►
From YouTube: Kubernetes SIG CLI 20200909
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
september
9th,
if
I'm
correct-
and
this
is
another
of
our
bi-weekly
sexy
live
calls-
my
name
is
matty
and
I'll
be
your
host
today.
A
So
our
agenda
is
pretty
light
today,
but
I
think
I
want
to
cover
a
couple
topics.
So,
first
of
all
those
announcements,
the
120
release,
dates,
are
official
I'll
link
the
full
schedule.
If
you're
curious,
the
most
important
dates
are
at
the
top,
which
is
code.
Freeze
is
in
the
middle
of
november
on
november
12th
to
be
precise
and
the
enhancements
freeze,
which
basically
means
when
you
have
when
you
need
to
have
your
kubernetes
enhancements
proposal
merged
is
october
6th.
B
No,
I
don't
okay,
so
yeah
go
ahead.
Sean.
C
Oh
no,
I
was
just
saying
that
I
don't,
I
think
my
my
sound
doesn't
work
very
well
with
this
web
version
of
zoom.
I'm
gonna
have
to
get
this
yeah.
A
It
is,
it
is
chopped
a
little
bit
just
like
it
was
last
time
when
I
was
telling
you
that
one
oh
I'll
work
on
that
the
app
works.
Okay
on
linux
as
well,
I'm
running
fedora.
It
works
fine.
A
A
A
The
biggest
issue
that
we
had
at
the
end
of
119
was
that
the
queue
was
very
likely.
There
was
a
lot
of
test
test
failures
and
so
forth,
and
there
are
discussions
about
making
120
stabilization
release.
Although
the
majority
of
people
claims
that
it's
a
little
bit
too
late
to
be
announced
a
civilization
release
because
pretty
much
everyone
had
their
plans
already
laid
out
in
front
of
them.
A
Sig
leads
to
be
mindful
about
the
area
and
I
linked
here
both
the
test
grid
for
six
cli
and
a
a
flake
tracker
tracker,
I'll,
probably
link
also
what
I
was
talking
about
last
time
is
a
document
from
jordan
about
deflaking
the
queue
and,
generally,
the
ask
is
to
ensure
that
your
area,
in
our
case
cube
cuddle,
related
tests.
A
Are
not
affecting
the
queue
significantly
and
ensure
that
we
have
sufficient
coverage
in
in
the
tests.
For
for
that
I
remember,
and
I
still
owe
brian
reviews
because
he
looked
out
at
the
test
score.
There
are
several
issues
related
and
I'm
so
deep
in
everything
that
I
need
to
do
that.
I
still
owe
him
a
couple
reviews
on
that
one,
but
I'm
hoping
probably
sometime
next
or
the
following
week.
I
should
have
my
my
backlog
cleaned
up
a
little
bit
to
be
able
to
go
through
these.
A
So
that's
pretty
much
all
that
I
wanted
to
bring
up
to
the
sick
if
anyone
have
any
topics
that
they
want
to
talk
with.
The
group
now
is
your
chance
or
if
you're,
a
new
person
that
hasn't
joined,
I'm
looking
at
the
ad
in
these
lists-
and
I
think
I
recognize
the
majority
of
the
names,
but
I
might
be
wrong
so,
if
you're
new
feel
free
to
speak
up,
introduce
yourself
tell
a
little
bit
about
yourself.
D
E
A
Okay,
welcome.
Do
you
have
any
particular
area
within
six
cli,
whether
that
will
be
cute
cuddle
or
projects
that
we
are
sponsoring
that
you're
particularly
interested
in.
D
A
D
A
Leaving
the
number
of
the
pull
request
and
the
slack
that'll
be
probably
the
easiest
for
me
to
track
to
track
it.
I'm
the
guy
that
announced
that
there's
a
60
lie
in
five
minutes.
A
Okay,
so
welcome
again
I'll
have
a
look
I'll
I'll.
Definitely
have
a
look
at
your
pr.
If
it's
a
bug
fix,
then
definitely
that's
something
that
we
can
merge
right
away.
Sure
sure.
