►
From YouTube: Kubernetes SIG CLI 20200408
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
Good
morning,
good
evening,
good
afternoon,
depending
on
where
you
are
today,
is
April
April
8th-
and
this
is
another
of
our
if
I
would
be
sick,
CLI
meetings.
My
name
is
ma
che
and
I'll
be
your
host
today,
so
I
don't
have
any
particular
announcements.
I
was
looking
for
119
dates
a
minute
ago
and
I
have
not
seen
I
have
not
seen
any
dates
published
yet
so
I'm
guessing
that
they
should
show
up,
probably
sometime
next
week
or
so
then
we
will
be
able
to
to
talk
a
little
bit
more
about
it
either
way.
A
B
Was
a
couple
of
issues
with
the
118
0qu
cuddle
and
we
they
just
cut
the
118
one,
which
has
some
of
the
those
fixes,
and
so
we'll
communicate
more
broadly
to
the
rest
of
the
community
kubernetes
community.
But
we
should
all
know
that
the
that
first
version
of
118
zero
had
some
issues
with
with
diff
and
with
apply,
and
so
those
should.
A
A
Okay,
with
that,
out
of
the
way,
I
think
we
can
jump
on
to
the
119
planning
from
the
60
alive,
leading
team,
the
1
and
most
important
bit
that
we
will
continue
our
journey
with.
Is
the
move
of
Cuba
caudal
to
this
aging
repository
as
it
as
it
stay
as
it
is
today
we
are
still
two
items
left
on
that
list,
and
the
biggest
offender
on
that
list
is
African
style.
A
A
A
We
put
together
a
proposal
about
using
femoral
containers
for
debugging
application
with
cue
cattle
debug,
and
he
was
very
diligent
to
start
cranking
the
code
behind
the
cubicle
debug
command
and
he's
already
pinging
me
about
follow-ups
that
will
be
that
will
be
added
to
the
119
so
he's
currently
working
on
those
bits.
The
cubicle
events
is
another
Alpha
command
that
we
will
be
adding.
Unfortunately,
we
didn't
make
it
for
118.
A
The
next
topic
is
a
follow-up
to
the
server-side
print,
as
you
might
be
aware
of,
when
you
are
when
you
are
cube,
cuddle
getting
resources
from
the
server,
it
is
the
server
that
dictates
the
get
command
how
these
resources
should
be
printed.
The
reason
for
that
is
one
we
want
to
have
a
consistent
way
of
printing
resources
across
all
des
all.
A
The
clients,
including
cube
huddle,
as
well
as
other
clients
that
can
be
written
in
different
languages,
programming
languages,
as
well
as
the
ability
to
for
the
server
to
tell
the
clients
how
to
print
resources,
also
allows
clients
to
understand
the
CR
DS
which
are
registered
on
the
server
without
having
extra
types
being
baked.
In
on
the
servers
on
the
client
side,
we
would
like
to
continue
the
same
steps,
but
this
time
with
the
with
the
Q
cuttle
describe
commit.
A
A
On
that
one,
and
at
least
put
a
proposal
for
this
one
and
maybe
start
some
initial,
some
initial
implementation
in
the
next
topic
and
then
I
promise
I'll
shut
up
is
a
common
sight,
command.
Metrics
Philip
put
a
proposal
about
half
a
year
ago
or
so
about
the
ability
to
send
over
to
the
server
the
information
about
the
comments
that
are
being
executed
by
the
users,
as
well
as
the
flags
and
additional
metadata
that
we
can.
Then
we
you
log
on
the
server
and
try
to
gather
some
statistics
around
those
data.
A
Those
statistics
should
allow,
in
the
long
run,
to
help
us
gather
information
about
how
cube
huddle
is
being
used
or
what
kind
of
functionality
is
the
most
useful
for
user
or
where
they're
having
some
issues
with
I
think
these
were
the
main
topics
that
I
wanted
to
put
brought
to
your
attention.
I
see
that
John
Ryan
put
together
one
another
I
want
to
talk
about
it
a
little
bit.
