►
From YouTube: Kubernetes Community Meeting 20170914
Description
We have PUBLIC and RECORDED weekly video meetings every Thursday at 10am US Pacific Time.
https://docs.google.com/document/d/1VQDIAB0OqiSjIHI8AWMvSdceWhnz56jNpZrLs6o7NJY
A
B
B
All
right,
hi
everybody
I,
am
Harris
Pittman
I
just
started
with
Google
under
Sara
Novotny
working
on
kubernetes
community
stuff
I
do
not
necessarily
represent
one
sig,
but
I
will
be
working
across
SIG's
on
many
upcoming
programs
like
awesome
mentoring,
this
election
and
many
other
things.
So
just
a
reminder
to
please
mute
when
you're
speaking,
we
do
have
a
demo
today
and
if
you
save,
updates
and
demo
first
up
is
Ed,
Lee
and
or
you
want.
D
C
C
C
It
kind
of
provides
another
choreography
for
works,
that
you
run
in
communities
to
make
it
easy
for
users
to
submit
jobs
and
to
make
sure
that
the
kubernetes,
your
covert
kubernetes
cluster,
is
well
fed
and
ended
for,
and
we
have
the
URLs
for
our
github
project
and
so
on,
which
I'll
touch
up
on
a
little
bit
later,
so
one
sec,
okay
next
slide.
So
what
is
the
continued
native
workflows
or
what?
What
this
arvo
do?
C
Argo
is
a
workflow
engine
and
where
each
step
in
the
workflow
is
a
container
which
you
can
define
to
be
whatever
you
want.
It
also
has
included
integrated
artifact
management
since
oftentimes,
when
you're
running
a
workflow
you're
going
to
generate
derived
objects
or
artifacts
or
the
output
by
one
step,
and
then
you'll
want
to
consume
them
in
another
step
in
order
to
do
additional
processing.
It
also
has
something
called
fixtures
which
you
can
think
of
it
as
a
easy
way
to
spin
up
services.
Their
comment,
like
you,
know
many
deployments.
C
If
you
will
select
right
now,
they
just
end
up
on
my
SQL
database
as
a
picture.
You
know
stored
data
and
it
run
some
tests
if
you
like,
and
then
the
picture
will
automatically
be
destroyed
on
your
workflow
finishes.
So
it
all
really
kind
of
manages
the
lifecycle
of
picture
to
make
sure
that
I
work
will
doesn't
start
until
it
sure
is
ready
and
then
destroys
the
fixture
at
there.
So
that
is.
B
C
C
C
C
C
It
gives
you
a
brief
overview
of
our
NGO
and
then
has
some
simple
instructions
to
get
started.
If
you
just
need
to
download
the
Argo
TOI
and
then
when
you
run
our
cluster,
it
will
install
both
a
kubernetes
cluster
as
well
as
the
Argo
system,
on
top
of
the
cluster,
we're
working
on
a
version
that
will
let
you
just
in
the
article
on
any
existing
kubernetes
cluster.
But
today
you
you
need
to.
It
will
create
a
separate
cluster
on
your
a
journalist
account
okay,
so
now
to
the
demo.
C
C
Take
a
look
at
the
mo
to
see
what
it
looks
like
so
here's
kind
of
the
high
level
workflow
it's
called
input,
DBC
I
hit,
say
example,
CI
reply.
You
say
CI
workflow,
mainly
because
most
people
are
familiar
with
how
CI
works,
although
I
really
think
that
workflows
longer
term
should
really
be
it'll.
Probably
used
more
for
things
that
are
not
CI,
especially
by
average
users,
they're
useful
release
to
get
any
kind
of
work
done
to
automate
the
jobs
and
so
forth.
C
You
can
kind
of
see
by
the
structure
depth.
First,
there
is
a
kind
of
a
description
section
at
the
top
and
then
there's
an
input
output
section
in
this
particular
case.
This
workflow
has
no
output,
so
don't
have
inputs.
The
two
inputs
are
the
commit
ID
and
the
repo
that
you
want
to
build,
and
then
the
steps
of
the
workflow
consists
doing
a
check
out
building
the
code
and
then
running
a
couple
tests.
These
two
tests,
you
can
see,
is
an
array
and
the
mo.
So
there
you
will
be
running
in
parallel
and
it's.