D
A
F
Oh
hey
all
I
am
from
vmware.
I've
been
involved
with
communities
for
a
few
months
now,
and
I
was
just
like
looking
at
all
the
sigs
and
attending
a
few
meetings
to
see
how
the
community
works
and
how
the
how
the
work
gets
distributed
and
how
just
getting
to
know
people
involved
in
the
community
itself.
A
I
wasn't
sure,
because
I
was
super
quiet
and
I
wasn't
sure
whether
you
wanted
to
say
something
or
I
was
unrelated,
okay,
cool
welcome.
Do
you
have
any
particular
area
after
your
initial
overview
that
you'll
be
interested
in
or
particular
feature
a
book
or
a
particular
work
within
cube,
cuddle
or
60
li
brother
that
you're
interested
in.
G
F
A
Okay,
I'll
definitely
encourage
both
of
you,
the
new
person
to
join
the
bucks
crap.
A
H
A
Okay,
so
welcome
and
definitely
join
the
call.
Next
week
we
are
usually
going
through
the
bugs
prs
and
issues.
F
A
I
Go
ahead,
I
just
if
you
scroll
up,
I
just
there's
a
couple
issues
on
there,
the
first
one
we
actually
should
talk
about
today.
The
others
can
wait
till
next
week.
If
we
don't
have
time,
but
but
yao
ping
me
right
before
the
call
he's
not
able
to
make
it
because
it's
midnight
for
him
yeah.
G
I
This
one
antoine
brought
up,
it's
been
open
for
a
bit
he's
suggested.
We
need
a
formal
discussion
on
this
change.
I
don't
have
any
other
context.
Besides
that.
A
I
would
need
to
have
to
specifically
the
pr
I
know
that
yelping
me
as
well
some
time
ago,
but
I
like,
I
said,
not
sure
if
that
particular
one
no.
A
Right
because
we
at
some
point
in
time,
there
was
no
option
to
pick
which
option
the
the
deletion
works
and
over
time
we
introduced
options
to
at
least
in
the
api,
how
you
want
to
progress
with
with
delay
deleting,
and
that
includes
foreground,
which
basically
means
you
throw
the
delete
operation
to
the
server
and
the
server
will
handle
everything
background
other
way
around
foreground
you're
waiting,
basically
for
the
entire
delete
process
to
happen
background.
A
It
happens
in
on
the
server
and
you
you
can
pull
the
server
and
orphaning
basically
means
that
the
surface
type
process,
which
is
responsible
for
tracking
all
the
children's
of
your
process.
For
example,
a
deployment
creates
a
set
of
pots
when
you
delete
deployments
it
also
by
default,
removes
all
the
auto
parts.
Orphaning
means
I'm
good.
I
just
want
to
remove
the
deployment,
but
the
parts
should
be
left
behind
for
whatever
reason,
and
if
you
invoke
the
api,
you
have
the
option
to
pick
one
of
the
three
and
keep
cuddle
does
not
have
that
option.
A
A
We
do
the
foreground
deletion
and
this
change
I'll
definitely
have
a
look
at
the
pr
and
then
we
can
decide
whether
if
this
is
a
user
facing
change,
a
significant
one,
we
might
want
to
have
a
a
cap,
but
I'm
not
100
sure.
So
I'm
not
I'm
not
saying
one
way
or
the
other
and
I'll
sync
with
with
antoine
with
regard
to
that
one
as
well
so
yeah.
It's
definitely
something
that
I
would
need
to
check.
Yeah.
C
Sorry
to
interrupt,
I
have
just
a
little
bit
of
context
on
that,
so
so
yeah
as
as
you,
I
think
you
mentioned
before,
there's
there's
only
a
boolean
now
for
whether
we
cascade
on
delete
and
it's
going
to
be
modified
to
be
a
string,
so
it
is
actually
kind
of
an
api
change
in
that
it's
changing
the
flag.
It's
so
so
it
is
relatively
significant
in
that
regard.
A
Okay,
we
still
have
lots
of
time,
so
we
can
definitely
check
the
remaining.
