►
From YouTube: Kubernetes SIG CLI 20190925
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
Okay,
good
morning,
good
afternoon,
good
evening,
depending
on
where
you
are
today,
is
September
25th,
and
this
is
our
bi-weekly
six
6cl
I
call.
My
name
is
Monty
and
I'll
be
your
host
today
and
before
we
dive
into
our
main
topics,
which
is
planning
for
117
a
quick
announcement,
slash
reminder
from
the
previous
sessions.
We
are
holding
a
six-year
line
90-minute
session
during
Kubik
on
North
America
in
San,
Diego
I
put
the
link
in
the
agenda
and
the
meeting
notes.
A
So
if
you
happen
to
be
at
the
at
the
cube
con
in
San,
Diego
in
November
feel
free
to
join.
We
will
be
covering
multiple
topics
depending
on
what
people
will
want
to
what
people
will
be
interested
in,
so
it
will
be
a
light
session.
It
can
be
life
coding,
life
code
reviews
and
literally
the
contents
of
that
session
will
depend
on
the
attendees
of
the
session.
A
A
B
Mo
J
yeah,
actually
as
much
a
mention
I
think
we've
mentioned
a
couple
times
earlier,
we're
very
close
to
being
completely
independent
of
kubernetes
core.
There
are
three
commands
in
particular:
sub
commands:
coop
cuddle,
sub
commands
in
particular
that
have
dependencies
on
kubernetes
core
those
are
coop
cuddle,
get
coop,
cuddle,
auth,
reconcile
and
coop
cuddle
convert,
and
so
I
have
updates
on
all
three
of
those.
So
on
get
there
is
word
to
PRS
away
from
removing
this
dependency.
I
have
a
PR
out.
B
Now
it's
been
out
for
a
little
while
and
Jordan
Liggett
is
aware
of
it,
and,
and
hopefully
we
could-
he
could
get
to
that
relatively
soon.
It
basically
rewrites
a
bunch
of
tests
to
to
allow
us
to
pull
the
table
printing
out
of
package
printers.
That
was
probably
way
too
many
details,
but
basically
were
to
pr's
away
from
getting
that
dependency
out.
B
So
so
so
David
heeds
actually
his
preferred
path
for
the
coop
cuttle
convert
functionality
is
to
put
it
into
the
API
server
and
so
I'm
actually
creating
an
issue
right
now,
so
that
Jordan,
David
and
I
can
resolve
exactly
what
which
one
of
four
different
options
now
for
how
to
wear
the
coop
cut
will
convert.
Functionality
goes
once
it's
moved
out
of
coop
got
old,
so
it
could
be
a
plugin.
It
could
be
a
standalone
binary,
it
could
go
to
the
API
server
or
we
could
just
drop
it
all
together.
B
B
Hurdles
to
overcome
it
sounds
like
now
for
coop
cuddle
convert.
Although
we're
we,
we
announced
deprecation
in
113,
so
we
are
technically
allowed
to
pull
coop
cuddle
convert
out
right
now.
It's
just
that.
If
we
do
without
putting
that
functionality
somewhere
else,
we
may
be
letting
some
of
the
the
community
down.
B
So
so
anyway,
if
you,
if
you
have
any
ideas
on
this,
if
you
think
we
should
just
drop
it,
let
it
let
us
know
again:
I'm
gonna
send
to
the
entire
six
CLI
the
this
issue,
so
that
we
can
discuss
and
hopefully
quickly,
come
to
a
consensus
on
if
and
where
that
functionality
will
end
up
once
it's
removed
from
coop
cuddle
convert.
So
the
final
earth.
B
A
Where
the
command
as
a
plugin
or
standalone,
it
doesn't
matter
honestly
because
both
will
work
the
same
way,
whether
it
will
live
in
the
main
kubernetes
repo,
whether
an
API
machinery
or
somewhere
else
honestly,
should
not
matter
that
much
okay,
it's
just
that.
We
will
not
include
it
in
keep
CTO
by
default.
