►
From YouTube: Kubernetes SIG CLI 20191106
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
Today
we
have
pretty
short
agenda,
but
I
still
wanted
to
meet
because
we
will
be
skipping.
The
next
meeting
and
I
should
probably
add
that
one
to
the
announcements,
so
the
next
meeting
in
two
weeks
on
November
20,
if
I'm,
counting
correctly,
will
be
cancelled
due
to
Q
Khan
coming
on
I,
do
hope
to
see,
if
not
all,
then
at
least
most
of
you
over
there
I
would
I'm
gonna.
Definitely
encourage
you
to
check
out
the
two
session.
A
Two
official
session
that
we
will
be
holding
one
will
be
during
the
main
conference,
which
is
the
90-minute
session,
where
we
will
be
talking
about
totally
everything
related
to
cube,
CTL
and
6li
overall
we're
hoping
to
cover
a
little
bit
of
the
past
12
to
18
months.
What
we've
achieved
we're
hoping
to
to
shed
some
light
on
the
future
of
the
six
CLI
as
well.
A
So,
if
you
will
be
part
of
the
contributors
summit,
definitely
check
this
one
out
as
well
after
I
put
in
information
about
his
crew
talk.
So
if
you
don't
know,
crew
is
a
plugin
manager
and
we
will
also
cover
plugins
for
sure
during
the
big
session,
but
what
ahmed
is
going
to
present
a
separate
session
just
with
regards
to
crew?
A
So
if
your
drink
cup
cone
make
sure
to
visit
those
those
talks
or
step
by
and
just
say,
hi
or
throw
whatever
you
want
to
throw
at
us
at
the
same
time,
although
I
hope
it
won't
be
rotten
tomatoes
or
old,
eggs
or
stuff
like
that
anyway,
anyhow,
let's
move
on
to
the
main
topics,
the
first
one
is
a
continuation
from
from
the
previous
session
from
two
weeks
ago.
It's
the
keep
CTL
debug,
Lee
and
I
had
a
very
short
conversation
about
the
future.
A
Let's
call
it
all
Turner
our
till
early
to
approach
any
possible
use
case
that
application
developer,
but
as
well
as
the
cluster
administrator's
might
have
when
it
comes
to
debugging
Lee.
Can
you
talk
a
little
bit
about
the
the
idea
that
we
want
to
put
into
into
kept
I?
Do
hope.
It'll
just
happen
over
the
next
couple
of
days
or
in
the
worst
case
weeks,
but
for
sure
118
is,
is
the
goal
for
this
one
yeah.
B
I'd
be
happy
to
so
I
kind
of
started.
Thinking
of
coop,
controlled
debug
as
a
facility
for
things
that
are
that
are
higher
level
than
container
debugging
I
mean
I've,
been
focusing
mainly
on
pod.
There's
two
ways
that
I
can
think
of
that
people
might
want
to
debug
pod
in
one
is
attaching
an
ephemeral
container
to
the
running
pod,
so
just
sort
of
inserting
the
container
in
the
simple
use
case,
and
then
the
other
is.
Maybe
people
want
to
change
some,
some
other
sections
of
the
pots
back
and
restart
it
in
place?
B
Maybe
that's
useful
for
switching
to
a
debugging
image
of
your
binary
or
something
like
that.
I
think,
there's
some
equivalent
functionality
and
in
openshift
that
had
been
mentioned
early
on
a
sort
of
prior
art
and
then
another
one
that
I
hadn't
thought
of
at
all,
but
seems
really
useful,
is
debugging
a
node
so
creating
maybe
a
privileged
container
in
the
host
name
spaces.
B
So
you
can
see
what's
going
on
with
the
node
without
having
to
track
it
down
and
an
SSH
I
wanted
to
ask
if
there
were
any
other
ideas
and
just
sort
of
open
the
open.
The
floor
to
general
discussion
or
or
any
questions
about
the
ephemeral
containers,
a
use
case
that
may
have
arisen
well
well,
I,
wasn't
here
before.