C
Test
as
well
as
a
kind
of
an
end-to-end
test,
end-to-end
test
is
going
to
actually
spin
update.
In
some
DB
instance,
I
was
built,
build
previously
and
then
run
run.
Some
requests
against
a
tinkerer.
Another
thing,
you'll
notice,
is
that
between
these
cells
were
testing
artifacts
as
arguments
between
the
stuff.
So
this
first
argument,
for
example,
is
a
code
artifact,
which
was.
C
Output
from
the
checkout
step,
which
is
the
previous
step,
the
unit
test
step
also
takes
about
the
same
artifact.
The
instructs
vve
to
each
have
takes
a
slightly
different
artifact.
Its
face
to
the
input
be
be
binary
which
will
generate
an
artifact
as
an
output
from
the
build
step.
So
this
is
the
kind
of
artifact
management
really
makes
it
easy
to
kind
of
have
derived
objects
between
cells
and
those
been
aiming
everything
for
you
and
delete
any
artifacts
that
are
not
necessary,
based
on
the
schedule
or
once
in
a
while.
They
take.
C
C
So
this
notes
F,
takes
as
input
the
code
artifact,
which
was
generated
by
the
check
out
that
in
this
case
and
places
that
as
the
specified
path
and
then
as
output,
it
generates
the
ins
of
DB
object,
which
is
apparent
at
the
Kogan
in
flux,
the
tap,
and
you
see
that
it's
going
to
be
going
to
stay
Golan
container
and
then
running
some
commands.
This
is
a
built.
C
Server
template
in
the
steps
portion
is
going
to
first
run
a
BIOS
lab
which
initializes
the
input
CD
and
then
it's
going
to
start
up
three
producers
in
parallel,
which
would
generate
records
in
the
influx
DB,
and
you
can
see
that
it
accesses
the
IP
adjust
the
input
VD
as
a
parameter.
Here
is
the
IP
address
of
the
input
CD
of
this
fixture,
which
we
defined
up
here,
so
the
ideas
will
be
assigned
automatically,
and
this
is
the
way
or
there's
network
access
this
service.
C
Once
the
data
has
been
loaded
instance
of
CD,
it
will
run
a
consumer
stuff
which
basically
does
some
gifts
on
the
input
CB
to
get
the
data
that
was
thought
in
system.
Okay,
as
we
look
at
the
news
box,
DB
server
stuff
itself,
you
see
it's
very
simple:
it
basically
takes
one
input,
which
is
the
binary,
which
was
the
input.
Cd
binary,
runs
a
standard,
Debian
container
and
then
basically
runs
let's
DB
surface
so
but
llamó.
Looking
over
other
example,
you.
C
C
This
at
a
local
like
it
sells
if
you
use
the
template,
that's
stored
locally
on
my
workspace.
Other
them
end
up
committed
it.
May
it
makes
it
very
easy
to
modify
your
templates
play
around
with
it
without
having
to
constantly
commit
them
and
create
kind
of
garbage
in
your
github
repository.
So
as
you
are
in
the
developing
your
templates,
this
is
a
very
convenient
feature.
Ok,
the
job
is
and
submit
it.
C
C
Yeah
you
can
see
that
this
is
a
workbook
that
we
just
saw.
You
can
see.
It
shows
a
tree
structure,
view
of
that
job
and
you
can
see
that
it's
currently
running
the
check
out
steps
so
I
do
it
again
is
still
running.
The
check
comes
now
Oh
check
out
that
job's
finished
and
then
it's
going
to
move
now
on
to
the
other
stuff.
So
let's
take
a
look
at
this
job
in
the
GUI
you
see
kind
of
a
graphical
representation
of
this
portal.
C
Kind
of
browse,
what's
happening
and
what's
depth
in
your
code
is
actually
being
executed
along
the
top
here,
you'll
see
a
few
other
tabs,
so
you
can
put
up
a
look
at
all
those
artifacts
that
are
being
generated
and
I
know.
These
are
the
stuff
that
you
like.
Okay,
let's
see
there,
you
could
also
look
at
all
the
logs
that
are
being
generated.
So
if
you
anything
goes
wrong,
the
log
file
see
what
happened
and
there's
also
a
job
summary
and
one.