You'll
have.
B
A
Have
you'll
have
to
do
that
additional
step
to
be
able
to
to
have
that
functionality
and
we'll
just
have
to
figure
out
and
that's
doable
within
117?
What
is
the
the
final
location,
not
sure
about
API
machinery,
though
I
was
under
the
impression
that
API
machinery
does
not
have
the
internal
types
that
are
required
for
convert
to
work,
probably
but
I
find
a
person.
Sure
I
think.
B
A
B
Okay,
all
uh-oh,
so
this
just
happened.
That's
why
I
haven't
been
in
contact
with
you
ma
che
apologize
for
for
dropping
this
on
you
right
here,
I
figured
we
were
meeting
it
might
as
well
bring
it
up
here
instead
of
anyway
yeah.
So
there's
so
there's
that
and
then
there's
the
auth
reconcile,
which
Jordan
has
has
kind
of
taken
the
lead
on,
but
but
he
just
he's
recently
mentioned
that
it's
it's
been
trickier
than
he's
imagined
and
that
he's
not
even
sure
that
he's
going
to
get
it
completed
by
this
release.
B
This
cleanup
that
will
remove
the
dependency
so
yeah.
So
that's
unfortunate,
but
yeah
I'm
not
I'm,
not
sure
how
we
could
it's
not
clear
that
that's
going
to
be
like
the
top
priority
for
for
Jordan
in
this
release
and
I'm,
not
sure
who
else
would
have
the
chops
to
to
to
approach
it
lemon.
Let
me
bring
that
up
again
with
Jordan
to
see
if
we
can
prioritize
it
for
117,
but
yeah
that.
B
A
A
B
That's
that's
definitely.
Our
plan
is
to
you
know
lease
by
the
end
of
this
release.
Have
it
all
out
yeah
but
I?
Think
as
I
mentioned,
I
think
that
there's
a
risk
there,
but
I'll
definitely
I'll.
Take
the
action
item
2
to
loop,
you
and
Sally
in
with
Jordan
to
make
sure
that
we
we
come
to
a
conclusion
on
at
least
we're
going
to
approach
it
yeah.
A
Yeah
yeah
definitely
Sally
and
myself.
We
can't
half
the
time
needed
to
to
make
it
happen
because
I
think
it
is.
It
is
a
time
so
that
we
can
have
the
keep
see
deal
we'll
have
plenty
of
additional
work
after
that
happens,
to
do
clean,
because
the
fact
that
when
we
will
be
in
a
staging
doesn't
mean
that
we
have
all
the
e
to
e
and
the
entire
interest
testing
infrastructure
and
the
release
infrastructure
figure
it
out,
and
that
will
take
some
additional
time.
A
A
Okay
hearing,
none
to
add
it,
I
added
the
second
I-10.
So
about
a
year
ago,
when
we
were
moving
customized
over
to
CLI
Ron
time,
I
had
a
discussion
with
Phil
I
do
hope.
He
will
show
up
here
shortly
that
we
will
move
the
core
of
customized
into
CLI
runtime,
but
leave
all
the
mechanics
of
customize
as
part
of
the
customized
project.
A
If
we
won't
be
involved
like
that,
includes
myself
and
Sally,
and
it
keeps
with
your
independence,
we'll
try
to
cut
the
core
of
customize
into
CLI
runtime,
so
that
people
who
are
currently
depending
on
Scylla
runtime
then
have
to
vendor
the
entire
customized.
At
the
same
time,
we
would
like
to
have
the
core
that
the
engine
of
customized
being
part
of
CLI
runtime,
like
it
was
discussed
several
times
during
our
sig
meetings
in
the
past
and
then
have
the
additional
bits
as
part
of
customized
as
it
is
these
days
and.
A
Still
are
on
time
being
only
the
core
elements,
because
currently,
if
you
work
with
you
at
runtime,
the
only
functionality
that
that
we
are
reusing
from
customeyes
is
the
the
core
element
that
allows
you
to
process.