C
Yeah,
the
the
summary
of
this
is
that
in
in
the
describe
command,
if
there
are,
is
there
sensitive
data
in
an
annotation,
we
display
that
it's
normally
not
a
problem
in
when
you,
when
you
run
the
apply
command,
there's
an
annotation,
that's
added
that
shows
a
copy
of
the
object
and
we
currently
skip
showing
that
annotation,
specifically
there's
other
tooling.
That
wants
to
do
the
same
trick
to
like
capture
the
state
of
the
object
on
an
annotation
absent
any
other
location.
C
A
A
A
We
are
considering
hiding
the
entire
metadata
that
is
server-side
related
because
currently,
if
you
keep
cattle,
apply
a
resource
and
then
you
open
that
that
same
resource
with
Q
cattle
added,
not
only
you,
will
see
your
actual
resource,
but
the
same
amount
of
data.
You
will
see
as
the
server
side
related
applied
data
that
will
be
confusing
to
a
lot
of
the
users.
A
It
is
basically
a
short
description
of
the
of
that
field,
and
we
are
thinking
that
maybe
placing
that
data
inside
of
the
cube
kernel
edit
resource
will
be
valuable
for
a
lot
of
users
when
they
are
editing
those
we
might
probably
start
with
hiding
those
behind
a
flag
before
and
will
enable
those
by
default.
But
that's
a
that's
a
future
goal
that
all
all
probably
should
have
to
do
the
list
as
well.
So.
B
B
C
D
That
secrets
have
data
that
we
typically
don't
want
to
print
to
the
console
like
if
you're
logging
it
from
a
CI
CD
system
or
something
like
that,
and
you
don't
want
that
annotation
printed
as
part
of
describe
and
maybe
would
help
to
have
a
little
bit
of
background
information.
What
do
you
are
you
doing?
C
C
It's
it's
it's
apply
plus
because
it
because,
through
through
this
additional
metadata,
a
a
set
of
resources
can
be
treated
as
a
group
and
operations
going
to
be
applied
against
it.
There's
semantics
in
this
tool
in
encap,
kaa
PP,
there's
semantics
in
the
tool
for
suggesting
or
requiring
ordering
of
resources,
cap
weights
for
customer
resource
definitions
to
be
applied
and
be
available
along
other
resource.
Where
there
are
some
dependencies,
you
can
name
some
of
that.
It
helps
out
with
merging
data.
C
D
With
server-side
apply,
do
you
see
so
specifically
with
them?
We're
like
a
lot
of
that
stuff
sounds
like
it's
not
impacted
by
this
issue.
It
really
seems
to
be
the
like.
Is
it
correct
that
the
this
issue
really
only
impacts
the
the
updating,
the
object
room
up
like
detecting
gifts
in
merging
in
local
configuration
to
the
remote
object
effectively
what
apply
does,
but
maybe,
if
that's
the
case,
then
do
you
still
like
the
server
side,
apply
impact
anything
about
how
you're
doing
cuz
the
way
server
side
apply
is
working?
D
D
C
That's
a
great
question:
I'm
I'm
personally,
just
becoming
aware
of
the
server
side.
D
Because
so,
like
my
high-level
thoughts
and
I
think
I
left
a
couple
of
these
on
the
issue
are
the
this
like?
It
definitely
make
sense
what
you're
asking
for
doing
it?
The
way
we
did
it
for
secrets,
is
it
hack
and
probably
one
I
wish
we
didn't
have.
This
is
going
to
impact
everyone
who
currently
describes
secrets
like
how
they
view
it
right
in
so
that's
going
to
disappear
for
them
from
an
API
perspective,
we
don't
we
don't
have
backwards
compatibility
constraints
on
the
output
of
describe.
D
So,
if
you
are
like
using
describe
to
parse
data
out,
then
we
may
break
you.
On
the
other
hand,
people
do
it
anyway,
we
see
them,
do
it
quite
a
bit,
we
occasionally
break
them
and
then
it's
a
bad
experience
for
them.
So,
even
though
we
don't
have
a
strict
guarantee,
we
promise
not
to
break
this.
I
would
say
it's
not
like
when
we
update
that
stuff.
D
We
try
to
do
it
in
a
way
that
isn't
going
to
break
people
that
are
our
person
data
out
of
there
and
then
the
other
thing
I
think
I'd
say
is
what
you're
doing
since
is
really
similar.