I
I
So
the
tldr
with
this
one,
I
believe
this
person
is
working
on
scaffold
at
google
and
they
realize
that
when
you
do
a
port
forward,
it
picks
the
oldest
pod
instead
of
the
newest
pod.
So
if
you
do
a
rolling
deployment
or
a
new
pod
is
being
created,
it
won't
grab
the
latest
change.
Pod
it'll
instead
grab
the
oldest
pod.
It
doesn't
have
your
newest
changes
and
he
kind
of
laid
out
how
all
that
actually
works,
and
so
his
proposal
is
to
flip
that
logic.
A
A
We
probably
would
have
to
figure
it
out
whether
because
that
particular
mechanism
is
then
being
used
for
getting
locks.
If
you
don't
specify
a
particular
container
and
pod
and
there's
more
than
one.
A
I
would
have
to
look
at
the
particular
details
why
we
came
up
with
this
and
not
the
other
way
approach,
but
at
the
same
time
yeah.
It.
B
Is
this?
Is
this
just
a
capability
they
need,
or
do
they
want
the
default
to
be
this
for
some
reason,
because
if
it's
just
something
they
need
that
isn't
currently
supported,
we
could
safely
add
like
a
flag
that
enabled
some
control
over
the
you
know
the
preference
for
which
pod
is
attached
to
or
forwarded
to
rather.
B
A
Yeah,
it's
it's
a
little
bit
more
complicated.
I
guess.
A
I
mean
the
question
is
more
about
their
particular
use.
Case
is
for
having
newer,
but
then,
if
you
look
at
it,
look
at
it
from
a
broader
picture,
it's
only
a
matter
of
time
when
someone
shows
up
with
oh,
I
want
to
have
that
particular
pod
over
newer
and
then
we
would
be
coming
up
with
yet
another
mechanism
and
yet
another
flag,
and
something
like
that.
So
I'd
rather
have
some
kind
of
a.
A
A
A
There
were
some
changes
with
regards
to
cube
cuddle
top
pod.
I
can't
remember
if
the
entire
backing
mechanism
is
being
removed
or
something
like
that.
I
I
had
a
pr
out
for
top
pod
for
sorting
it
had
to
do
with
sorting,
I
think,
does
your
pr
switch
it
to
use
the
builder
resource
builder
thing,
because
currently
it's
not
right
now,
I
I
looked
into
this
a
little
bit.
It's
just
looks
like
we
need
to
use
the
builder
and
it's
kind
of
using
old
logic.
Instead.
A
A
A
Someone
was
literally
mapping
all
of
the
options
that
the
client
go
has
over
to
to
the
keep
cuddle,
including
tokenfile,
but
I'll
I'll,
dig
through
it
and
find
where
we.
What
was
the
answer,
then
it
might
be
that
I
was
not
too
super
in
favor
of
having
all
the
possibilities.
I
I
Popping
up
this
ci
cd
has
popped
up
a
few
times,
one
with
the
base64
encoded
token.
So
I
think
we're
eventually
going
to
have
to
talk
about
an
actual
supported
use
case
for
this.
B
Yeah
well,
there's
I
think,
like
the
coup
control
like
the
rebase
on
hugo,
that
siam
is
doing
as
part
of
the
season
of
docs.
B
Where
we're
like
re,
the
control
book
is
being
rebased
on
hugo
and
then
the
plan
is
to
fold
the
customizing
control
content
into
a
single
page
single
site,
so
that
site
like
as
part
of
that,
we
could
try
and
figure
out.
Where
is
the
right
place
for
this
sort
of
stuff?
And
maybe
it's
a
matter
of
it
just
not
being
surfaced
at
the
right
particular
location.
A
I
mean
the
this
is
the
type
of
a
feature,
all
the
feature
requests
and
everything
like
that.
We
were
talking
about
it
to
put
it
in
the
faq
which
would
be
living
inside
of
the
cube
color
repo,
and
the
premise
was
that
we
would
have
a
template
when
people
will
be
opening
features.