Customized
files
and
I
would
like
to
have
that
piece
of
code
live
inside
of
this
year.
Light
runtime,
not
sure
how
much
we
will
be
able
to
go
with
the
plan
within
the
117
time
frames,
but
I
would
like
to
start
start
that
process
and
the
rest
are
drinks
in
that
one
aren't
there.
A
With
regards
to
cap,
I
think
it
would
be
worth
highlighting
the
main
goals.
It
is
not
strictly
required,
but
I
think
as
a
good
community
member.
We
should
create
a
cap
describing
the
goals
and
the
desired
outcomes
for
this
process.
So
I'll
sync
with
your
salary
and
we'll
try
to
draft
the
cap.
Okay,.
E
E
Currently,
Java
is
working
on
refactoring
customized
so
since
currently
because
of
the
dependency
circular
dependency.
Previously
we
had
so
we
couldn't
remember
customized
again,
so
Java
is
working
on
refactoring
that
so
that
we
have
multiple
modules
and
which
can
be
perfect
there
for
consumers
to
use.
E
So
with
that
change,
think
it'd
be
easier
to
just
lender
that
so-so
to
us,
so
they
say
customized
with
its
core
functionality
and
it's
called
transformers
are
plugins
in
the
same
project
is
it
can
do
it
has
larger
velocity
and
we
can
do
other
testing
and
read
all
the
dogs
and
they're
the
same
project.
I.
A
A
I
would
like
to
revert
that
dependency
so
that
we
only
rely
on
CL
at
runtime
and
customized
can
also
rely
on
CL
runtime
and
all
the
other
projects
can
have
a
lighter
dependency
tree
for
when,
when
consuming
the
co
at
runtime,
but
I'll
synch
with
Jeff,
that's
a
good
pointer,
I'll,
synch
with
Jeff.
What
his
plan
is
on
that
one
and
we
will
try
to
come
up
with
a
reasonable
plan
how
to
best
about
this.
One.
A
F
Hi
this
is
Himanshu,
so
I
want
to
discuss
a
PR
that
I
created
recently.
The
PR
is
about
default,
sorting
of
QC,
I'm,
sorry
cube,
City
I'll
get
events,
and
the
ask
is
to
when
there
is
no
sort
by
flaga
supply,
then
it
should
be
defaulted
to
last
time.
Staff
so
has
a
little
bit
idea
about
it.
I
discussed
with
him.
So
if
you
guys
wanna
discuss
I
can
give
you
the
PR
number
or
yeah.
C
So
I
think
the
issue
here
is
that,
like
when
you
do
to
control,
get
for
a
resource
like
it's
the
same
command
for
all
the
resources
and
most
resources,
we
don't
want
to
look
by
creation
timestamp,
but
events
is
unique
in
that
we
most
users
probably
typically
want
to
the
events
sort
of
by
timestamp,
rather
than
know
something
else,
and
so
the
proposal
is
to.
If
we
see
that
users
are
getting
events
to
control
people,
he
fought
the
sorting
order
to
timestamp
for
that
resource.
Only.
C
A
C
The
sort
order
the
default
sort
order,
yeah
I,
assume,
if
you
and
I
think
probably
our
stance
should
be
that
if
you
you
are,
depending
on
the
particular
sorting
order,
do
some
scripting
to
like
get
all
the
events
or
something
that
you
should
specify
the
sort
order
that
you
want.
Rather
than
just
relying
on
that.
C
C
B
B
B
B
A
A
Most
of
the
time,
I'm
interested,
in
particular
events,
either
from
a
particular
namespace
or
related
to
a
particular
object
or
something
else
and
get
is
definitely
not
gonna.
Allow
us
to
do
to
do
it
yet
should
be
generic
as
much
as
possible
and
I
would
just
leave
it
as
is,
and
I'm
inclined
to
implement
something
something
less
generic
specifically
targeted
at
akattak.
If
I
remember
correctly,
I
would
have
to
grab
through
our
agenda
notes,
I
like
them.