A
This
also
covers
the
case
when
you
are
trying
to
debug
some
kind
of
a
connectivity
issues
that
your
application
is
struggling
with,
or
configuration
issues
that
you
haven't
envisioned,
such
as
how
to
access
a
8
different
servers,
whether
the
DBE
that
you
set
up
under
this.
It's
actually
labeled
through
these
environment
variables
through
these
bits
that
you
hoped
for
because
it
is
sometimes
like
I,
said
valuable
to
have
the
application
with
extra
debugging,
but
also
for
debugging.
The
connectivity
with
other
components
in
the
cluster.
B
What
do
you
envision?
The
infirm
role,
containers
for
being
particularly
useful
for
network
connectivity
debugging,
since
there
they're,
always
in
the
same
network,
namespace
and
I,
did
also
want
to
mention
that
process
namespace
targeting
is
a
feature
that
I
and
soon
in
alpha
I
have
a
PR
out
for
that
as
well.
That
will
allow
you
to
create
an
ephemeral
container
for
pods
that
don't
share
process
namespaces,
create
an
ephemeral
container
in
a
targeted
container
namespace
and
a.
A
Very
important
use
case
for
debugging
note
is
what
we
do.
An
open
ship
for
specifically,
is
that
the
nodes
of
the
cluster
by
default
are
not
reachable
through
an
external
API
IP
address,
so
you
need
to
set
up
a
bash
and
hugs,
usually
if
you
want
to-
and
that
requires
some
additional
steps.
But
if
you
want
to
have
a
quick
access,
OCD
bug
note,
which
was
available
you
know,
see
for
quite
a
bit-
is
a
possible
option
for
for
doing
that
and
yeah.
B
Cool
so
I
think
my
next
steps,
then,
will
be
to
finish
writing
filling
out
the
the
different
sections
of
the
cap.
There's
still
plenty
of
to-do
items
in
there,
but
hopefully
we
can
finish
that
during
during
the
code,
freeze
would
be
convenient,
I'd
be
happy
to
have
any
co-authors.
If
anyone
is
interested
in
working
on
this
more
there's
plenty
of
work
here
to
do
way
more
than
then
I'll
be
able
to
do
in
any
short
period
of
time.
Please
just
ping
the
feature.
A
You've
mentioned
in
the
discussion
that
we
had
earlier
today,
reaching
out
to
Mike
Allen
for
his
name,
but
maybe
reaching
out
to
the
authors
of
the
QC
field.
Plugins
may
be
worth
pinging
them
at
least,
and
the
cap
would
be
beneficial
and
figuring
out
whether
they
are
either
interested
in
and
co-authoring.
The
proposal,
or
maybe
interested
in
in
helping
with
the
implementation
I'll,
definitely
devote
some
time
to
to
review
the
path
and
suggest
options,
not
sure
if
that
I
can
you
can
consider
myself
as
a
goth
or
my
thing.
Yeah
yeah.
B
C
C
A
D
B
My
opinion
is
I'd
love
to
see
a
way
that
we
could
implement.
This
would
be
sort
of
a
suite
of
debugging
container
images
that
could
be
shared
in
some
fashion.
I
mean
I,
guess
it's
sort
of
similar
to
plugins,
but
like
a
common
toolkit
that
could
be
referenced
by
the
debug
command.
But
I
haven't
really
thought
that
through
I.
A
So
currently
the
locks
command
has
an
additional
flag
which
allows
bypassing
the
security
of
the
certificate
check,
and
this
should
allow
you
to
get
the
logs
from
the
running
container,
but
this
is
more
towards
when
the
infrastructure
is
failing
due
to
some
certificate.
Issues
such
as
the
cubelet
is
using
either
outdated
or
or
unverified
certificate.
So
there's
there
are
connectivity
issues
between
the
cube,
API
server
and
the
cubelet
and
getting
logs
and
normal
circuit
circumstances
would
not
be
possible.
With
this.
E
A
This,
in
combination
with
what
we
are
proposing
as
the
bug
I,
think
that
the
bug
at
least
a
coop
ctod
bug
is
more
towards
well
more
towards
I.
Think
the
application
developers
need
a
nurse
that
are
running
their
apps,
although
the
the
debugging
node
part
will
be
towards
the
cluster
administrators
cluster
operators.