A
That
would
point
them
to
this
fact,
stating
if
you're
trying
to
open
issues
that
are
related
to
one
of
the
questions
that
we
have
over
there
make
sure
that
it
is
covered
there
and
if
it's
covered
don't
open,
because
there
is
99
chance
that
we
will
close
it
so
that
that
was
the
goal
when
eddie
started.
Writing
this.
This
fact.
B
Yeah,
no,
I
think
I
remember
that
now.
That
makes
a
lot
of
sense.
I
think
I
was
thinking
like
if
there's
like
a
bunch
of
if,
if
the
answer
for
a
lot
of
these
issues
is
use,
this
set
of
bash
commands.
Instead,
we
might
want
to
just
add
a
page
of
like
here's.
All
the
different
bash
commands
that
we've
recommended
in
the
past.
A
That
reminds
me
that
that
reminds
me
ahmed's
the
he
has
a
script
that
generates
aliases
for
all
the
commands
as
a
shortcuts.
So
you
can
have
very
efficient
commands
through
simple
command
through
very
simplistic
commands.
So,
instead
of
writing
entire
cube
cuddle
get
pods
with
an
ancient
or
whatever
you
you
can
type
kgp
and
there's
so
many
more
combination
in
an
acme
solution.
A
So
that's
pretty
neat,
but
the
downside
is
that
the
europe
rc
grows
significantly
and
I'm
worried
if
we
would
come
up
with
all
the
possibilities
that
people
come
up
with
for
injecting
I
don't
know
files
and
whatnot.
A
I
might
grew
the
our
documentation
more
than
the
actual
content
that
we
would
want
to
care
about.
I
This
one
specifically,
is
where
I
get
stuck
and-
and
I
I
get
that
we
want
people
to
use
bash,
but
like
this
functionality
already
lives
in
client
go
and
all
we
would
need
to
do
is
wrap
this
with
the
flag
and
use
the
client
go
logic.
So
if
it's,
this
is
hard
right.
This
is
where
it
gets
hard
and
I,
like
I
said
I
honestly
think
that
we're
going
to
see
ci
cd
stuff
pop
up
way
more
because
people
are
starting
to
adopt
it
more.
I
A
So
here's
the
thing
because
he
is
when
you
embed,
cube
cuddle
in
the
cluster
and
he
specifically
speaks
about
a
an
in
cluster
use
case
because
he's
injecting
var
run
secrets
kubernetes.
I
o
service
account
token
when
you
inject
cube
cuddle
in
your
cluster
and
you
don't
specify
any
flags
for
configuration
or
anything
like
that.
The
default
mode,
and
probably
a
lot
of
people
are
not
aware
of.
It
is,
as
the
first
thing
cube.
A
A
You
are
currently
running
in
so
it
only
it's
only
a
matter
of
injecting
a
proper
service
account
into
your
pod
and
then
invoking
the
queue
catal
commands
like
you
would
normally
do,
and
it
will
pick
the
right
in
cluster
configuration.
You
can
easily
verify
that
by
bumping
the
verbosity
level
to
I
think
at
two.
We
are
printing
which
configuration
we
are
picking
and
you
will
notice
as
one
of
the
first
elements
that
it
will
say
that
it
is
using
in
cluster
configuration.
A
A
Unless
somebody
is
trying
to
do
something
different
and
it's
not
injecting
the
the
search.
The
way
that
that
we
initially
thought
of.
A
I
A
C
One
other
quick
question:
there's
a
minus
minus
coupe
config
flag
to
point
to
a
specific
coupe
config
and
within
the
coupe
config
you
could
have
a
token.
Is
that
correct,
yeah.
A
A
Oh
this
one
psy
wants
to
talk
about
the
local,
it's
not
quite
local
across
all
the
commands.
What
about?
I
was
recently
talking
about
it
with
my
co-workers.
J
So
I
was
thinking
that
we
could
make
a
structure
of
common
options
and
then
populate
them
in
the
complete
function
call
and
then
all
the
different
subcommands
would
use
this
particular
function
to
populate
the
common
fields
and
then,
if
any
specific
two
sub
commands
are
there,
they
can
do
that
after
the
function
call
and
then
basically
all
we
don't
have
to
duplicate
a
lot
of
code
in
all
the
subcommittee.
