.
E
G
Yeah
I
mean
the
path
forward,
is
clear:
let's,
let's,
let's
go
ahead
and
make
a
attempt
to
build
it
satisfy
that
use
case
that
you
have
and
see
what
the
answer
looks
like
and
if
it's
really
really
ugly,
then
maybe
we
need
to
introduce
the
the
customization
patch
I
mean
it
sounds
plausible.
It
sounds
like
a
good
idea.
I'm,
not
saying
it's
a
bad
idea.
It
sounds
like
a
good
idea.
I
just
want
to
erect
a
barrier
to
adding
new
features
to
the
language.
If
you
will
right
yeah.
E
E
There's
nothing
I
think
I'm,
not
sure
Jeff.
Maybe
you
seen
this
tomorrow,
I'm,
not
sure
if
a
lot
of
people
use
this
feature
of
customized,
where
if
you
points
to
the
same
base
from
different
overlays,
but
they
use
a
different
name,
reflux
or
suffix
or
namespace,
but
you
want
you
with
one
command
deployed
to
different
of
three
different
applications:
I'm,
not
sure
it's
that
much
used
examples
to
be
frank,.
E
E
The
erroneous,
no,
the
thing
that
I
believe
it's
not
common
is
that
thinking
customized,
where
to
overlays
different
overlays
point
to
the
same
base,
but
it's
a
different
name
prefix
and
thus
deploy
to
different
applications.
In
our
case,
it's
it's
one.
Application
that
has
been
edited
by
two
different
staff
notes:
two
completely
different:
no,
no
to
the
notes
of
the
cases
with
different
names,
the
same
application
with
two
different
names:
I'm,
not
sure
how
can
explain
it
there.
D
E
B
Yeah
I,
don't
it'd,
be
great
to
have
more
usage
statistics.
I.
Think
Jeff
is
actually
working
on
for
some
time
developing
a
project
which
will
get
data
from
github
on
different
custom
like
instances
of
people
using
customize.
So
maybe
that
we
could
use
that
data
to
figure
some
of
this
stuff
out.
It.
B
B
D
B
Is
correct
right
without
going
into
the
details
of
how
its
represented
specifically,
then
the
next
step
would
be
to
let
me
post
that
back
to
any
issue,
because
there's
a
bunch
of
people
on
the
issue
and
a
bunch
of
conversation
and
just
close
the
loop
there
and
say
this.
Is
the
plan
fixed
for
the
problem
in
this
issue
right
and
then
and
then
at
that
point
in
time,
dig
into
the
fine.
G
Details,
yes,
so
sorry
to
interrupt,
but
concretely
so
chime
and
try
to
provide
an
actual
example
in
the
existing
try
to
solve
the
problem
statement
in
the
existing
framework-
and
you
know
see
if
that's
a
possible
and
B
how
ugly
it
looks
and
then
that
would
be
it
then
the
next
step
would
be.
Everybody
decides
what
to
do.
I
don't
really
want
to
decide.
Anything.
I've
want
to
first
provide
an
example
of
what
the
what
the
alternative
isn't
our
existing
framework
before
we
expand
it.
B
Parts
of
customization
files
around-
and
so
you
have
you
know
you
have
customized
files
that
were
like
the
way
you
do
manual
composition
right
is,
instead
of
referencing
a
customized
file,
that's
shared
and
composed
into
three
different
other
customized
files.
You
just
copy
and
paste
the
contents
of
that
file
into
three
different
places.
Is
that
correct?
That's.
G
B
So
that
would
be
super
helpful
and
then
is
there
for
posterity
and
people
know
you
can
do
it
like
if
this
was
a
blog
post,
so
I
love
the
way
you
wrote
it
and
like
it's,
a
great
is
user
documentation
and
maybe,
along
the
same
theme
of
how
writing
proposals
in
the
form
of
the
sort
of
documents
our
end
users
will
eventually
consume
like
a
blog
post.
If
this
was
written
as
a
blog
post,
it
would
say
like
before.
You'd
have
to
do
such-and-such
right
now
we
have
introduced
this
new
feature
that
provides
this
right.
B
G
E
B
Think
you've
been
to
it
right,
like
you,
put
a
lot
of
effort
into
this
already
and
it
looks
great
but
just
making
it
clear.
So
everyone
knows,
like
maybe
a
paragraph
that
you
know
with
a
couple
of
lamp
points
and
say
you'd
copy,
like
four
or
a
table
that
says
where
things
are
copy
and
paste
it
or
something
like
that,
or
do
the
pull
examples
whatever
it
is,
but
yes
write
something.
A
B
Cool
I
mean
this
doesn't
go
in
the
notes,
but
Alex
it's
it's
worth
noting
that,
like
the
the
infrastructure
customize
is
built
on,
is
Ben
is
going
through
it
like
every
right
and
it's
quite
far
along.
So
we
just
want
to
make
sure
that
if
we
were
to
implement
this
proposal
that
it
would
be
implemented
in
the
new
style
so
that
we
don't
have
to
migrate
that
as
well.
So
we
can
follow
up
out-of-band
about
what
that
looks
like.
E
B
A
At
least
from
my
loose
understanding
of
the
differences
of
your
proposal,
versa,
what
currently
proposed
isn't
the
difference,
a
matter
of
how
much
actual
customization
you
will
have
her
per
deployment,
slaver,
meaning
its.
If
there's
only
one
thing
you're
changing
in
between
the
different
deployment
flavors,
then,
obviously
the
composition
would
be
in
favor,
but
if
more
than
50%
of
the
customized
file
is
different
in
each
case,
I'm
sure,
if
composition
will
help
you
with
this
and
then
the
old
approach
will
be
better,
am
I
missing,
something.