C
C
I'm
almost
done
this
is
on
one,
you
can
yeah,
and
so
this
was
a
particular
was
a
CI
workflow,
but
obviously
you
can
run
any
kind
of
workflow.
In
particular,
I'd
like
to
run
on
a
buck
simulation.
Did
he
have
a
lot
of
like
parameters?
You
want
to
run
simulations
on,
create
a
very
simple
workflow
that
runs
at
all
and
Terrell
will
get
it
done
really
quick.
So
that's
it.
Thank
you
very
much.
D
Those
meetings
have
been
happening,
Monday,
Wednesday
and
Friday
string
last
week
in
this
week
and
next
week
they
will
go
to
a
daily
cadence
because
we're
going
to
be
getting
closer
and
closer
to
crunch
time
on
the
release.
So
in
my
update,
you'll
see
we're
still
sorting
out
a
lot
of
broken
tests,
especially
upgrade
tests.
My
commentary
is
this
happens.
D
Github
is
always
the
record
of
note,
but
there
are
occasionally
times
where
people
update
a
spreadsheet
and
not
necessarily
reflect
that
change
in
github,
and
that
seems
to
be
what
has
happened
this
time.
So
we're
taking
this
as
a
retro
item
and
we're
going
to
work
very
hard
to
reconcile
that
for
1.9
and
moving
Ford,
there's
no
risk
there
necessarily
to
the
release
other
than
that
we're
just
trying
to
have
full
understanding
of
of
what
is
what
is
different
between
those
two
things.
D
So
the
idea
of
a
centralized
feature
management
process
feels
like
it's
a
little
out
of
sync
with
what's
happening
in
this
release,
so
that,
in
some
ways
is
good.
I
think
that
that
the
people
who
are
contributing
code
and
we're
participating
in
sakes
should
have
a
direct
feeling
of
tying
their
work
to
what
those
themes
are
that
are
being
delivered.
D
D
and
also
cigarette
sure,
and
also
cig
release,
because
these
program
management
types
sakes
all
have
a
lot
of
cross-cutting
concerns,
and
we
need
to
make
sure
that
those
are
being
spoken
for,
because,
though,
the
work
of
product
management
team
is
extremely
important
and
is
how
a
lot
of
great
work
gets
done
to
share
with
the
community
what
what
delivery
are
valuable
or
delivering.
So
it
doesn't
in
any
way
mean
that
SiC
PM
is
not
relevant.
It
just
means
that
there's
ways
in
which
the
work
is
being
completed
and
and
managed
that
are
diverging.
D
There's
also
the
fact
that
one
body
was
supposed
to
be
a
quote
unquote,
a
stabilization
release
did
not
feature
focus,
yet
we
are
sitting
at
a
backlog
of
30
alpha
features
going
into
this,
so
it
doesn't
feel
like
we
necessarily
accomplish
that
goal
per
se.
That
said,
I
want
to
get
my
standard
disclaimer
and
reassurance
that
this
is
very
high
quality
code.
There's
nothing
about
it
that
needs
stabilization.
It's
not
an
unstable
release.
It's
simply.
Where
are
we
focusing
our
efforts?
D
For
that.
So
as
a
release
team,
we
have
tried
very
hard
to
do
late,
late
changes
in
this
process.
So
what
we're
going
to
do
is
we're
going
to
directly
reach
out
to
six
and
to
solicit
feedback
on
what
sig
level
prioritization?
What
for
the
release?
So,
in
some
cases
like,
for
example,
cig
note,
their
their
release
theme
is
actually
more
of
a
rationale
for
the
city
existing
in
the
first
place,
which
is
providing
really
high-quality,
stable
extensible
code
at
the
node
level
and
less
like
specific
features
related
to
to
their
to
this
release.
D
D
So
don't
worry
if
there's
nothing
really.
Concerning
about
this.
It's
the
natural
and
healthy
evolution
of
our
processes,
and
thankfully
we
have
a
community
and
a
bunch
of
governance
around
this.
That's
emerging,
that's
helping
guide
this
work
and
I
feel
confident
that
we're
going
to
deliver
high-quality
release
regardless
so
before,
I
move
on
any
questions
about
all
that,
because
I
was
a
lot.
D
Okay,
silence
equals
lazy
consensus
except
Robert
Bailey.