A
I
think
you,
you
left
that
kind
of
a
comment.
A
Oh
I'm
definitely.
I
would
like
to
see
this
in
an
action.
So
if
you
could
sketch
up
a
proof
of
concept
pr
and
then
we
can
look
at
it
during
one
of
our
next
6eli
calls,
I
think
it's
it's
a
reasonable
approach,
probably
the
simplest
one
that
we
could
come
up
with.
A
A
Eddie
you
were
looking
at
is
there
anything
else
that
we
wanted
to
talk?
Okay,
those
are
the
three
that
I
needed.
Some
input.
I
A
Stand-Ups,
do
people
have
topics
and
I'm
looking
at
crew,
koi
and
customize
that
I
want
to
share
with
the.
K
K
B
So
a
number
of
years
back
as
part
of
the
probably
second
version
of
the
kubernetes
stocks,
they
were
developed
by
a
company
that
we
contracted
with.
I
think
we
paid
them
to
develop
those
and
we
spent
some
time
co-developing
or
like
giving
giving
feedback
and
helping
directing
those-
and
I
don't
think
they've
really
been
changed
much
since
so
we
are
free
to
change
those.
We
are
free
to
do
effectively
anything
we
want
with
them.
K
And
like
in
terms
of
just
the
coup,
the
six
cli,
for
example,
customized.
Is
there
any
desire,
even
forget,
even
about
intensive
any
desire
to
you've
mentioned?
You
know
some
of
the
transition
from
the
git
books
to
hugo?
Is
there
any
desire
or
need
perceived,
need
to
work
on
the
catacota
side,
or
do
you?
Is
it
already
foreseen
that
hugo
solution
will
be
more
than
good
enough.
B
I
think
the
hugo
will
probably
be
good.
I
mean
if,
if
we
had
more
like
like
a
pool
of
folks
wanting
to
improve
stuff,
I
think
there's
areas
I
could
direct
folks
to.
I
wouldn't
pick
catacota,
as
my
first
choice
like
I
think.
Catacota
is
a
great
experience
for
I'm
coming
to
the
website
and
I
want
to
have
successfully
run
something
in
a
way
that
feels
like
it's
running
on
kubernetes
right
and
then
that's
pretty
much
what
you
get
out
of
that
he's
like
I
ran
a
pod.
B
You
know
what
I
mean,
and
I
can
do
that
again.
I
think
that's
that's
critical
point
in
the
user's
journey,
but
but
it
only
lasts
maybe
an
hour
right
and
then
I
don't
envision
anyone
ever
going
back
to
that
page
after
that
hour
is
done.
B
Right
right,
like
it,
the
there's
part
of
the
reason
the
catacomb
too
right
is
because,
like
with
kubernetes,
is
like
okay.
First,
you
have
to
get
a
cluster
and
when
that
stuff
was
actually
written
like
I
don't
I'm
not
even
sure
mini
cube,
existed
and
definitely
not
like
docker
on
desktop
kubernetes
or
or
kind
or
any
of
these
other
things.
So
but
I'd
say
it's
a
combination
of
as
the
rest
of
the
system
has
matured,
there's
probably
less
need
for
bootstrapping
folks
with
a
sort
of
clustery
thing,
but.
B
But
also
that
the
like,
I
don't
think
it
scales
to
learning
about
the
apis,
for
instance
right.
So
so,
like
okay,
like
what
do
you
do
after
that?
It's
like?
Well,
I
need
to
learn
about
deployments.
I
need
to
learn
about
services.
I
need
to
learn
about
config
maps
and
at
that
point
in
time
I
think
it's
fair
to
tell
someone
like
go,
download,
mini
coop
or
like
go,
go,
download
something
and
get
your
own
cluster,
because,
like
you're
not
going
to
be
wanting
to
do
all
this
in
this
browser.
K
Yep
yeah,
I
guess
the
the
thing
that
occurred
to
me
is
I
was
reading
through.