C
A
A
C
Yeah
I
agree
with
that,
like
you
prop
the
workflow
for
listing
deployments
or
pods
of
these
sorts
of
things
is
different
than
the
workflow
for
listing
events,
because
with
pods
and
deployments
you
like
know
either
the
name
or
the
name
space
or
though
labels
right,
there's,
you
know
which
ones
you're
interested
in
and
so
you're
requesting
those,
whereas
with
events
they're
actually
generated
based
on
other
objects.
And
so
you
don't
know,
and
instead.
A
A
C
But
there's
like
another
thing,
I've
seen
is
optional
components
that
get
installed
as
part
of
an
app
like
the
app
CRD
is
an
example
of
this.
Where,
when
you
package
up
your
workload,
you
may
want
the
apps,
you
may
want
an
absolute
heat
graded
or
an
app
created
for
it
if
it
exists,
but
then
you
don't
want
applying
that
to
fail
and
the
user
doesn't
have
absolutely
installed
right
so
being
able
to
mark
things
as
don't
fail.
C
C
I
think
there's
a
collection
of
directives
we
need
to
give
to
apply
like
one
is:
skip,
apply
completely
a
customization.
What
would
fall
into
this
case?
Another
is
ignore.
Don't
try
to
apply
this
if
the
resource
API,
if
the
API
doesn't
exist
on
this
particular
cluster,
it's
optional
right.
The
applications
here
D
would
fall
into
that
context.
C
Another
would
be
like
don't
delete
this,
even
if
I
tell
you
to
delete
it
because
I
may
have
screwed
up
and
that
would
be
like
the
staple
set
you
may
want
to
annotate
like
certain
staple
sets.
You
don't
delete
your
persistent
storage
like
I'm,
not
sure
what
all
of
them
are
I.
Think
that's
where
the
work
needs
to
be
done
to
figure
out
the
types
of
things
we
want
and
we
probably
don't
need
to
define
all
of
them
up
front.
C
But
we
probably
do
when
to
from
a
high
level
spell
out
the
sort
of
categories
of
things
we
want
to
do
and
then
pick
out
one
or
two
like
the
customization
yamo.
One
I
think
is
going
to
be
important
if
we
are
like
for
customized
generators
and
transformers,
and
these
sorts
of
things
may
want
to
be
hidden.
C
The
I
could
I,
don't
think
trait
is
critical,
I
create
and
recreate
or
thing
is.
What's
called.
We
really
discourage
folks
from
using
those
anyway.
So
if
it
happens
to
work
on
them,
great
I
think
if
we
have
an
easy
way
of
making
it
work
for
all
the
appropriate
plans
great.
But
if
that's
going
to
increase
the
workload
yeah
I,
don't
think
it's
certainly.
A
A
Pretty
sure
that,
as
you
start
writing
this
down
a
couple,
things
might
get
clarified
and,
as
the
rest
of
us
will
be
reading
through,
we'll
get
some
other
ideas
and
we
can
shuffle
through
those
and
figure
out
the
end
goal
but
I
overall,
the
the
idea
does
make
sense.
I
know
that
I
have,
for
example,
somewhere
at
the
back
of
my
head.
A
One
example
is
within
the
entire
cube:
CTL
I
would
like
to
be
able
to
ignore
printing
the
last
suppliant
annotation
that
is
put
together
on
every
single
resource
when
you're
using
keep
CTL
apply
you
guys
in
the
majority
of
cases,
this
will
end
up
printing
the
same
object
twice
and
in
99%
of
cases
you
didn't
care
about
that
particular
annotation.
When
looking
at
the
object,
yeah.
A
G
Little
thing
just
tell
me:
this
is
Nick
here
on
the
topic
of
events,
I'm
hoping
I'm,
just
missing
something
like
this
here,
but
we
thought
recently.
So
the
default
table
structure
you
get
from
gate.
Events
does
not
include
any
identifying
elements
for
the
rows
and
so,
for
example,
the
event
ID
is
not
there.