What
makes
us.
F
I
guess
I
was
just
curious.
You
say
we
think
it's
better
for
end-users.
Why
do
we
think
it's
better
for
end-users
and
who
are
those
end
users
right?
Do
we
expect
end?
Users
are
saying
I
really
care
about
a
new
networking
feature.
Therefore,
it's
better
that
they're
really
start
to
organize
into
networking
being
separate,
that's
storage,
and
that
could
totally
be
the
case,
but
I
would
be
surprised
if
any
users
want
or
should
need
to
care
about
the
substructure
we
create
within
our
project
to
be
SAS
I.
G
Guess
I
have
some
thoughts
on
that
I
think
it's
important
to
start
educating
the
users
of
the
project
so
when
they
inevitably
run
into
bugs
and
ask
for
feature
requests,
they
know
they
have
at
least
some
idea
in
their
in
their
head
and
I.
Don't
think
we
would
ever
move
away
from
having
component
based
release
notes
as
well,
but
just
a
heightened
level
like
what
the
sig
does.
What
they're
focused
on
and
I
linked
in
the
cig,
node
sig
level
theme,
austere
lifecycle.
G
My
slack
chat
just
like
a
link
to
where
that
landing
pages
in
the
community,
because
so
they
can
start
to
build
these
mental
models
in
their
head.
So
they
can
start
associating
that
the
couplet
is
owned
by
techno
term
ATM
is
owned
by
hustler,
cycl,
say
them
with
cops
just
to
help,
because
we
I
mean
we're
going
to
start
breaking
down
the
mono
repo,
and
so
they
will.
G
People
will
definitely
need
a
guide
to
help
them
and
so
I
think
everything
we
do
should
reinforce
the
underlying
structure
of
the
project,
and
just
this
is
just
one
way
one
attempt
of
trying
to
do
it
and
just
we're
gonna.
Try
it
we'll
see
if
people,
if
it's
useful
people
will
see
but
I
took
a
point.
Yeah.
D
Thanks
kill,
that's
that's
spot
on
also
that
what
we
didn't
want
to
do
was
try
and
manufacture
themes
like
look
at
this
and
try
and
figure
out
what
those
things
might
have
been
I
think
that's
worse
the
whole
the
whole
intent
of
release
notes
is
really
to
serve
people
who
are
trying
to
evaluate
communities
as
far
as
doing
one
upgrade
to
this
version
are
the
things
that
I
need
to
know
about
this
that
are
irrelevant
to
my
use
case.
What
is
the
impact?
D
One
of
the
things
that
I'm
getting
so
it's,
it's
really
less
developer
focused,
and
so
it's
a
really
interesting
balance
that
we
have
there
right.
What
is
the
right
level
of
detail
but
I
think
the
things
like
in
networking,
for
example,
see
and
I
work
is
just
absolutely
that's
crossing
release
boundaries.
It's
it's
huge!
It's
it's
so
important
to
the
to
the
the
lifecycle
of
communities
as
a
whole
that,
in
my
opinion,
deserves
to
be
called
out
that
isn't
that's
a
big
deal
and
there
are
there
sinks
with
equal.
D
D
All
right
moving,
hopefully
rapidly
along
that
there's
a
time
line
in
the
kubernetes
community
meeting
notes,
basically
that
links
to
the
timeline
for
the
release
that
we
developed
eons
ago,
which
is
varying
levels
of
correct,
but
I
share
it
because
that's
a
that
is
our
playbook.
To
date.
The
big
things
this
week
are
that
we
are
trying
to
do
code
unfreeze.
D
You
want
to
have
to
have
the
requirements
in
place
for
that,
so
1.9
essentially
will
the
work
on
that
will
begin.
The
moment
that
that
gets
cut
over
we're
doing
tech
review
on
documents
and
there's
we're
reviewing
and
triaging
open
issues
that
open
issue
count
is
currently
70,
which
is
a
little
concerning,
but
we're
going
to
go
through
those
and
see
what
actually
are
released
blocking.
They
are
considered
release
blocking
by
default
until
they're,
either
pulled
out
of
the
milestone
or
closed.
D
So,
if
you're,
a
cig
with
one
of
those
issues,
please
put
all
hands
on
deck
to
resolve
or
pull
those
out
of
the
milestone
as
soon
as
possible.