I
was
trying
to
do
a
compare
and
contrast
of
all
these
different
forms
of
docs.
We
have
out
there
and
it
does
seem
like
a
lot
of
the
more
recent
docs
we've
written
are
very
good
but
they're
written
sort
of
in
that
kind
of
kind
of
tutorial
kind
of
style.
You
know
they
have
execute
this
command.
You'll
see
this
output.
K
Here's
an
example,
input
file,
they're
kind
of
had
they
had
that
notebook
kind
of
structure,
but
they're
written
as
static
documentation,
which
that
seems
pretty
effective
just
as
curious
as
whether
that
was
an
intentional
choice
to
do
that
in
the
world
of
html
rather
than
category.
B
I,
like
we
have
not
changed
those
to
my
knowledge
since
they've
been
written
like
so
many
years
back,
I
think
catacota
may
do
like
upgrade.
We
may
upgrade
the
kubernetes
version
and
do
some
of
that
stuff,
like
some
very
fine,
like
some
very
focused
sort
of
changes
to
those,
but
we
haven't
really
changed
the
content.
I
think
it's
it's
costly
to
change
that
content.
It's
difficult,
yeah,
like
it's
costly
from
an
engineering
perspective
right
like
I
can
write
in
the
markdown
file.
B
If
something
takes
me
an
hour
to
do
right
to
put
the
doc
together,
like
paste
input
output,
it
I
think
it's
fine
like
that.
I
could
do
learning
catacota,
it's
like.
How
do
I
even
test
this
thing
and
how
do
I
push
it
and
I
think
it
requires,
like
maybe
the
the
cat
like
we
have
to
upload
binaries
or
these,
which
are
maybe
not
binaries.
We
have
to
upload
the
whatever
the
scripts
are,
that
kind
of
describe
how
the
catacombs
systems
run
and.
K
Makes
sense,
okay,
this
is
very
helpful.
Thank
you.
Yeah
we've
been
working
on
a
few
things,
sort
of
separately
for
different
use
cases,
but
it
occurred
to
me
as
we
were
working
through,
that
there
might
be
some
intersection
with
some
of
this
so
and
maybe
on
next
next
week's
call
a
little
premature
now,
but
maybe
next
next
times
call
we
can.
I
can
give
you
some
demos
of
some
of
the
stuff
we've
been
thinking
about
dinner.
B
B
I
think
my
my
thoughts
on
docs
are
like
the
two
places
where
you
get
the
most
return
on.
Your
investment
is
one
is
like
simple
reference
and
guides
like
you
want
your
reference
to
be
really
good,
because
that's
what
people
I
think,
are
going
to
space,
but
that's
where
people
are
going
to
spend
the
most
time
over
the
course
of
their
like
experience.
Working
with
your
product.
B
B
I
learned
about
so
really
just
like
having
pretty
thorough
reference
documentation
which,
which
we
don't
have
as
much
as
I'd
like
now,
but
we're
working
on
then
the
the
guides
after
that,
which
just
make
sure
that
those
route
to
the
right
reference
documentation
and
provide
the
right
content
concepts
in
context
for
all
the
information
in
the
reference,
so
that
you
can
make
that
transition
without
being
frustrated
and
have
the
right
mental
model
and
then
actually
the
biggest
one
that
we
don't
do.
A
Also
important
factor
is
that,
if,
if
the
docs
is
a
simple
md
file
that
I
can
fix,
if
there's
a
typo
there's
an
error
or
anything
like
that,
if
it's
a
simple
md
file
and
I'm
I'm
gonna
close
the
repo
clone,
the
repo
fix
it
within
30
seconds-
ends
up
in
apr
versus
if
it's
a
katakura,
then
I'll,
probably
just
won't
which
which
which
which
greatly
improves
the
overall
quality
of
the
docs,
because
the
entry
barrier
for
fixing
a
text
document
is
super
simple
yep.
A
K
That's
a
great
point
yeah.
So
in
the
huge
one
last
question
on
the
hugo
work,
I
I'm
definitely
a
hugo
newbie
here,
but
as
far
as
I
can
tell
from
hugo,
you
have
the
option
of
either
giving
it
straight
up.