B
Something
that
that
could
be
useful,
not
so
much
as
part
of
coop
control
to
bug
but
more
use
of
ephemeral
containers
would
be
to
attach
some
sort
of
inspection
container
image
that
could
connect
to
a
broader
debugging
tool.
But-
and
you
can
imagine
that,
like
you,
you
run
this-
it
targets
a
particular
process
and
extracts
its
log
files
and
displays
it
to
you.
D
There's
a
lot
of
aspects
to
debugging.
There's
the
remote
attached,
there's
log
analysis:
there's
schema
validation,
yeah,
so
there's
a
lot
of
things
you
don't
get
to
throw
at
the
problem
in
order
to
actually
debug
it,
and
so,
as
its
phrase
scoot
build
about,
sounds
like
it's
a
broader
topic,
then
it
might
be
focusing
on
now,
but
I.
A
It
will
be
inspection
tolling,
so
he
will
debug
get
into
that
container
immediately
and
then
he
can,
just
you
know,
do
any
kind
of
introspection
that
their
desire.
Of
course
you
got
it.
It
is
still
possible
using
chip,
CTL,
Ron,
plus
interactive
fine,
but
I
think
we
can
make
some
additional
functionality
into
the
debug
command
itself
makes.
A
C
Sure
so,
as
I
mentioned
before,
we
let
me
just
give
the
info
on
the
contributor
summit
workshop
on
app
the
contributors
summit.
We
have
a
90-minute
workshop
where
we
hope
to
address
or
give
us
some
information
on
coop
cuttle
development,
and
that
will
include
you
know
several
different
sections
that
have
slides
and
code
labs
and
and,
as
most
of
us
probably
already
know,
there's
quite
a
bit
of
common
structures
and
conventions
to
the
pheasants
and
dozens
of
different
coop
kuddle
commands.
C
And
what
we're
trying
to
do
is
we
don't
we're
trying
to
develop
a
a
more
robust,
coop,
cuttle
development
community
and
increase,
try
to
get
some
more
reviewers
and
approvers
and
update
our
reviewers
and
approvers
lists,
maybe
afterwards,
so
that
you
know
we
can
we
we?
We
will
address
the
the
kind
of
that
the
common
cases
for
coop
cuddle
development,
so
that
you
know
we
can.
We
can
actually
have
this
stronger
development,
community
and
and
lower
the
kind
of
address
the
hurdles
for
the
learning
curve
for
coop
cuddle
development.
C
Like
you
know,
runtime
object,
unstructured
the
strongly
typed
objects,
and,
and
so
anyway,
each
each
one
of
those,
those
sections
we're
kind
of
hoping
to
have
a
slick
presentation
and
then
a
code
lab
for
those
four
and
then
we
are
hoping
to
once.
We
kind
of
finished
up
on
this
legacy.
Coop
cuddle
development
area,
kind
of
open
it
up
and
start
digging
into,
or
at
least
addressing
and
bringing
up
kind
of
the
future.
C
Coop
cuddle
development,
including
you
know,
verbs,
debug
efforts,
but
also
kind
of
observability
efforts
to
add
headers
into
the
coop
cuddle
command
so
that
we
can
scrape
them
in
logs
and
find
out.
Actually,
what
you
know
what
the
coop
color
usage
is
so
yeah,
I,
guess
I
kind
of
discussed
many
different
areas
here.
C
What
I'm
going
to
do
is
add
a
link
to
this
to
our
notes.
That
includes
the
outline
and
we
I'd
like
to
get
feedback
from
the
community
on
what
you
think
about
this
effort.
I
think
one
of
the
most
important
things
about
this
contributor
summit
is
that,
since
we're
going
to
be
videotaping
and
recording
it,
anybody
who
has
who
wants
to
get
into
coop
cuddle
development
or
core
six
CLI
development
can
actually
watch
this,
and
hopefully
they
would.
C
You
know
again
be
able
to
bypass
these
the
the
learning
curve
hurdles
and
be
significantly
more
productive,
sir.
So
so
why
don't
I?