Because
if
the
the
release
timing
slips,
it
will
absolutely
be
as
a
result
of
this,
and
if
it
is
a
result
of
that,
I'd
like
it
to
be
because
we're
actually
doing
the
right
thing
and
fixing
critical
issues
and
having
a
high-quality
release
as
opposed
to
just
atrophy
or
maybe
issues
not
getting
terminated
properly
or
resolved
properly.
D
There's
more
detail
in
this
about
the
nuts
and
bolts
about
this
week
feel
free
to
ask
any
questions
in
the
sig
release
channel.
If
you
have
those
next
week,
we'll
have
daily
burn
down,
meetings
will
cut
the
1.8
release
Kenan
on
Wednesday
and
the
my
big
ass
surfin.
The
CUNY
is
please,
please
start
playing
with
that
release.
Put
it
through
its
paces.
Do
whatever
you
need
to
do,
because
that
is
a
key
indicator
of
release
health
and
we'll
definitely
need
that
as
soon
as
possible.
D
E
Hi
this
is
anthony,
we're
four
one.
Six
great
thanks
for
1/6
n
is
alpha,
but
I
don't
have
anything
particular
to
call
out
there,
but
I
do
want
to
take
this
chance
to
reiterate
that
there's
an
issue
with
stateful
said
if
you're
starting
on
something
lower
than
1/7
you'll
need
to
watch
out
for
this
if
you're
upgrading,
either
to
one
seven
or
even
to
one
eight.
So
please
check
out
the
release
notes
for
169
there's
some
links
there
to
explain
what
the
action
is
at
work.
A
A
Ok,
so
first
it's
a
keep
control
plugin,
it
is
alpha
in
1:8
and
the
client
will
be
paid
in
one
night
and
the
document
already
under
the
next
series
to
keep
control
to
its
own
repo.
So
it
had
to
start
first
eliminate
eliminate
but
dependencies,
it's
working
in
progress
and
the
second
step
is
figure
out
how
we
can
test
it
and
how
to
release
it
after
move
it
out.
A
So
for
the
first
step,
so
we
need
to
break
back
bad
dependency,
including
not
depend
on
the
internal
tabs
and
also
move
some
shared
utility
head
to
to
repo,
which
is
criminality,
utils
and
divinity
common.
The
first,
the
YouTube
repo
has
been
set
up,
but
the
common
repo
is
not
yet
grant
who
is
set
up
at
the
beginning
of
q4
and
next
is
exploring
Bryan's
declarative
application.
Application
management
approach
basically
have
a
base
layer
at
the
bottom,
and
then
you
can
apply
multiple
overlay
on
top
of
the
under
this.
A
A
A
Currently,
the
code
base
for
this
is
really
hard
to
maintain
under
extent
and
then
reactor
Rea
architecture
in
the
apply
logic,
so
the
current
hell
would
flowing
applies
first
to
aget
and
then
calculate
the
gift
and
the
dance
and
the
patch
did.
The
new
approach
is
to
a
get
and
calculate
the
days
and
applied
patch
locally
and
Dan
to
a
food.
A
A
D
Totally
make
Brian
do
this,
but
he's
got
a
cold
and
he
can't
sew
as
a
dutiful
sick
person
we're
giving
our
update
quarterly
here,
which
is
basically
that
I've
got
a
link
to
our
meeting
minutes
and
recordings.
Our
meeting
time
is
60
minutes
and
links
every
Thursday.
It
happens
at
8:30
in
the
morning
sometime
when
p.m.
is
Eastern
and
1700
UTC.
We
have
an
alternating
schedule
of
office
hours
on
the
weekend.
The
full
meetings
on
the
other
this
week
constituted
a
full
meeting
next
week
will
be
office
hours
just
a
quick
bit
about
office
hours.
D
We
use
that
time
to
discuss.
That's
not
a
time
where
we
would
typically
make
decisions
or
have
procedural
related
things
happen
it's
really
about.
If
people
have
architecture,
questions
that
they
need
answered,
or
they
just
wanted
to
run
things
by
the
the
group
or
the
sick
leaders
or
sick
lead
just
generally
speaking,
and
that's
the
time
to
do
it,
it
also
is
sometimes
will
organize
the
backlog
of
things
that
we
need
to
bring
to
the
next.