E
Let's
say
that
you
want
to
enable
a
feature
for
an
application
and
some
options.
You
don't
want
you
once
this
options
change.
You
don't
want
to
grab
her
and
said
your
way
around
this
issue
across
overlays.
You
want
to
change
this.
Only
in
one
location,
it's
handling
the
it's
it's
drying,
it
drying
your
place
and
moving
things
from
just
one
location.
B
Yeah
I
can
imagine
like.
If
you
have
is
I
mean
it's
like:
do
you
copy
and
paste
code
or
create
a
function
that
you
know
it
has?
The
shared
code
is
kind
of
my
I
guess
the
metaphor
analogy
right
and
it's
like
well.
It
depends,
but
certainly
like
in
the
case,
provided
in
the
example
like
if
the
reCAPTCHA,
for
instance,
variant
required
a
new
patch
or
something
like
that
right.
If
there
were
three
variants
that
use
the
reCAPTCHA,
that
required
recap
should
be
enabled
for
them.
You
would
not.
B
E
A
D
D
You
should
not
be
able
all
right,
that's
good
enough,
so,
as
you
may
know,
QC
TX
is
this
project
that
is
semi
popular
because
switching
between
clusters
and
namespaces
actually
something
that
people
do
pretty
frequently,
but
in
QC
TL
as
far
as
possible,
it's
kind
of
lengthy
to
write
that
command
to
actually
switch
between
namespaces
and
contexts.
So
up
until
this
point,
this
project
was
actually
written
purely
in
bash.
D
It
had
a
pretty
good
success,
although
it
was
actually
kind
of
slow,
because
we
were
shelling
out
to
keep
control
under
the
covers
and
we
were
doing
multiple
calls
and
then
which
in
turn
read
the
config
file.
As
you
know,
so
what
I
did
was
I
basically
thought
about
writing
the
single
for
a
while
and
I.
Finally
did
it
and
the
current
version
is,
as
expected,
it's
a
bit
faster.
So
if
I,
actually,
you
know
if
I
were
to
show
you
how
this
project
works
is
basically
the
program
that
you
run
and
you
can
switch.
D
You
know
back
and
forth
between
contexts
or
you
can
have
batch
completion.
It's
not
a
little
here
was
normally
on
on.
You
know
you
can
choose
between
context,
etc.
So
is
it
you
know
what
I
actually
wanted
to
share
is
something
that
we
do
in
this
version,
which
is
we
actually
do
now,
have
the
capability
to
edit
the
yellow
files
directly
without
losing
their
format
or
contents.
So
we
totally
respect
the
user
in
terms
of
what
they
do.
D
So,
if
I,
for
example,
if
I
come
here
and
if
I
actually
like
write
an
example
comment
here,
if
I
exit
and
like
switch
the
context
again
and
if
I
look
at
it
again,
you'll
see
that
the
comment
is
still
there
and
the
way
this
works
is
under
the
covers
there
is
this:
go
I'm,
a
library
that
you
might
know
from
Gustavo
new
mayor
and
this
library
has
a
new
version.
That
is
not
the
default,
it
is
the
v3
version,
and
this
retrieve
version
actually,
if
I
can
find
the
API
documentation.
D
This
new
version
exposes
a
type
that
lets
you
directly
deal
with
llamó
without
on
marshaling
it
to
the
final
strut.
So
there's
this
thing
called
node
and
you
know
you
can
actually
control
each
field
and
which,
what
content
does
it
have?
What
sort
of
string
it
is,
what
sort
of
you
know
multi-line
string?
It
is,
and
you
can
control
them.
You
think
you
know
it
set
the
position
of
the
fields
you
can
preserve.
The
ordering
comment
so
I
believe
this
customized
actually
used
this
by
the
way
I
was
gonna.
Ask
ya.
B
D
D
B
B
B
G
Yeah
I
can
I
can
chat
on
this,
so
the
underlying
mechanics
of,
manipulating
yamo,
as
you
know,
llamo,
is
a
kind
of
a
bizarre
wrapper
around
Jason
and
has
all
these
funky
things
that
people
may
be
used.
Maybe
don't
use
like
back
references
and
whatnot,
so
getting
a
good
underscore
underscore
underlying
library
that
allows
us
to.
G
You
know
manage
feels
order
and
retain
comment
structures,
as
we
you
know,
shove
that
yamo
through
some
kind
of
transformation
and
get
it
aside,
is
you
know
the
the
reason
why
we're
rewriting
the
rebasing
customized
on
a
much
more
powerful
yamo
library,
so
yeah,
it's
gonna,
be
a
handy
for
any
sorts
of
application
like
this,
where
you
want
to
mess
with
the
mo
maintain
comments.
Yes,
yeah.
D
A
A
D
So
I
would
say
that,
like
there
are
other
things
that
touch
cue:
config
files
like
the
g-cloud
command
line
tool,
I
think
other
clock
writers
have
similar
tools.
I
think
all
of
them
are
probably
going
to
destroy
that
file.
Although
I'm
hearing
in
the
community
that
there's
there
are
a
lot
of
like
handcrafted
files
that
you
know
operations,
people
get
from
their
admins
or
something
like
that.
A
A
D
I
saw
the
demo
for
C
gaps
for
cats.
Actually
I
think
my
use
case
is
a
little
bit
different
in
this
case.
My
main
problem
is
like
how
cute
control
actually
reads
from
multiple
cube,
counting
files
and
how
which
one
does
it
choose,
which
won't
write
to
that's
something
that
we'll
be
looking
at
next,
because
that
part
is
like
a
little
bit
undefined
in
terms
of
behavior.