Html
or
giving
it
markdown,
and
it
kind
of
it
can
kind
of
mix
and
match.
K
Is
the
goal
for
the
for
science
and
the
work
for
the
summary
of
summer
of
docs?
Is
that
the
goal
to
be
purely
marked
down
speaking
to
mache's
point,
or
is
there
going
to
be
some
lingering
html
raw
html.
B
The
goal
is
so
currently
it's
in
markdown,
it's
in
a
markdown
with
the
get
book
format
it's
being
migrated
to
markdown
in
the
hugo
format.
I
think
the
big
difference
there
is
sort
of
how
stuff
is
how
a
particular
markdown
file
may
be
rendered.
B
B
They
support
like
here's,
how
you
draw
a
graph
in
one
versus
the
other,
so
there's
not
a
lot
of
those
in
the
content.
I
think
there's
very
few
actually,
but
those
have
to
be
migrated.
It's
not
a
big
portion
of
the
work.
The
real
big
chunk
of
the
work
is
going
to
be
basically
fall
into
two
categories.
One
is
restructuring
the
get
book
to
factor
guides
and
reference
separately
right
now,
they're
kind
of
mixed
together.
B
I
think
it
was
an
interesting
experiment.
It's
written
as
kind
of
a
book
format
and
books
tend
to
mix
like
if
you
look
at
the
inaction
books,
they
progress
without
having
separate
referencing
guides,
but
looking
at
kind
of
the
doc
sites.
My
experience
since
then
has
been
that
having
those
separate
actually
is
very
valuable.
People
expect
to
have
a
strong
reference
section,
so
it's
splitting
the
content
there
and
then
the
second
piece
is
the
kind
of
the
order
things
were
written
right.
B
Is
you
had
there's
a
bunch
of
documentation
on
the
coop
control
website
on
the
cncf
or
on
the
kate's
cncf
website
that
is
intermingled
with
all
sorts
of
other
stuff
about
kate's
right,
because
anytime,
you're
teaching
anything
to
someone
about
kubernetes.
You
kind
of
use
coup
control
to
demonstrate
that
right.
So
that's
all!
There's
a
bunch
of
documentation
there
plus
there's
documentation
that
was
written
early
on
purcha
control.
B
There's
this
stuff
in
the
book
that
was
kind
of
an
attempt
to
reorganize
a
lot
of
the
stuff
that
was
either
emitted
from
the
website.
The
kubernetes
website,
or
just
not
well
structured
and
bring
in
the
customize
content,
because
when
that
was
written,
like
customize
was
hitting
a
point
where
it
needed
to
be.
B
Introduced
to
coupe
control,
I
think,
and
so
it
was
an
attempt
to
bring
those
two
together
and
then
the
third
thing
you
have
is
you
got
the
customized
docs,
which
exist
as
markdown
files
in
the
typical
git
approach
with
is
just
a
bunch
of
like
md
files
that
aren't
rendered.
They
don't
have
any
platform
for
using
them
and
that
content
spans
old
new,
but
it
hasn't
really
been
strongly
organized.
B
And
then
I
took
a
stab
at
writing
like
an
actual
reference
section
for
customize
before
it
was
mostly
just
examples
of
customization
files
and
guides,
but
there
wasn't
like
a
great
here's
every
field
in
the
customization
enamel
and
what
it
does
example
using
that
field.
B
So
I
took
a
stab
at
that
and
it
was
mostly
taking
existing
content
with
a
little
bit
of
writing.
More
put
it
in
the
cubicle
site.
You
can
go.
Look
at
the
customized
site
now,
maybe,
as
ever
presenting,
could
bring
it
up,
and
so
the
the
more
difficult
of
the
tasks,
but
I
think,
is
going
to
pay
off,
is
in
addition
to
breaking
the
code
control
book
into
guides
and
reference
also
at
the
same
time
integrating
it
with
the
customized
guides
and
reference.