What
like
I
get
the
point
where
we
special
case
the
apply
annotation,
because
you
know
two
reasons
one
is
we
control
it?
We
know
it's
going
to
be
there.
D
D
A
D
B
A
A
It's
pretty
much
similar
to
what
we
did
with
export
and
a
lot
of
people
were
complaining
that
we
are
deprecating
and
removing
the
expert
functionality
and
the
problem
with
that
was
that
each
person's
idea
of
exporting
resources
from
the
server
was
different.
One
was
using
export
as
a
backup.
Others
were
using
sport
as
a
well
I
want
to
get
those
resources
but
stripped
metadata
or
just
strip
namespace
or
something
else,
and
and
we
figure
it
out
that
there
are
many
possible
use
cases
how
users
can
or
want
to
consume.
A
D
A
F
F
B
A
D
D
D
The
next
step
would
be
server-side
describe,
which
is
what
we're
working
on,
which
is
dynamic,
but
per
type
still
right,
so
the
client
doesn't
know
how
to
display
anything.
The
server
does,
but
the
server
just
knows
how
to
display.
This
is
what
a
secret
looks
like
right,
which
is
better
but
like
I
think
would
be
really
interesting
to
me
is
if
you
go
one
step
further
and
server
can
say
not
just.
D
Maybe
this
this
the
way
this
beam,
salt
and
I
think
John,
maybe
I
guess
what
I'd
suggest
is
like
as
an
immediately
maybe
check
out
the
apply
pieces.
The
server-side
apply
and
consider
like
is,
is
that
a
route
that
maybe
server-side
apply
may
alleviate
the
need
for
this.
In
the
long
term-
and
so
maybe
we
don't
need
this
feature
right
now
and
then
and
then,
if
we
really
think
that
we
need
this
think
about
the
thing
about
the
problem
in
the
context
I
just
kind
of
described,
yeah.
C
That's
really
helpful.
Philip
both
clearly
got
some
homework
here
to
to
do
that,
and
then
understanding
kind
of
where
your
head's
at
in
terms
of
the
life
of
this
tool
chain
and
in
seeing
that
it's
at
this
point
in
its
maturation,
we're
providing
first-class
extensions
for
things
like
this
is,
is
where
you're
at
not
where
you
were
that's
really
helpful.
Yeah.
Thank
you
appreciate
that
all.
D
E
A
D
B
Just
just
the
those
two
small
issues
that
we
discussed,
one
about
the
version,
skew
warning
which
might
be
you
know
it's
gonna,
be
have
a
high
impact,
but
it
would
be
a
small
effort
to
get
that
out.
It's
already
an
issue
for
for
putting
a
version,
a
warning
out
into
cuddle.
Whatever
the
version
skew
is
out
of.
B
Client
server
and
then
there
was
the
there's,
probably
going
to
be
a
few
issues
that
some
of
the
some
of
our
members
here
would
want
to
address,
removing
the
Cobra
from
the
options.
Those
are
also
going
to
be
I
think
pretty
straightforward
PRS
for
people
who
want
to
get
some
of
those
we
want
to
contribute.
A
Actually,
we
already
have
an
example
of
the
separation
between
Cobra
and
options.
If
I
remember
correctly,
I
was
when
we
were
talking
about
it
last
time.
I
was
also
checking
because
I
remember
David,
put
together
the
the
wait
grant,
which
is
still
alpha.
The
way
logic
was
used
is
used
within
the
delete
command
so
that
we
could
maintain
the
old
behavior
of
waiting
until
the
resource
is
being
removed.
A
A
Maybe
it
will
be
worthy
and
we
should
probably
create
an
umbrella
issue
for
switching
to
the
to
the
flow
where
we
have
the
option
structure
and
those
three
methods
or
well.
The
three
is,
it
will
always
depend
on
the
on
the
command,
but
at
least
to
where
there
is
a
complete,
the
valley
which
is
optional
most
most
frequently
and
then
there's
the
actual
run
command.
A
Eddie
can
I
ask
you
to
create
two
umbrella
issues
under
cube
Caudill
for
these
two
or
you
will
create
the
one
about
the
new
structure
and
Shaun
could
create
the
one
about
the
separation
with
pointers
or
two
specific
commands.