G
You
can,
but
then
it's
difficult
to
if
I
want
the
default
column
plus
an
ID
I,
think
that
is
difficult
to
express
because
there's
no
way
to
say
I
have
these
add
one
extra
column
to
the
default
set.
So
just
a
it's
a
relatively
minor
point,
but
it's
something
that,
if
we're
considering
small
changes,
it
might
be
helpful.
That
would
be
something
of
that
nature
so
that
consumers
can
know
in
with
some
easy
and
he's
some
easy
fashion
from
the
defaults
or
extended
defaults
having
some
identifying
elements
in
their
rows.
So.
A
G
C
Probably
worth
that's
an
interesting
point:
I,
don't
think
that's
something
we
thought
about.
It
might
be
worth
broadening
that
discussion
to
what
should
the
output
look
like
in
order
to
be
processed
by
other
tools
right
because
I
imagine
there's
like
I.
Imagine
you're
doing
like
parsing
on
your
end
right
that
maybe
you
wouldn't
need
to
do
if
it
was
just
formatted
differently
right
or
like
I,
don't
know,
there's
things
that
I
don't
know,
but
maybe
just
starting
from
like
we
could
have
a
new
like
we
have
wide.
We
have
custom
things.
We
have
different
outputs.
C
G
B
B
B
G
A
Any
objections
to
extend
what
I
remember
Clayton,
what's
troubling
resurface,
our
printing,
we
just
reused
the
same
columns
that
we
are
printing
these
day
as
we
were
printing
back,
then
nothing
has
changed
in
between,
but
I
don't
see
any
objections
for
us
to
adding
more
columns
if
it's
more
appropriate
and
the
idea
of
the
surface
area
was
to
expose
the
printing
mechanism
to
a
broader
audience
of
customers.
So
if
there
is
a
strong
value
in
it,
I
personally
don't
see
any
objections
that
I
could
like
push
the
the
pr4
to
get
it
through.
B
E
Thank
you
so
for
customized
we
in
the
past
a
week
we
release
a
new
version,
3.2
dollar.
So
the
link
points
to
the
release
note
and
in
this
release
we
have
two
new
features.
One
is
the
inline
patch
which
supports
you
can
put
the
whole
patch
as
a
string
in
the
customization
dot
yeah
mobile
instead
of
putting
that
patch
into
a
separate
file.
So
the
other
feature
is
a
new
sub
command
in
customized.
Now,
under
a
directory
you
can
run
customize
create.
E
Then
it
will
create
a
customized
in
yellow
file
with
all
the
resources
automatically
added.
So
it
will
scan
your
current
directory
either
only
on
the
top
level
or
recursively
and
pick
up
all
the
resources
fell
and
list
out
into
in
customization
removal.
So
that's
for
the
new
release
and
another
thing
is
the
beauty
in
plug-in
dock.
E
H
Really,
just
just
to
maybe
rephrase
the
plugins,
the
ability
to
configure
built-in
plugins
is
really
a
way
to
make
customizes
native
behavior
behave,
however,
you'd
like
so
it's
a
mechanism
to
can
basically
to
customize
customize.
So
the
dad
provides
examples
of
how
to
do
this,
and
it
gives
you
completely
fine
detailed
control
over
what
gets
customized
and
and
how
that's
all.
Thank.
E
A
Okay,
thanks
does
anyone
else
and
I'm
looking
at
your
Nick
I
want
to
provide
some
updates.
G
A
B
One
sorry
to
interrupt
I
just
wanted
to
give
a
quick
shout
out
to
one
of
our
members,
Richard
Marshall,
who
I
think
is
still
on
the
call
he's
been
helping
resolve
some
of
the
coop
cuddle
issues
and
has
been
really
helpful
and
constructive
in
in
his
comments
in
helping
to
answer
and
to
resolve
some
of
these
issues.
So
if
you're
out
there
I
think
you're
still
are
I
really
appreciate
your
efforts.