B
So
the
customized
site
now
has
guides
and
reference,
and
if
you
look
at
what's
in
the,
if
you
just
look
at
the
content
in
the
get
book
for
cube
control
in
the
content
of
the
customized
guides
and
reference,
there's
a
lot
of
overlap
there
now,
because
different
sites
got
developed
at
different
times.
But
you
couldn't
really
admit
like
it
wasn't
pos.
You
wouldn't
have
a
good
experience
emitting
that
content
when
they
were
on
different
sites
right
because
it'd
be
going
back
and
forth
between
different
sites.
B
But
now,
with
the
sites
being
integrated,
you
can
kind
of
dedupe
and
say:
okay,
let's
just
create
a
good
reference
section
on
generating
config
maps
and
secrets,
and
then
we
in
the
guides.
We
won't
talk
about
as
much
of
all
the
config
map
and
secret
generating
options
and
what
they
are
we'll
simply
just
give
a
thinner
example
of
doing
it
once.
I
I
want
to
throw
out
there
real
quick.
I
don't
know
if
it's
true
anymore,
because
catacota
is
now
owned
by
o'reilly,
but
ben
hall.
The
creator
of
catacota
was
super
helpful
in
regards
to
implementing
things
with
it.
Kubernetes
is
actually
one
of
their
featured
courses
that
they
use
in
sales,
demos
and
stuff.
So
if
there's
anything,
we
want
like
tweaked
on
that
stuff.
I'm
sure
we
can
reach
out
to
ben
and
see
if
he
can
get
some
resources
to
help
with
that.
B
Yeah,
I
I
want
to
make
sure
thank
you
eddie,
like
I.
Our
experience
with
catacota
has
been
very
good
from
perspective.
Ben's
been
very
helpful
and
it
does
it
works
as
advertised
and
if
you
want,
if
you
want
to
develop
an
interactive
experience,
I
feel
positively
about
our
experience.
Working
with
them.
That's.
A
Okay,
thanks,
we
have
five
minutes
still
left
stand.
Ups.
Does
any
of
the
sub
projects
want
to
share
with
folks.
G
I
can
take
30
seconds
and
talk
about
customize.
G
I
don't
know
primarily
it's
an
appeal
for
some
help.
Let
me
throw
something
in
the
chat
here.
G
So
we're
focusing
on
that
and
focusing
on
stability
fixing
the
bugs
that
are
being
reported
by
you
know.
Super
users,
like
haley,
really
well
documented
bugs
that
we
can
fix
quickly,
but
the
focus
is
on
reintegration
with
coop
cuddle.
So
if
anybody's
got
some
bandwidth
for
help,
you
would
be
appreciated.
E
I
should
actually
have
some
time
for
this.
Iteration
we've
been
in
crunch,
working
on
our
first
ga
and
unfortunately
we're
having
depend
3.61
of
customized
for
that
but
yeah.
Definitely
our
goal
is
to
get
back
on
the
latest.
I
I
want
to
bring
this
up
at
probably
the
controvex
meeting,
but
tacktoberfest
is
next
month
and
if
we
tag
some
issues
with
hacktoberfest
we'll
get
magical
contributions,
because
the
oktoberfest
website
surfaces
all
the
issues
tagged,
so
that
might
be
something
to
bring
up
with
sig
controbex.
But
I
feel
like
we
get
a
lot
of
help
there.
If
we
tag
some
issues.
I
I
will
drop
you
a
link,
it's
it's.
An
open
source
online,
hackathon
and
contributors
make
four
or
five
pull
requests
in
the
month
of
october
and
they
get
a
free
t-shirt.
It's
run
by
github
and
digitalocean.
It's
been
going
on
for
like
six
or
seven
years
now,
and
I
think
last
year
we
had
a
question
when
I
still
worked
there,
we
had
like
80
000
contributors
like
work
on
projects
and
get
t-shirts.
So
there's
some
cool
stuff.
A
I
A
Okay,
I
think
that's
pretty
much
all.
Thank
you
very
much
for
today,
a
reminder
we
have
a
box
crop
next
week,
so
I'm
hoping
to
see
you
next
next
wednesday.