How
this
should
look
like,
and
then
we
can
point
people
or
eventually
we
can
have
a
checklist
of
what's
still
left.
G
B
A
A
Okay,
so
I'll
I'll
ask
the
other
projects
Kukui
curved
and
customized,
about
their
plans
for
119,
although
I'm
guessing
that
you're
not
strictly
tied
to
two
they're.
Really
these
cycles
and
you're
always
working
tirelessly
to
you,
know
to
release
your
bids
pretty
much
continuously.
I
know
the
crew
is
doing
so
and
kuya
the
same
and
customized
pretty
much
as
well
yeah.
Very
much
so
that's
the
benefit
of
being
outside
of
the
main
repo.
Probably
that's
where
we're
targeting
with
cube
cuddle
as
well,
and
then
we
run
but
I
think
it
would
be.
D
A
D
D
The
second
thing
it's
going
to
get
us
is
that
we've
been
developing
different
libraries
built
on
top
of
this,
these
yellow
native
libraries
that
allow
us
to
do
some
pretty
interesting
things
with
respect
to
schema
and
embedding
schema
/
within
resources
themselves
and
doing
per
object,
schema,
and
so,
for
instance,
you
should
be
able
to
specify
on
a
particular
field
use
this
merge
key
for
this
field,
which
will
allow
you
to
one
override
things
where
the
merge
key
is
probably
not
the
merge
key.
What
port
is
an
example
of
this?
The
port.
D
H
That's
that's
that
camera.
That's
actually
not
really
on
it
here,
so
small
update
of
our
crew.
We
have
refreshed
our
adopt
site.
We
have
pretty
good
organized
documentation
now,
although
most
things
users
need
to
know
is
in
one
page
called
Quick
Start,
which
is
pretty
short,
and
the
rest
is
mostly
about
how
developers
should
distribute
plugins
and
best
practices
around
that.
We
actually
took
the
time
to
also
to
sit
down
and
write,
how
you
should
approach
plug-in
development
and
like
what
you
should
reuse
and
why
shouldwe
you
should
pay
attention
to.
H
So
that's
it's
a
pretty
good
guide
to
code
of
plugins
in
general,
so
the
next
big
work
that
is
happening
right
now
is
Kris
Kim,
who
is
also
in
the
call
right
now,
is
working
on
supporting
a
multiple
plug
in
the
Nexus
I
added
a
link
to
the
umbrella
issue.
If
we
finally
created
writing
out
we're
pretty
good,
we
have
a
pretty
solid
prototype
that
working.
This
is
gonna.
Allow
people
to
distribute
plugins
that
we
have.
You
know
previously
rejected
from
the
centralized
index,
because
we
do
that.
Actually
quite
a
lot.
H
People
come
up
to
plugins,
that
are,
they
thought
they
don't
have
a
particular
focus,
or
they
are
just
doing
something
that
another
plug-in
is
already
doing.
So
you
know
we're
typically
rejecting
those
plugins,
but
mainly
right
now,
we're
enabling
people
to
you
know
destroy
plugins
if
their
cadence
we're
also
trying
to
add
tooling
so
that
other
play
gonna.
You
know,
index
owners
can
actually
validate
plugins
as
well.
Still
that's
crappie.
I
So
you
have
maintained
on
the
call
as
well
here
she's
on
my
team.
Just
want
to
give
another
shout
out:
we've
been
working
out,
yes,
but
near
future
I
think
we're
going
to
continue
to
work
with
disinterring,
but
I
summarized
in
the
in
the
the
docs.
So
we've
recently
added
a
lot
of
updates
to
our
read
knees
and
documentation
and
maintain
his
hard
work
to
finish
that
work
up.
I
I
know
that
Eddie
you
volunteered
last
week
to
help
out
a
bit
writing
some
support
and
I
sort
of
put
back
I
delayed
that
a
bit
toka
week.
You
can
fish
docks
once
we
have
the
docks
in
here
for
help
developers
on
board,
then
we
can
open,
open,
I
think
the
case.
A
little
more
I
rely
with
a
little
more
confidence,
so
I
think
we're
pretty
close
to
that
she's
maintained
great
job
on
the
documentation.