C
D
There
is
a
architectural
backlog
in
github
and
there's
a
link
to
that
in
the
many
minutes
right
now.
Our
strategy,
since
we're
relatively
new
sig,
is
to
get
processes
in
place,
and
then
we
want
to
run
proposals
through
it
to
validate
our
approach.
The
processes
are
mainly
related
to.
How
do
we
review
proposals?
How
do
we
take
in
those
things
and
get
commentary
set
timelines
on
ratifying
or
asking
for
more
question
or
it's?
D
You
know
how
do
we
handle
the
entire
lifecycle
of
a
proposal
and
once
that's
in
place,
then
we're
going
to
go
back
and
retroactively
lis.
Look
at
the
the
proposals
and
make
sure
that
they
have
the
metadata
that
we
need
encapsulated
in
those,
and
the
metadata
proposal
is
is
out
in
that
meeting
minutes
as
well
and
we're
working
on
finalizing
the
sync
proposal
directory
structure
by
tomorrow,
Friday
the
15th,
and
essentially
that
is
there's
the
the
version
there
actually
is
in
flight.
D
So
there's
a
few
things
in
there
bugs
that
Brian's
going
to
work
on
today
in
general,
but
essentially
trying
to
make
sure
that
proposals
have
a
home
in
each
sig
and
that
allows
us
to
to
make
sure
that
we
know
who
to
contact
for
the
evolving
architecture
there
and
again.
So
ownership
is
clear,
there's
also
a
proposal
in
process.
It's
community
pull
request,
nine
six,
seven
and
that
is
related
to
how
we
handle
RFC's
as
a
community.
D
So
I,
hopefully,
will
get
some
good
feedback
on
that
and
close
that
out
by
the
29th
of
this
month,
that's
being
driven
largely
by
Caleb
and
yeah.
If
you
have
questions
about
any
of
this
hit
me
or
Brian
up
and
we'll
be
more
than
happy
to
help,
there's
also
the
cig
architecture,
slack
Channel
and
the
cigarette
texture
mailing
list.
Any
questions
about
that.
B
H
So
border
registration
closed
yesterday,
so
everyone
who's
either
applied
for
membership
status.
You
should
have
gotten
an
email
with
the
confirmation
of
your
status
and
if
you
see
the
second
link
there
on
the
notes,
Paris
has
put
together
a
voter
guide
that
has
a
list
of
all
the
important
dates,
including
when
the
election
is
going
to
start
who,
what
address
you're
going
to
get
the
email
from
which
is
a/cs,
cornell
edu
email
address,
so
you
might
have
to
whitelist
that
so
expect
something
from
that
address.
H
H
Candidates
are
gonna,
be
submitting
their
little
platform
pages,
so
you
can
kind
of
get
to
know
them
a
little
bit
better
know
why
they're
running
and
anything
that
you
need
will
be
put
in
this
voter
guide,
including
at
the
very
bottom,
a
link
to
the
email
address
of
the
current
bootstrap
committee.
If
you
have
any
questions,
concerns
or
anything
like
that
feel
free
to
ping
us
on
slack
or
via
email.
Thank
you.
B
B
All
right,
we
are
going
to
start
engaging
with
an
organization
called
outreach
org.
This
may
have
already
happened
in
the
past
with
certain
individuals,
slash
things,
get
a
quoit
and
get
acquainted
with
this
site.
This
is
a
fully
funded
stipend
internship
for
underrepresented,
folks
and
text.
It
acts
like
an
internship,
but
they
help
us
contribute.
The
docs
team
is
now
putting
together
a
project
proposal.
B
So
if
anybody
on
here
is
interested
in
mentoring,
anyone
would
be
an
underrepresented
individual
intact.
Please
get
in
contact
with
me.
The
excuse
me,
the
outreach
e
o
RG
information
is
going
to
be
in
the
notes
section
as
well,
but
this
is
not
a
call
to
action.
This
is
just
a
call
to
see
if
you're
interested
but
reach
out
to
me
individually,
either
slack
or
I'll.
Put
my
my
Google
email
address
on
the
docks
as
well,
but
anyone
else
have
any
other
quick
announce
actually,
no
I'm.
Sorry,
we
have
a
election
question.
Yeah.