Let
me
add
this:
the
outline
link
to
these
notes
and
I'd
love
to
hear
any
of
your
guys
feedback
on
on
what,
on
you
know,
details
or
even
the
high
level.
You
know
what
you
think
should
be
addressed.
What
are
the
most
important
parts
of
our
six
CLI
development
that
you
think
should
be
addressed?
Let
me
let
me
find
that
and
add
that.
C
A
C
So
so
maybe
we
could,
you
know,
drop
off
that
the
last
you
know
unit
testing
part
if
we
want
to
spend
more
time
on
the
future
development
projects.
To
be
honest,
I'm,
not
even
you
know,
I'm
not
even
quite
sure
how
we
were
going
to
how
we
were
gonna
address
those
future
development
projects
or
how
much
time
we
want
to
spend
on
it.
C
Do
what
I'm
not
I'm,
not
whatever
said
sure
how
we
were
going
to
address
the
you
know
what
the
future
development
projects
are
other
than
just
giving
a
brief
overview
of
what
we
think
they
are.
You
know
for
the
first
one
Phil
already
has
a
cap
on
the
observability
and
the
the
X
headers,
and
it
actually
you
know
it'll
we'll,
have
to
probably
work
pretty
closely
with
API
machinery
on
that
one
yeah
yeah.
C
Let
me
know
what
you
guys
think
about
the
balance
of
you
know:
legacy
coop,
cuttle
development
versus
future
coop,
cuddle
development,
and
if
we
don't
have
enough
time,
then
you
know
I'm
I'm
gonna
propose
that
it's
it's
weighted
towards
the
legacy.
Coop
cuddle
development,
just
because
this
is
going
to
be
recorded
and
used
as
our
kind
of
introduction
to
coop
cuddle
and
even
see
CLI
development.
Yeah.
A
C
A
What
is
a
cap
and
stuff
like
that,
so
that
people
that
are
new
have
the
the
initial
jump
start
very
quick.
I
know
that
for
a
lot
of
people,
the
initial
bids.
Oh,
how
do
I
do
this?
How
do
I
that
do
that?
Oh,
what
is
a
cap
that
people
are
talking
about
and
those
you
know
simple
FAQ
for
beginners
would
be
much
more
valuable
than
even
as
a
written
dog
that
we
could
search
during
the
session
or
even
post
session
and
having
both
the
the
recording
and
that
kind
of
FAQ.
A
E
Think
yeah
one
thing
I
suggested
to
Sean
was
that,
if
you're
new
to
being
started,
there's
kind
of
two
or
even,
if
you're
an
experienced
contributor,
but
if
you're
trying
to
contribute
there's
kind
of
two
pieces
you
need.
The
first
is
like
what
should
I
be
working
on
and
the
second
is:
how
do
I
do
it
and
of
the
two?
How
do
I
do
it
with
something
that
is
possible
for
people
to
figure
out
on
their
own,
like
it
might
not
be
fun?
E
A
Definitely
I
just
realized
that
something
that
is
completely
missing
from
this
and
I
would
not
even
say
that
it
requires
a
significant
portion
of
time,
but
it's
so
worthy
is
covering
how
you
can
submit
your
fixes
or
improvements
to
our
keep
CTL
book,
which
is
covering
how
to
use
which
is
basically
documentation
if
there
are
people
that
are
interested
in
documenting
cube
sitio.
That
is
completely
valid
approach
and
that
will
be
supervised
super
valuable
for
this
egg
as
well.
A
A
So
if
you
know
how
to
any
have
the
the
ability
to
ride,
that
is
perfectly
valid
contributions
that
that
we
are
missing
currently
I
know
it
falls
under
the
sig
Docs,
but
I
think
you
keep
CTO
book,
it's
something
that
we
have
significant
ownership
of
so
being
able
to
to
show
people
how
to
contribute
to
that
part
of
the
year.
Our
our
projects
is
also
perfectly
valid.
C
Okay,
so
if
I
here
fill
correctly,
he
he's
voting
more
towards
weighing
it
towards
future
cout
cuddle
development.
So
if
that's
the
case,
then
we
can,
you
know
we
can
reduce
the
scope
of
the
legacy.