I
We
also
recently
added
some
support:
initial
support
for
augmenting
and
visualizing,
some
of
the
crew
aspects
of
crew
like
crew
list
and
and
you
can
get
a
nice
sidecar,
showing
a
summary
of
the
plugins
on
the
status
of
the
plugins
and
so
on,
and
also
for
customized.
We
have
some
additional
support
there
as
well.
So
if
we
say
customize,
you
know
on
the
directory,
you'll
get
back
a
I,
think
a
pretty
nice
summary.
It's
a
start,
at
least
for
summarizing
the
output
of
the
customize
command.
I
In
addition,
when
you
do
a
create
or
apply
Kay
you'll
now
get
the
full
GUI
status
table.
So
if
you
don't
apply
Kay
on
directory,
you'll
get
a
nice
summary
of
everything.
It's
it's,
not
as
everything
that's
being
deployed.
So
the
hope
is
that
now
we
have
a
firm
foundation.
We
can
really
start
you
bringing
people
on
board
to
fill
in
these
kinds
of
gaps
for
customizing
crew
and
other
kinds
of
new
features.
I
I
So
they
we
don't
really
have
a
really
great
spi
for
the
multi-tenancy
story
and
that's
part
of
the
part
of
the
work
we
need
to
round
this
out.
For
example,
one
of
our
users
has
a
very
kind
of
a
teasing
kradic
multi-tenancy
model,
other
Azipod
base
multi-tenancy
model,
and
so
we
see
to
work
out
some
of
the
details
there.
How
we
can
extract
that
out,
but
the
idea
is
that
we
can
do
it.
We
can
solve
it
in
a
couple
of
different
ways.
I
A
Okay,
so,
okay,
we're
are
slowly
closing
our
meeting.
There's
anyone
I
wanted
to
I,
don't
know
sure
some
note
some
ideas,
I
remember
that
Trinidad
is
a
new
one.
He
mentioned
shortly
that
at
the
beginning,
if
you
want
to
say
hi
to
everyone
or
you
want
to
you,
know,
you're
looking
for
some
of
the
advise
or
stuff
like
that.
Just
that's
your
time
to
do
it
sure.
J
I'm
Trinidad
I'm,
located
in
San,
Antonio,
Texas
and
I'm,
oh
transitioning,
from
a
retail
company
to
an
insurance
company,
but
with
the
growth
of
containers
and
cloud
technologies,
I
think
it's
something
that
you
know
as
someone
who
applies
a
lot
of
this
stuff
needs
to
get
a
little
bit
deeper
into
and
I
decided
to
kind
of
kind
of
work
around
someone
a
said
group
see
where
I
can
help
and
get
more
familiar
with.
I
was
thinking
about.
A
B
A
F
I'm
working
at
circle
CI
right
now
we're
exploring
customized
a
little
bit
also
just
as
a
lark
I
saw
the
KML
that
is
coming
out
of
the
customize
project
and
I'm,
putting
together
a
linter
based
on
that,
because
trying
to
do
that
before
with
the
unstructured
package
was
a
nightmare
and
I
gave
up
halfway
through.
It
was
awesome.
D
Yeah
I'm,
actually,
if
you're
doing
a
linter,
we're
actually
just
digging
some
some
really
interesting
stuff
there
with
how
do
I
get
specific
line,
numbers
and
columns
and
fields
and
that
sort
of
thing
and
then
have
like
warning
versus
error
versus
info
messages
and
this
sort
of
stuff
and
the
idea
is
that
we'll
be
able
to
actually
develop
really
high
fidelity
linters.
You
can
take
the
output
of
and
and
plug
it
into
your
system.
Yeah.
If
you
want
to,
you,
can
feel
free
to
connect
with
me
and
we
can
about
that.
More
yeah.
F
D
J
Basically,
I'd
like
to
help
out
where
I
can
I
was
thinking
like
I
said
earlier,
joining
the
docs
group
may
be
helping
out
there,
but
you
know
something:
that's
low
level
or
or
something
yeah
cuz
I'm
just
getting
into
the
kubernetes
area,
but
I'm
just
gonna
dive
into
the
pool
and
try
and
help
out
where
I
can
yeah.
G
D
G
J
Hi,
how
do
I,
how
do
I
tap
into
that
that
group
you
just
mentioned
if.