Cuddle
development
I,
think
that
you
know
developing
the
slides
and
the
code
lab
for
four
more
areas
rather
than
fewer
would
be
fine.
It's
just
that.
You
know.
Once
we
get
to
a
certain
time
point
in
that
contributor
summit,
we
can
move
on
to
future
coop
cuddle
development.
C
C
A
Okay,
cool
Oh,
unless
there
are
other
questions
or
suggestions
with
regards
to
the
contributor
session
I
think
we
can
move
on
to
the
stand-up
section.
I
see
that
crew
and
cuy
put
their
sent
ups
crew
aqua.
Do
you
want
to
talk
about
I,
see
her
on
the
cam?
Can
you
hear
me
yeah
loud
and
clear?
Okay,
awesome.
A
If
I
can
come
in
on
that,
one
that
seemed
like
themselves
were
their
source
of
our
biggest
issues
with
keep
CTL
copy
command,
we're
for
the
past
year,
or
so
we've
been
involved
in
several
each
and
every
single
one
of
them,
including
handling
symlink.
So
definitely
I
I
can
advise
extra
caution
on
the
same
link
handling.
It
can
easily
get
pretty
nasty,
so
yeah.
D
Yeah
and
thanks
all
right
further
review
and
watching,
unfortunately,
it
looks
like
for
electron
builds
on
Mac
OS.
Sim
links
are
used
fairly
extensively
for
the
frameworks
to
link
together
the
frameworks.
It
doesn't
affect
clinics
of
Windows
systematic
issue,
but
it
is
an
issue.
There
I
think
the
tar
balls
explode
by
about
a
factor
of
three
if
we
dereference
all
the
sim
links.
So
a
non-trivial
issue
but
I
also
agree
a
siblings
are
danger
zone.
D
So,
let's
be
careful
so
in
preparation
for
crew
supporting
symlinks
we've
been
doing
work
to
factor
out
the
kubernetes
visualization
support
that
up
until
now
has
been
inside
of
the
big
Kui
mono
repo
there's
now
a
standalone
repo
that
I've
linked
in
the
and
the
Google
Docs
that
you
can
try
today.
D
So
you
can
get
clone
that
NPM,
CI
and
NPM
start
and
you'll
have
Kui
come
up
within
about
10
seconds
or
so
15
seconds,
and
then
you
can
do
all
your
coop
commands
visualizations
that
clients
you
can
now
separately
billed
to
standalone
electron
client.
There
are
build
targets
for
targeting
accrue
turbo,
which
bundles
everything
up
nicely
back
as
you
that
perfectly
for
crew.
D
So
we're
mostly
secured
up
for
enabling
the
coop
visualization
as
a
plugin
pending,
is
some
like
the
importance
in
like
work
to
sort
of
flesh
out
some
of
this
work.
We've
started
on
a
few
other
prototype
plugins
log
visualization
is
one,
that's
obviously
been
in
frequent
demand.
The
Kubb
UI,
basically
looking
at
some
very
primitive
support
for
offering
log
support
along
visualizations.
D
One
of
the
goals
of
maybe
we
can
shoot
forward
with
a
log
UI
plug-in
is
just
to
consolidate.
I
know,
there's
a
lot
of
work
out
there
for
block
analysis.
You
know
occupation,
filtering
and
even
visualization,
so
maybe
we
can
start
consolidating
some
of
that
and
offering
new
stuff
if
needed,
so
that
we
can
have
quick
pop
up
support
for
answering
questions
about
law.
Contents,
visualizing
long
timeline,
things
like
that,
so
I've
linked
those
two
in
the
docs.
D
A
Okay
hearing
none
with
that
I'm
gonna
close
the
call.
Thank
you
very
much,
I
hope
to
see
all
of
you,
or
at
least
majority
of
you
in
two
weeks
in
San
Diego
and
the
regular
meetings
will
happen
in
four
weeks
time
frame.
So
if
you
have
anything
that
you
want
to
learn
before
the
code,
freeze
just
feel
free
to
ping
ping
myself
and
other
others
in
there
slack
thanks
a
lot
have
a
good
one.